A SUPPLEMENT TO · 2019. 6. 12. · P. David Adelson, MD, FACS, FAAP University of...

GUIDELINES FOR THE ACUTE MEDICAL MANAGEMENTOF SEVERE TRAUMATIC BRAIN INJURY IN INFANTS, CHILDREN,AND ADOLESCENTS

P. David Adelson, MD, FACS, FAAP

University of Pittsburgh/Children’s Hospital

Susan L. Bratton, MD, MPH

University of Michigan Health System

Nancy A. Carney, PhD

Oregon Health & Science University

Randall M. Chesnut, MD, FCCM, FACS


Hugo E. M. du Coudray, PhD

Portland State University, Oregon Health & Science University

Brahm Goldstein, MD, FAAP, FCCM


Patrick M. Kochanek, MD, FCCM

Safar Center for Resuscitation Research, University of Pittsburgh School of Medicine

Helen C. Miller, MD, FAAP


Michael D. Partington, MD, FAAP, FACS

Gillette Children’s Specialty Healthcare

Nathan R. Selden, MD, PhD


Craig R. Warden, MD, MPH, FAAP, FACEP

OHSU/Doernbecher Children’s Hospital, Tualatin Valley Fire & Rescue

David W. Wright, MD, FACEP

Emory University School of Medicine

A SUPPLEMENT TO

A JOURNAL OF THE SOCIETY OF CRITICAL CARE MEDICINE,THE WORLD FEDERATION OF PEDIATRIC INTENSIVE AND CRITICAL CARE SOCIETIES,AND THE PAEDIATRIC INTENSIVE CARE SOCIETY UKJuly 2003 Volume 4, Number 3

Foreword

Guidelines for the acute medical management of severe traumaticbrain injury in infants, children, and adolescents

Practice guidelines for physicianswho treat children with braintrauma are long overdue. A sig-nificant barrier to producing

guidelines has been the lack of data fromwell-designed, controlled studies that ad-dress each specific juncture of the acutetreatment phase. Our goal with this docu-ment was to assimilate the scarce data thatexist and present it within an evidence-based framework in order to provide treat-ment guidelines. With topics for whichthere were no evidence-based data, weworked as a group to achieve consensusand provided treatment options. To accom-plish this, we assembled a multidisciplinaryteam of clinicians and researchers, keepingin mind that the presence of multiple per-spectives would minimize bias. Althoughwe recognize that this list is not complete,it represents a multidisciplinary group ofclinicians and scientists with considerableexpertise in key areas relevant to the project.

We consider this work a “document inprogress” and are committed to its ongoingrevision and to incorporating additional ar-eas of expertise that may not currently berepresented. It is our goal that these guide-lines be used to distinguish important areasof research, so that future revisions willcontain more substantial evidence.

A number of acknowledgments must bemade. The template for our work has beenthe adult guidelines of the Brain TraumaFoundation. This previous work made im-portant distinctions in treatment that weused to formulate pediatric topics. As such,we are indebted to the Brain Trauma Foun-dation for their organization and supportfor the adult severe head injury guide-lines—and to the authors of that docu-ment. Due to the relatedness of the twodocuments, we have frequently referred toand quoted from the original adult docu-ment and its more recent revision (1, 2).

The International Brain Injury Associa-tion assembled the first working group forthis project in March 2000 and provided

additional funding for subsequent meetingsof the research team. We also wish to thankJansjörg Wyss (Chair and CEO), Tom Hig-gins (President, Spine), Steve Murray(President, Maxillofacial), and Paul Gordon(Group Manager of Programs) of SynthesUSA for their generous, unrestricted con-tributions for meeting and administrativecosts. The National Center for Medical Re-habilitation Research in the National Insti-tute of Child Health and Human Develop-ment and National Institute of NeurologicDisorders and Stroke contributed the fundsfor joint supplemental publication of thisdocument. We are indebted to Drs. MichaelWeinrich and Mary Ellen Michel at these twoinstitutes, respectively, for their assistance.

We wish to acknowledge the willingnessof Dr. Joseph Parrillo, editor in chief ofCritical Care Medicine, and Dr. Basil Pruitt,editor in chief of the Journal of Trauma, topartner with us in this important andunique triple supplement in the journalsCritical Care Medicine, Pediatric CriticalCare Medicine, and the Journal of Trauma.This was also facilitated through the effortsof John Ewers, publisher at Lippincott Wil-liams and Wilkins, and Deborah McBride,Director of Publications at the Society ofCritical Care Medicine. Simultaneouslypublishing this document as supplementsin three journals is unique and will ensuredissemination to an international audienceof more than 30,000 subscribers, crossingmultiple disciplines.

One of the greatest challenges in pro-ducing an evidence-based document withmultiple authors is the administrativemanagement of the project. The expertiseto meet this challenge was provided bypersonnel of the Evidence Based PracticeCenter of Oregon Health and ScienceUniversity, and the project would nothave succeeded without this resource.

Members of the research team belong toa number of important medical societiesthat have provided a background of supportthroughout this project. Included are theSociety of Critical Care Medicine, the Sec-tions of Neurotrauma and Critical Care andPediatrics of the American Association of

Neurologic Surgeons, the Congress of Neu-rologic Surgeons, the American Academyof Pediatrics, the American College ofEmergency Physicians, and the World Fed-eration of Pediatric Intensive and CriticalCare Societies. Because of temporal con-straints inherent in the preparation of thisdocument, it was not possible to obtainformal endorsement by all of the relevantsocieties. We thank these societies alongwith the Child Neurology Society, theAmerican Association for the Surgery ofTrauma, the International Trauma Anes-thesia and Critical Care Society, and theInternational Society for Pediatric Neuro-surgery for important feedback and are es-pecially grateful to those that gave expe-dited approval to the document.

Finally, and most sincerely, we thankeach person who served as an investigatorand coauthor on this project. We trustthat the uncompensated time and abso-lute commitment over three years willresult in improved outcomes for childrenwho sustain traumatic brain injury.

Nancy A. Carney, PhDDepartment of Medical Informatics

and Clinical EpidemiologyOregon Health & Science UniversityPortland, OR

Randall Chesnut, MDDepartment of Neurological SurgeryOregon Health & Science UniversityPortland, OR

Patrick M. Kochanek, MDSafar Center for Resuscitation

ResearchUniversity of Pittsburgh School of

MedicinePittsburgh, PA

REFERENCES

1. Bullock R, Chesnut R, Clifton GL, et al: Guide-lines for the management of severe head in-jury. Brain Trauma Foundation, American As-sociation of Neurological Surgeons, JointSection on Neurotrauma and Critical Care.J Neurotrauma 1996; 13:641–734

2. Bullock R, Chesnut R, Clifton GL, et al: Guide-lines for the management of severe head in-jury. Brain Trauma Foundation, American As-sociation of Neurological Surgeons, JointSection on Neurotrauma and Critical Care.J Neurotrauma 2000; 17:449–627

Copyright © 2003 by Oregon Health & ScienceUniversity, 3181 SW Sam Jackson Park Road, Port-land, Oregon 97201.

DOI: 10.1097/01.CCM.0000067635.95882.24

S1Pediatr Crit Care Med 2003 Vol. 4, No. 3 (Suppl.)

Introduction

Chapter 1: Introduction

P. David Adelson, MD, FACS, FAAP; Susan L. Bratton, MD, MPH; Nancy A. Carney, PhD;Randall M. Chesnut, MD, FCCM, FACS; Hugo E. M. du Coudray, PhD; Brahm Goldstein, MD, FAAP, FCCM;Patrick M. Kochanek, MD, FCCM; Helen C. Miller, MD, FAAP; Michael D. Partington, MD, FAAP, FACS;Nathan R. Selden, MD, PhD; Craig R. Warden, MD, MPH, FAAP, FACEP; David W. Wright, MD, FACEP

A comparison of pediatric withadult trauma, using the Na-tional Pediatric Trauma Regis-try, indicates that a greater

proportion of pediatric than adult traumainvolves traumatic brain injury (TBI) (1).However, because of difficulty in evaluat-ing treatments across age groups and de-velopmental phases for children withTBI, little substantial research has beenconducted to specify standard treatmentsfor acute care as well as inpatient andoutpatient rehabilitation (2). The persis-tence of some federal funding agencies todevote their resources exclusively tostudies of adult populations with TBIhighlights the fact that pediatric braininjury remains underinvestigated. Pediat-ric physicians are left to make deductionsfrom guidelines developed for adult pop-ulations (3) or to call upon their clinicalexperience to make individual treatmentdecisions.

Evidence-based medicine is playing anincreasing role in the direction of medi-cal practice. As an empirical, unbiasedanalysis of the state of the literature on agiven topic, an evidence report summa-rizes vast bodies of literature such thatthe practitioner can more confidentlybase therapeutic choices on a scientificfoundation. By this means, one is able tounderstand not only the evidence sup-porting various therapeutic options butalso the rigor of the evidence. A manage-ment decision based on solid (class I)evidence should offer the highest degreeof confidence that the correct choice hasbeen made. At the other extreme, a deci-sion based on much less solid (e.g., classIII) evidence may be equally efficacious,but it alternatively may not be the bestchoice, and the results of that decisionshould be followed closely so that alter-ations can be expediently made if neces-sary. The majority of management deci-sions are not and never will be supportedby class I evidence. As such, knowing thedegree of support for a given decision, aswell as the alternatives, should be an ex-tremely valuable adjunct to the practiceof medicine.

The evidence-based “Guidelines forthe Management of [Adult] Severe Trau-matic Brain Injury” for adults (4) havebeen exceptionally well received by alarge audience of physicians, paramedicalpersonnel, and administrators around theworld. Although to a lesser extent thanwould be optimally expected, the adultguidelines have changed practice and im-proved outcome from severe TBI. Nota-bly, however, the evidentiary foundationon which they are based did not specifi-cally address the pediatric population.One of the most common questions askedfollowing presentations of the “Guide-lines for the Management of [Adult] Se-vere Traumatic Brain Injury” addresses

how they apply to children. The properanswer to date has been that this is un-known. Since it is not true that “childrenare just little adults,” it is not proper togeneralize from the adult literature to thepediatric population. The present effort isan attempt to address this problem byproviding a companion to the “Guidelinesfor the Management of [Adult] SevereTraumatic Brain Injury” that addressesissues from that work as they apply tochildren while additionally covering as-pects of TBI care that are unique to ordifferently developed in the pediatric pop-ulation. A number of the pediatric recom-mendations at the option level mirrorthose in the adult guidelines. In both theadult and pediatric guidelines, these werederived based on consensus.

These guidelines address key issues re-lating to the management of severe TBIin pediatric patients with a GlasgowComa Scale score of 3–8. Pediatric isdefined as �18 yrs of age. Traumaticbrain injury is defined as primary or sec-ondary injury to the brain resulting froma traumatic etiology. Abusive head injury(identified by a variety of different termsincluding inflicted traumatic brain injuryand shaken baby syndrome, among oth-ers) is included in this category. Injuriesdue to mechanisms such as drowning,cerebrovascular accidents, or obstetricalcomplications are not addressed here.

There are a number of shortcomingsof relying solely on evidence reports todirect medical decision making. A majorfactor is that such works can only statewhat the literature supports. If an issuehas been only poorly or incompletely re-searched, little strong support for it willbe gleaned from the literature even if it isa currently widely accepted therapy orone that is very strongly believed to beeffective on a purely clinical basis. Anexample in the TBI literature is the use of

From the University of Pittsburgh (PDA); Universityof Michigan (SLB); Department of Medical Informaticsand Clinical Epidemiology (NAC), Department of Neu-rological Surgery (RMC, HEMdC, NRS), Department ofPediatrics (BG), Department of Emergency Medicine(HCM), Division of Pediatric Neurosurgery (NRS), Pedi-atric Neurotrauma (NRS), and Pediatric EmergencyServices (CRW), Oregon Health & Science University;Safar Center for Resuscitation Research, University ofPittsburgh School of Medicine (PMK); Pediatric Neuro-surgery (MDP), Gillette Children’s Specialty Healthcare;Emergency Medicine Research Center (DWW), EmoryUniversity School of Medicine.

The sections Degrees of Certainty, Classification ofEvidence, and Correlation Between Evidence and Rec-ommendations closely mirror the adult “Guidelines forthe Management of Severe Traumatic Brain Injury.”

Address requests for reprints to: Nancy Carney,PhD, Department of Emergency Medicine, OregonHealth Sciences University, 3181 SW Sam JacksonPark Road, Portland, OR 97201-3098.

Copyright © 2003 by Oregon Health & ScienceUniversity, 3181 SW Sam Jackson Park Road, Port-land, Oregon 97201.

DOI: 10.1097/01.CCM.0000066600.71233.01

S2 Pediatr Crit Care Med 2003 Vol. 4, No. 3 (Suppl.)

mannitol to treat intracranial hyperten-sion. This is one of the oldest treatments,and there is generally a strong medicalconsensus that it is an effective andproper response to elevated intracranialpressure. Because it became “clinicallyestablished” before today’s more rigorousmedical environment, there is surpris-ingly little empirical, strong scientific lit-erature on its use. This contrasts sharplywith the literature on a much newertreatment for intracranial hypertension,hypertonic saline, for which there is a fairbody of reasonably strong research. Suchnewer treatments, having to meet higherscientific standards, are often supportedby higher classes of evidence. This leadsto the commonly encountered but lessthan satisfactory situation where olderand more established therapies will befound to have less empirical support thannewer treatment modalities (e.g., manni-tol vs. hypertonic saline or ventriculardrainage vs. lumbar drainage). Theproper manner of addressing this issuewhen formulating treatment recommen-dations remains an unsettled and difficultproblem in developing treatment guide-lines.

Another difficulty is that generalizingfrom a study is inconsistent with a rigor-ous evidence-based process. Even if it isrecognized that a given factor is highlygermane to a specific question, it cannotbe addressed in the absence of data onthat factor in particular. An example fromthe “Guidelines for the Management of[Adult] Severe Traumatic Brain Injury” isthe list of factors that support intracra-nial pressure monitoring in patients witha Glasgow Coma Scale score of �8 and anormal admission computed tomographyscan at the level of a treatment guideline.These include the presence of two ormore of the following: age �40 yrs, uni-lateral or bilateral motor posturing, andthe presence or history of a systolic bloodpressure �90 mm Hg. In no way arethese believed to be the most definitive oronly factors that suggest a high likeli-hood of intracranial hypertension in co-matose patients with normal computedtomography scans—they are merely thefactors that arose from the class II studyupon which the management guidelinewas based. Such a restriction on general-ization maintains the scientific rigor ofthe evidence-based process. It is, how-ever, also extremely unsatisfying becauseit severely limits our ability to addressissues that are commonly thought to beimportant and it may produce statements

that superficially appear to be incom-plete. It is hoped that such shortcomingswill be corrected through future re-search.

Process Used in theDevelopment of TheseGuidelines

The project was initiated in March2000 during the 5th Annual Aspen Neu-robehavioral Conference. Three phases oftreatment for pediatric TBI were identi-fied as priorities for guidelines develop-ment—acute medical management, re-habilitation, and school/family/community reentry. Participants at theconference who are also investigatorswith the Evidence-Based Practice Center(EPC) of Oregon Health & Science Uni-versity (OHSU) agreed to take acute med-ical management as their topic. Theycontacted individuals who had previouslyworked on the topic and included addi-tional investigators from OHSU to assem-ble a multidisciplinary team.

Using the 14 topics from the adultguidelines as a place to begin, we addedpediatric-specific questions and arrived ata set of 18 key topics, including one thataddresses a critical pathway for treat-ment, or a treatment algorithm. Eachtopic was assigned a primary and second-ary author. With the assistance of theEPC’s reference librarian, we conductedMedline searches from 1966 to 2001 byusing a broad search strategy for eachquestion. Blinded to each other, and us-ing predetermined criteria, the primaryand secondary authors read abstracts andidentified studies for which full-text arti-cles would be retrieved. They then readthe studies and eliminated another set.They used as their general criterion thegoal of obtaining the best possible evi-dence. If studies at the best level of ran-domized trials were not available, thenthose at the next level down were ac-cepted. Thus, the level of evidence in-cluded for each topic in this documentvaries with what was available in the lit-erature.

A baseline requirement for a study tobe used for recommendations at the levelof standards or guidelines was that thestudy be about children or contain dataabout children that was reported sepa-rately from data reported about adults. Ifwe could not distinguish adult data fromchild data in a study, or if a study did notinclude children, we did not include it inthis review. In the case where there was

no pediatric literature for a topic or for asubset of a topic, we reviewed the adultguidelines and by consensus elected howto refer to that document in terms ofrecommendations for children. Thus, anumber of the pediatric recommenda-tions at the options level mirror those inthe adult guidelines.

Degrees of Certainty

Regarding the degree of certainty as-sociated with a particular recommenda-tion, the following terminology is themost widely accepted and is used in thisdocument.

● Standards: Accepted principles of pa-tient management that reflect a highdegree of clinical certainty

● Guidelines: A particular strategy orrange of management strategies thatreflect a moderate clinical certainty

● Options: The remaining strategies forpatient management for which there isunclear clinical certainty

Note that “guideline” is used both in aglobal sense, that is, clinical practiceguidelines, as well as in a more specificsense, as noted previously.

Classification of Evidence

When assessing the value of therapiesor interventions, the available data areclassified into one of three categories ac-cording to the following criteria:

● Class I evidence: randomized con-trolled trials—the gold standard ofclinical trials. However, some may bepoorly designed, lack sufficient patientnumbers, or suffer from other method-ological inadequacies.

● Class II evidence: clinical studies inwhich the data were collected prospec-tively and retrospective analyses thatwere based on clearly reliable data.Types of studies so classified includeobservational studies, cohort studies,prevalence studies, and case controlstudies.

● Class III evidence: most studies basedon retrospectively collected data. Evi-dence used in this class indicates clin-ical series, databases or registries, casereviews, case reports, and expert opin-ion.

Correlation Between Evidenceand Recommendations

Standards are generally based on classI evidence. However, strong class II evi-


dence may form the basis for a standard,especially if the issue does not lend itselfto testing in a randomized format. Con-versely, weak or contradictory class I ev-idence may not be able to support a stan-dard.

Guidelines are usually based on classII evidence or a preponderance of class IIIevidence.

Options are usually based on class IIIevidence and are clearly much less usefulexcept for educational purposes and inguiding future studies.

Pediatric Conceptual Model

Three dimensions constitute the con-ceptual model necessary to evaluate andunderstand outcomes from brain injuryin children and adolescents and the effectof interventions on those outcomes:

● Developmental category of outcome● Developmental phase of child at time of

injury● Injury severity

Because outcomes from brain injuryare observed in cognitive and behavioraldimensions, as well as somatic dimen-sions, evaluation of the recovery processin children is confounded by cognitiveand behavioral changes that occur as a

function of normal development (2). Fur-thermore, development within each di-mension accelerates and decelerates dur-ing different developmental phases.Injury severity, and the presence or ab-sence of multiple-system injuries, willalso interact with the child’s age to influ-ence outcome.

The matrix in Figure 1, adapted fromseveral sources (5–7), offers a frameworkfor considering the complexity of mea-suring outcomes from brain injury inchildren.

The horizontal axis represents the var-ious categories within which outcomescan be observed. The vertical axis repre-sents developmental phases that influ-ence expected performance within devel-opmental category. The third dimensionof injury severity, not represented on thismatrix, introduces more complexity tothe consideration of outcome. This modelapplies to all of the questions addressedin these guidelines. Its utility is to mapexisting research for each question;empty cells indicate topics for futurestudy.

An additional issue specific to pediat-ric trauma is that of intentional injury.The nature of the trauma and secondarycomplications are thought to be distinctin ways from many unintentional inju-

ries, requiring corresponding differencesin appropriate treatment. The delay be-tween injury and hospital admission, orbetween admission and recognition thatthe injury was intentional, can thwarttimely treatment decisions. Although atthe beginning of the project we agreed topay careful attention to this issue in ourreview of the literature, no studies pro-vided data of any kind to assist in makingdistinctions about treatment for inten-tional injury.

REFERENCES

1. Tepas JJ, DiScala C, Ramenofsky ML, et al:Mortality and head injury: The pediatric per-spective. J Pediatr Surg 1990; 25:92–96

2. Carney N, du Coudray H, Davis-O’Reilly C, etal: Rehabilitation for Traumatic Brain Injuryin Children and Adolescents. Evidence ReportNo. 2 (Suppl) (Contract 290–97–0018). Rock-ville, MD, Agency for Healthcare Research andQuality, 1999

3. Lueressen T: Clinical Neurosurgery. Volume46. Proceedings of the Congress of Neurolog-ical Surgeons. Seattle, WA, Congress of Neu-rological Surgeons, 1998

4. Bullock R, Chesnut RM, Clifton G, et al:Guidelines for the management of severe trau-matic brain injury. J Neurotrauma 2000; 17:451–553

5. Berger K: The Developing Person Through theLife Span. New York, Worth, 1998

6. Berk L: Development Through the Lifespan.Boston, Allyn & Bacon, 1998

7. Sigelman C, Shaffer DR: Life-Span HumanDevelopment. Pacific Grove, CA, Brooks/Cole,1995

Figure 1. Matrix.

T hese guidelines ad-

dress key issues re-

lating to the man-

agement of severe traumatic

brain injury in pediatric pa-

tients with a Glasgow Coma

Scale score of 3–8. Pediatric

is defined as �18 yrs of age.


Guidelines

Chapter 2: Trauma systems, pediatric trauma centers, and theneurosurgeon

I. RECOMMENDATIONS

A. Standards. There are insufficientdata to support a treatment standard forthis topic.

B. Guidelines. In a metropolitan area,pediatric patients with severe traumaticbrain injury (TBI) should be transporteddirectly to a pediatric trauma center ifavailable.

C. Options. Pediatric patients with se-vere TBI should be treated in a pediatrictrauma center or in an adult trauma cen-ter with added qualifications to treat chil-dren in preference to a level I or II adulttrauma center without added qualifica-tions for pediatric treatment.

D. Indications from Adult Guidelines.The adult guidelines (1) recommend orga-nized trauma systems as a guideline andthe services of a neurosurgeon as an optionin the treatment of brain trauma. They citestudies that demonstrate overall reductionin mortality rate after implementation oftrauma systems, and they use the Re-sources for Optimal Care of the InjuredPatient of the American College of Sur-geons Committee on Trauma (2) as a foun-dation for their recommendations.

II. OVERVIEW

Although a number of studies reportdecreased mortality rate with implemen-tation of trauma systems and use of pe-diatric trauma centers (3–6), recent re-search suggests that survival in certainsubgroups may not be improved. Mann etal. (7) found a significant increase indeaths due to TBI from pre- to postimple-mentation of Oregon’s trauma system forpatients who were injured in rural areasand transferred to a higher level of care.For patients who died, transfer time fromlevel 3 and level 4 rural hospitals in-creased after the trauma system was es-tablished. The number of patients withTBI who were transferred for neurologicexamination also increased. Authors sug-gest that trauma system protocols for ex-

peditious transfer may have the unin-tended result of subjecting unstablepatients to premature transfer.

Trauma systems, pediatric traumacenters, and caregivers who are specifi-cally trained to treat children are all com-ponents of a system of care designed toprovide better outcomes for patients. Forthis section, studies were identified thataddress isolated components of this sys-tem of care and present the findings. Itmust be emphasized that, ultimately,outcome is a function of the system andnot of its isolated components.

III. PROCESS

We searched Medline and Healthstarfrom 1966 to 2001 by using the searchstrategy for this question (see AppendixA) and supplemented the results with lit-erature recommended by peers or identi-fied from reference lists. Of 24 potentiallyrelevant studies, three were used as evi-dence for this question (Table 1).

IV. SCIENTIFIC FOUNDATION

Three studies, two retrospective (4, 6)and one prospective (5), provide limitedevidence of the influence of trauma systemsand pediatric trauma centers on mortalityrates for children who sustain moderate tosevere TBI. One of the three (6) also eval-uates the effect of being a pediatric traumacenter on the number of neurosurgical pro-cedures provided to patients. The numberof procedures could be considered a surro-gate indicator of intensity of treatment andtherefore an indirect link to outcome.

Pediatric Trauma Centers

Potoka et al. (6) conducted a retro-spective review of medical records of pa-tients 0–16 yrs old treated at pediatric oradult trauma centers in the state of Penn-sylvania between 1993 and 1997. Fourpatient groups were specified, according

to the type of trauma center in whichthey were treated:

● PTC: pediatric trauma center (n � 1,077)● ATC AQ: adult trauma center with

added qualifications to treat children(n � 909)

● ATC I: level I adult trauma center (n �344)

● ATC II: level II adult trauma center (n� 726)

Whereas the study included patientswith mild and moderate TBI, this evalu-ation is based on the patients with severeTBI (Glasgow Coma Scale score 3–8). De-pendent variables were mortality rate,number of neurosurgical procedures, andmortality rate for patients who receivedneurosurgical procedures.

Method of and criteria for referral andtransfer within the statewide system arenot discussed in this publication. Distri-butions for injury severity based on in-jury severity score are presented for theparent group of all traumas but not forthe subgroup of TBI.

This class III study suggests the fol-lowing:

1. Pediatric patients with severe TBIare more likely to survive if treatedin PTCs, or ATC AQs, than in level Ior level II ATCs.

2. The pediatric patient with severeTBI who requires neurosurgicalprocedures has a lower chance ofsurvival in level II ATCs vs. theother centers.

Johnson and Krishnamurthy (5) con-ducted a prospective, nonrandomizedcomparison of mortality rate among ad-mitted patients, some of whom weretransported directly to Children’s Hospi-tal in Washington, DC, a level I PTC, andsome of whom were first transported toother hospitals and then transferred toChildren’s Hospital in Washington, DC.


Patients included children 1–12 yrs of agetreated in neurosurgical services between1985 and 1988.

Severity stratification included mild(Glasgow Coma Scale score 13–15), mod-erate (Glasgow Coma Scale score 9–12)and severe (Glasgow Coma Scale score�8) TBI. Our present interest is only the

Table 1. Evidence table

Reference Description of StudyDataClass Conclusion

Potoka (6)2000

Retrospective medical record review of children treated for headinjury (GCS score range, 3–15; age, 0–16 yrs) at accredited traumacenters in Pennsylvania with data entered in the PennsylvaniaTrauma Outcome Study registry between 1993 and 1997. Data forthis review include moderate (GCS, 9–12; n � 588) and severe(GCS, 3 to 8; n � 2,468) patients, n � 3,056. GCS score is notspecified.

Independent variable: level of pediatric accommodation in traumacenter (PTC, ATC AQ, ATC I, ATC II).

Dependent variables: mortality, neurosurgical procedures, mortalityfor patients receiving neurosurgical procedures.

Analysis: Student’s t-test, Mann-Whitney U, �2, Fisher’s exact.No use of multivariate statistics. Results were not stratified by age.

Baseline differences in ages between groups are not accounted for.

III Survival higher in PTC or ATC AQ than levelI or II ATCs for severe TBI.

Equal chance of survival for severe TBIrequiring neurosurgery in PTC, ATC AQ,or level I ATC, but not level II ATC.

Equal chance of survival for moderate TBI,regardless of facility.

For moderate TBI, more likely to haveneurosurgery in PTC or level I ATC, and ifthey do, less likely to die; less likely tohave neurosurgery in ATC AQ or level IIATC, and if they do, more likely to die.

Johnson et al.(5) 1996

Prospective, nonrandomized comparison of direct (n � 135) vs.indirect (n � 90) transports to level I PTC. Children (n � 225;age, 1–12 yrs) seen by neurosurgical services at Children’s Hospitalin Washington, DC, between 1985 and 1988.

Severity stratified by admission GCS (moderate, 9–12; severe, �8).Independent variable: direct vs. indirect transport.Dependent variable: mortality.Analysis: �2, Mann-Whitney U, Fisher’s exact.No use of multivariate statistics. Baseline differences: for entire

sample (including milds), LOS in PICU was significantly shorterfor direct transport group; percent intubated at arrival was greaterfor indirect transport group; significantly greater child abuse andchild abuse as cause of death in indirect transport group; forpatients with GCS 3–8, trauma score was significantly higher indirect transport group (score � 9) than indirect transport group(score � 7).

II For severe TBI, survival higher for directtransport patients than indirect transportpatients.

Equal chance of survival for moderate TBI,regardless of transport method.

Hulka et al.(4) 1997

Population-based (Washington and Oregon) retrospective medicalrecord review of children �19 yrs old, hospitalized with at leastone discharge diagnostic code between 800 and 959 (excluding905–909, late effects of injury; 930–939, foreign bodies;958–trauma complications. Severity stratified by ISS (minor, 1–15;serious, �15).

Independent variable: presence or absence of trauma system.Dependent variable: mortality.Compared mortality between Oregon and Washington before (1985–

1987) and after (1991–1993) Oregon implemented its statewidetrauma system.

All traumas/Oregon/before: 14,082All traumas/Washington/before: 18,525All traumas/Oregon/after: 8,981All traumas/Washington/after: 12,991Numbers for head injury not reported per group.Analysis: Multiple logistic regression model to calculate risk adjusted

odds of death. IVs in model: trauma system, age, gender, severity,AIS scores, and multiple injuries (AIS score of 2 or more in morethan one AIS region).

III For all severity levels and all injuries, nosignificant difference between states inmortality before or after trauma system.

For severe traumas and all injuries, nosignificant difference between states inmortality before trauma system; mortalitysignificantly higher in Washington thanOregon after trauma system.

Before trauma system, decreasing age, andmaximum AIS head, chest, and abdomenassociated with risk of death. After traumasystem, maximum AIS head and abdomenremained.

GCS, Glasgow Coma Scale; PTC, pediatric trauma center; ATC, adult trauma center; AQ, with added qualifications; TBI, traumatic brain injury; LOS,length of stay; PICU, pediatric intensive care unit; IV, independent variables; AIS, Abbreviated Injury Score.

Table 2. Pediatric mortality after acute trauma in Oregon and Washington, before and afterimplementation of a statewide trauma system in Oregon

Interval

State

Oregon Washington

1985–1987 No trauma system (n � 14,082) No trauma system (n � 18,525)1991–1993 Statewide trauma system (n � 8,981) No trauma system (n � 12,991)


patients in this study who sustained se-vere brain injury: 56 who received directtransport and 42 who received indirecttransport. However, statistical signifi-cance was only reported for the overallgroup, which included patients with mildand moderate TBI. Mortality rate for allpatients was significantly greater in theindirect transport group (4.7%) than thedirect transport group (1.9%).

An important baseline difference be-tween groups was noted for severe TBIpatients. The trauma score was signifi-cantly higher in the direct transportgroup (n � 9) than the indirect transportgroup (n � 7), indicating that the pa-tients in the latter group were less stablephysiologically. Authors suggest that thisis better viewed as an outcome than abaseline difference and that the physio-logic deterioration occurred as a functionof delays in appropriate treatment due tothe transfer.

This class II study suggests that in thismetropolitan area, pediatric patients withsevere TBI are more likely to survive iftransported immediately to a PTC than iftransported first to another type of centerand then transferred to a PTC.

Trauma Systems

Hulka et al. (4) compared mortalityrates between two states (Oregon andWashington) during two periods of time:before (1985–1987) and after (1991–1993)Oregon implemented a statewide traumasystem (Table 2). This retrospective medi-cal record review was an evaluation of allinjured pediatric patients �19 yrs of agehospitalized with at least one discharge di-agnostic code indicative of acute trauma.The sample sizes for subgroups of patientswith TBI were not reported. Multiple logis-tic regression modeling was used to calcu-late the risk-adjusted odds of death. Inde-pendent variables were trauma system, age,gender, severity, Abbreviated Injury Sever-ity Scores, and multiple injuries.

For all severe traumas, the risk ofdeath was significantly higher in Wash-ington than Oregon after Oregon imple-mented its trauma system. For TBI, max-imum Abbreviated Injury Severity Scorefor head was the strongest predictor ofrisk of death both before and after imple-mentation of the trauma system, withlittle change in the odds ratio (1.25 be-fore and 1.29 after the trauma system).Thus, this class III study suggests no ef-fect of the trauma system on risk of mor-tality from TBI.

V. SUMMARY

Children with severe TBI are morelikely to survive if treated in pediatrictrauma centers or in adult trauma cen-ters specially equipped and staffed to ac-commodate pediatric patients. In a met-ropolitan area, direct transport to a PTCappears to increase survival rate overall.There has been no evaluation of func-tional outcome for this topic.

VI. KEY ISSUES FOR FUTUREINVESTIGATION

Large data sets have accumulated fromstudies evaluating trauma systems thatcontain sufficient sample size and variables

to allow multivariate analyses focused onspecific subgroups of patients. These datasets should be used to identify pediatricpatients with TBI, to stratify by age andinjury severity, and to evaluate outcomebased on differences in care such as traumasystems and PTCs. Unfortunately, outcomemeasures in existing studies are limited tomortality or very short-term morbidity.Prospective studies that link acute medicalmanagement with long-term outcome areneeded to understand the effect of systemsof care on children with TBI.

Novel methodological technology forevaluating systems from the discipline ofsystems science could be directly appliedto questions about medical systems ofcare to provide a better understanding ofboth the intended and unintended resultsof implementation of new systems.

REFERENCES


2. ACS-COT: American College of SurgeonsCommittee on Trauma. In: Resources for Op-timal Care of the Injured Patient. AmericanCollege of Surgeons (Eds). Chicago, AmericanCollege of Surgeons, 1999

3. Hall JR, Reyes HM, Meller JL, et al: The out-come for children with blunt trauma is best ata pediatric trauma center. J Pediatr Surg1996; 1:72–77

4. Hulka F, Mullins RJ, Mann NC, et al: Influenceof a statewide trauma system on pediatric hos-pitalization and outcome. J Trauma 1997; 42:514–519

5. Johnson DL, Krishnamurthy S: Send severelyhead-injured children to a pediatric traumacenter. Pediatr Neurosurg 1996; 25:309–314

6. Potoka DA, Schall LC, Gardner MJ, et al: Im-pact of pediatric trauma centers on mortalityin a statewide system. J Trauma 2000; 49:237–245

7. Mann NC, Mullins RJ, Hedges JR, et al: Mor-tality among seriously injured patients treatedin remote rural trauma centers before andafter implementation of a statewide traumasystem. Med Care 2001; 39:643–653

See APPENDIX on Next Page

C hildren with se-

vere traumatic

brain injury are

more likely to survive if

treated in pediatric trauma

centers or in adult trauma

centers specially equipped

and staffed to accommodate

pediatric patients.


APPENDIX: LITERATURE SEARCH STRATEGIES

SEARCHED MEDLINE AND HEALTHSTAR FROM 1966 TO 2001

Chapter 2. Trauma Systems, Pediatric Trauma Centers, and the Neurosurgeon

1. trauma centers/2. trauma systems.tw.3. 1 or 24. exp craniocerebral trauma/5. head injur$.tw.6. brain injur$.tw.7. 4 or 5 or 68. 3 and 79. limit 8 to human

10. limit 9 to (newborn infant �birth to 1 month� or infant �1 to 23 months� or preschool child �2 to 5 years� or child�6 to 12 years� or adolescence �13 to 18 years�)


Chapter 3. Prehospital airway management

I. RECOMMENDATIONS

A. Standards. There are insufficientdata to support treatment standards forthis topic.

B. Guidelines. Hypoxia must beavoided if possible and attempts made tocorrect it immediately. Supplemental ox-ygen should be administered.

There is no evidence to support anadvantage of endotracheal intubation(ETI) over bag-valve-mask (BVM) ventila-tion for the prehospital management ofthe airway in pediatric patients with trau-matic brain injury (TBI).

C. Options. If prehospital ETI is insti-tuted for pediatric TBI patients, then spe-cialized training and use of end-tidal CO2

detectors is necessary.D. Indications from Adult Guidelines.

Under Guidelines (1), the authors state“hypoxemia must be avoided, if possible,or corrected immediately. . . .Hypoxemiashould be corrected by administeringsupplemental oxygen.” Under Options,the authors state that the “airway shouldbe secured in patients who have severehead injury (GCS �9), the inability tomaintain an adequate airway, or hypox-emia not corrected by supplemental oxy-gen. Endotracheal intubation, if avail-able, is the most effective procedure tomaintain the airway.”

II. OVERVIEW

Large prospective randomized studiesin the prehospital setting addressing theeffects of hypoxia, abnormal ventilation,and inadequate airway and possible ben-efits of invasive airway management havenot been conducted in either adult orpediatric populations. A large prospectiveobservational cohort study using theTraumatic Coma Data Bank that includedsome prehospital information showedthat hypoxemia in the prehospital settingwas associated with worse outcomes inTBI patients (2).

Several studies suggest that hypoxiaduring prehospital care of children withsevere TBI is common. A small study of 25

consecutive pediatric trauma patientsshowed that 16% had pulse oximetry read-ings �75% and an additional 28% had areadings of 75–90% during prehospital care(3). Another study also demonstrated thehigh frequency of hypoxemia in prehospitalTBI patients. Of 131 patients who were ret-rospectively reviewed, 27% were hypoxicon arrival to the emergency department(4). A retrospective chart review of 72 pe-diatric patients admitted to a single level Ipediatric trauma center with a postresusci-tation Glasgow Coma Scale (GCS) score of6–8 demonstrated that 13% had a docu-mented hypoxemic episode somewhereduring the continuum of care from theprehospital setting to the intensive careunit. However, the presence of hypoxia wasnot statistically related to outcome in thisrelatively small study (5).

III. PROCESS

We searched Medline and Healthstarfrom 1966 to 2001 by using the searchstrategy for this question (see AppendixA) and supplemented the results with lit-erature recommended by peers or identi-fied from reference lists. Of 35 potentiallyrelevant studies, four were used as evi-dence for this question (Table 1).


Few studies have been conducted inthe prehospital airway management ofpediatric patients and only one that spe-cifically evaluated the role of intubationin severe pediatric TBI patients. The onlyprospective class II study involved a ran-domized controlled trial of the airwaymanagement of all pediatric patients seenin two large urban emergency medicalsystems. A total of 830 patients �12 yrsof age were randomized to airway man-agement by BVM vs. ETI on an odd-evenday allocation with a total of 23 (2.8%)protocol violations. No pharmacologicadjuncts were used, and end-tidal CO2

monitoring was documented in 77% ofintubated patients. Of the 420 patientsassigned to ETI, 115 (27%) only received

BVM; 177 of the remaining 305 were suc-cessfully intubated (success rate 73%),and three had unrecognized esophagealintubations. Overall, no benefit was foundin ETI in this study or in any of theprospectively derived patient subgroups.Among children with severe TBI, the sub-group analysis, using a strict intention-to-treat analysis, showed that eight of 25in the BVM group vs. nine of 36 in theETI group survived (odds ratio, 0.71; 95%confidence interval, 0.23–2.19), and twoof 25 in the BVM vs. four of 36 in the ETIgroup had a “good neurologic outcome”(odds ratio, 1.44; 95% confidence inter-val, 0.24–8.52). Although this was thelargest prospective prehospital study todate, there is significant risk of a type Ierror in the subgroup analysis for TBIpatients alone (6).

A large retrospective study that usedthe National Pediatric Trauma Registryphase 3 abstracted all records of patientswith Abbreviated Injury Severity score�4 if they received either BVM or ETI byprehospital providers. A total of 578 caserecords met this eligibility out of the totalregistry population of 31,464. Endotra-cheal intubation was used in 479 (83%)and BVM in 99 (17%). The two cohortsdid not differ in injury severity or mech-anism but did differ in age stratification(ETI group older), the use of intravenousfluids (81% of ETI, 71% of BVM, p � .05),the use of intravenous medications (39%of ETI vs. 23% of BVM, p � .05), andtransport by helicopter (67% of ETI vs.27% of BVM, p � .01). Forty-eight per-cent of each cohort died. Injury compli-cations of any organ system were lessfrequent in the ETI group (58%) vs. BVMgroup (71%, p � .05). Functional out-come using the Functional IndependenceMeasure in patients �7 yrs old showed anonsignificant trend in improved out-come in the ETI patients.

Another smaller retrospective study ata single tertiary referral level I pediatrictrauma center during 1987 evaluated ETIin patients in the field, at the referringfacility, or in the institution’s own emer-


gency department. The authors reporteda success rate of only nine of 16 in thefield (including one successful and oneunsuccessful cricothyrotomy) com-pared with 36 of 36 at the referringfacility (p � .0002) and 17 of 17 in alevel I emergency department (p �.001). The authors attributed fourdeaths in patients who received prehos-pital ETI to “major airway mishaps.”The small numbers did not allow ad-justment for injury severity (7).

Another registry-based retrospectivecohort study that used patients admittedto the 13 trauma centers in Los AngelesCounty from 1995 to 1997 evaluated pre-hospital ETI in patients with a GCS �8and a head Abbreviated Injury Severityscore �3. There were a total of 137 pa-tients aged 11–20 yrs old (ETI was notallowed in younger patients), amongwhom 22 were successfully intubated.Mortality rate was 19 of 22 in successfulETI group vs. 57 of 115 in BVM group(unadjusted mortality rate, 1.74; 95%confidence interval, 1.36–2.23; p � .001),and mortality rate was seven of ten inunsuccessful intubation vs. 57 of 115 inthe BVM group (unadjusted mortalityrate, 1.41; 95% confidence interval, 0.90–2.21; p � .325). There was no risk strat-ification within this pediatric age cohortin the study.

Key Elements From the AdultGuidelines Relevant to PediatricTBI

Several studies have documentedthat adult patients with severe TBI whodevelop hypoxia have decreased risk ofsurvival. A large prospective observa-tional study of 717 patients with severeTBI admitted to four level I acutetrauma centers that participated in theTraumatic Coma Data Bank reportedthat an episode of hypoxia (PaO2 �60mm Hg or apnea or cyanosis in thefield) was independently associated witha significant increase in morbidity andmortality rates from severe head injury(2).

Another large retrospective cohortstudy of patients with severe TBI (GCS�9 and head or neck Abbreviated InjurySeverity score �4) involved a total of1,092 TBI patients of whom 351 had iso-lated TBI. The investigators found that26% of patients intubated in the prehos-pital setting died compared with 36.2% ofnonintubated patients (p � .05), and inthe subgroup analysis of isolated TBI pa-tients, 22.8% of patients who were intu-bated before admission to a hospital diedcompared with 49.6% of nonintubatedpatients (p � 0.05) (8).

A study of 50 consecutive TBI pa-tients who were transported by an aero-

medical service and were intubated(median GCS 7, SD 2, age 5– 84 yrs)were prospectively evaluated by pulseoximetry. Among patients with a pulseoximetry reading �90%, 14.3% (threeof 21) died and 4.8% had severe disabil-ity, whereas 27.3% (six of 22) of pa-tients with a pulse oximetry of 60 –90%died and 27.3% (six of 22) had severedisability. Finally, among patients withpulse oximetry �60%, the mortalityrate was 50% (three of six) and severedisability rate was 50% (three of six, p� .005). This study reports a strongassociation of preintubation hypoxemiawith poor neurologic outcome in severeTBI patients, but there was no evalua-tion of different methods for airwaymanagement. The outcomes of pediat-ric patients were not reported sepa-rately (9 –11).

V. SUMMARY

There are no well-conducted, prospec-tive outcome studies with sufficientpower to evaluate the role of various air-way maneuvers in pediatric prehospitalTBI care. On one hand, there is clearevidence that hypoxemia leads to poorerneurologic outcome in both pediatric andadult TBI patients. There is ample evi-dence also that hypoxemia frequently oc-curs in the prehospital setting in this



Nakayama et al.(7), 1990

Trauma registry of all hospitalized patients atone tertiary pediatric center over 1 yr whounderwent ETI of which 14 of 605 total wereprehospital attempts.

II Only 8/16 successful with two cricothyroidotomy attempts (1/2successful), whereas 36/36 (p �.0002) at referring hospitaland 17/17 (p �.001) at tertiary hospital were successful.Insufficient numbers to stratify by severity.

Gausche et al.(6), 2000

RCT of prehospital airway managementalternating BVM vs. ETI by odd/even days over3 yrs. Total 820 eligible patients, of which 61were TBI alone.

II Using intention-to-treat analysis, 8/25 BVM vs. 9/36 ETIsurvived (OR, 0.71; 95% CI, 0.23–2.19) and 2/25 BVM vs.4/36 ETI had “good” neurologic outcome (OR 1.44, 95% CI0.24–8.52).

Murray et al.(10), 2000

Retrospective registry-based review of all severeTBI patients (age �11 yrs, GCS �8 with headAIS �3) admitted to a single combined adultand pediatric level I trauma center over 3-yrperiod.

II For patients 11–20 yrs non-risk-adjusted MR of prehospitalintubated patients 19/22 vs. nonintubated 57/115 (MR, 1.74;95% CI, 1.36–2.23; p � .001) and unsuccessful prehospitalintubation 7/10 vs. nonintubated 57/115 (MR, 1.41; 95% CI,0.90–2.21; p � .325). Risk adjustment only done for entirecohort. No survival benefit to attempted prehospitalintubation in severe TBI patients.

Cooper et al.(11), 2001

Retrospective analysis of 31,464 records ofNPTR-3 for severe TBI (head AIS �4,578patients) stratified by field airway managementETI (479/578) vs. BVM (99/578).

II Similar injury severity and mechanism, ETI group older, moreoften received intravenous fluids and medications, and moreoften transported by helicopter. MR 48% both groups; FIM�6 in ETI group 65.7% vs. BVM 65.2%, p � NS; ETI lesscomplications (58%) vs. BVM (71%), p �.05.

ETI, endotracheal intubation; RCT, randomized controlled trial; BVM � bag-valve-mask intubation; TBI, traumatic brain injury; OR, odds ratio; CI,confidence interval; GCS, Glasgow Coma Scale; AIS, Abbreviated Injury Severity; MR � mortality ratio; NPTR-3, National Pediatric Trauma Registry phase3; FIM, Functional Independence Measure; NS, nonsignificant.


patient population. On the other hand,there is evidence that successful prehos-pital intubation of infants and childrenrequires specialized training, and reportedsuccess rates are generally less than inadults. Furthermore, there is much lessevidence that aggressive prehospital airwaymanagement changes outcome for eitheradults or children. In the largest prospec-tive airway study (adult or pediatric) yet,there was no benefit overall or to any sub-group analyzed (including TBI) attributedto endotracheal intubation by paramedicsin pediatric patients (6).


As with most areas of clinical medi-cine, large randomized, controlled tri-als are lacking that would clearly definesuitable interventions including pre-hospital airway management for the pe-diatric TBI patient. Given that the larg-est prehospital randomized, controlledtrial study to date of pediatric emer-gency airway management includedvery few children with TBI, it is difficultto conceive that an even larger study ofTBI patients will be accomplished soon.

Short of this type of study, clinicianswill need to infer from smaller studieswhether there is a role for prehospitalintubation in pediatric TBI patients. Thepotential benefits of intubation in pre-venting and treating hypoxemia vs. theknown risks (unrecognized esophageal ormainstem bronchus intubation) will needto be carefully balanced.

REFERENCES

1. Bullock R, Chesnut RM, Clifton G, et al: Guide-lines for the management of severe traumaticbrain injury. J Neurotrauma 2000; 17:451–553

2. Chesnut RM, Marshall LF, Klauber MR, et al:The role of secondary brain injury in deter-mining outcome from severe head injury.J Trauma 1993; 34:216–222

3. Silverston P: Pulse oximetry at the roadside:A study of pulse oximetry in immediate care.BMJ 1989; 298:711–713

4. Cooke RS, McNicholl BP, Byrnes DP: Earlymanagement of severe head injury in North-ern Ireland. Injury 1995; 26:395–397

5. Kokoska ER, Smith GS, Pittman T, et al:Early hypotension worsens neurological out-come in pediatric patients with moderatelysevere head trauma. J Pediatr Surg 1998;33:333–338

6. Gausche M, Lewis RJ, Stratton SJ, et al: Ef-fect of out-of-hospital pediatric endotrachealintubation on survival and neurological out-come: A controlled clinical trial. JAMA 2000;283:783–790

7. Nakayama DK, Gardner MJ, Rowe MI: Emer-gency endotracheal intubation in pediatrictrauma. Ann Surg 1990; 211:218–223

8. Winchell RJ, Hoyt DB: Endotracheal intuba-tion in the field improves survival in patientswith severe head injury. Trauma Researchand Education Foundation of San Diego.Arch Surg 1997; 132:592–597

9. Stocchetti N, Furlan A, Volta F: Hypoxemiaand arterial hypotension at the accidentscene in head injury. J Trauma 1996; 40:764–767

10. Murray JA, Demetriades D, Berne TV, et al:Prehospital intubation in patients with se-vere head injury. J Trauma 2000; 49:1065–1070

11. Cooper A, DiScala C, Foltin G, et al: Prehos-pital endotracheal intubation for severe headinjury in children: A reappraisal. Semin Pe-diatr Surg 2001; 10:3–6



Chapter 3. Prehospital Airway Management

1. exp craniocerebral trauma/2. head injur$.tw.3. brain injur$.tw.4. 1 or 2 or 35. limit 4 to (human and English language and “all child �0 to 18 years�”6. vascular access.mp.7. bone marrow.mp.8. intraosseous.mp.9. exp intubation, intratracheal/ or “intratracheal intubation”.mp.

10. exp Airway obstruction/th [Therapy]11. “AIRWAY MANAGEMENT”.mp.12. 6 or 7 or 8 or 9 or 10 or 1113. 5 and 12

S hort of this type of

study, clinicians

will need to infer

from smaller studies

whether there is a role for

prehospital intubation in pe-

diatric patients with trau-

matic brain injury.


Chapter 4. Resuscitation of blood pressure and oxygenation andprehospital brain-specific therapies for the severe pediatrictraumatic brain injury patient

I. RECOMMENDATIONS


B. Guidelines. Hypotension should beidentified and corrected as rapidly as pos-sible with fluid resuscitation. In children,hypotension is defined as systolic bloodpressure below the fifth percentile for ageor by clinical signs of shock. Tables de-picting normal values for pediatric bloodpressure by age are available (1). Thelower limit of systolic blood pressure (5thpercentile) for age may be estimated bythe formula: 70 mm Hg � (2 � age inyears) (2). Evaluation for associated ex-tracranial injuries is indicated in the set-ting of hypotension.

C. Options. Airway control should beobtained in children with a GlasgowComa Score �8 to avoid hypoxemia, hy-percarbia, and aspiration. Initial therapywith 100% oxygen is appropriate in theresuscitation phase of care. Oxygenationand ventilation should be assessed con-tinuously by pulse oximetry and end-tidalCO2 monitoring, respectively, or by serialblood gas measurements.

Hypoxia (defined as apnea, cyanosis,PaO2 �60–65 mm Hg, or oxygen satura-tion �90%) should be identified and cor-rected rapidly. Hypoventilation (definedas ineffective respiratory rate for age,shallow or irregular respirations, fre-quent periods of apnea, or measured hy-percarbia) is also an indication for airwaycontrol and assisted ventilation with100% oxygen in the resuscitation phaseof care.

Blood pressure should be monitoredfrequently and accurately. Timely fluidadministration should be provided tomaintain systolic blood pressure in thenormal range. Charts with normal valuesbased on age are available (1). Median(50th percentile) systolic blood pressure

for children older than 1 yr may be esti-mated by the formula: 90 � (2 � age inyears) (2).

Sedation, analgesia, and neuromuscu-lar blockade can be useful to optimizetransport of the patient with traumaticbrain injury (TBI). The choice of agentsand timing of administration are best leftto local Emergency Medical Services pro-tocols.

The prophylactic administration ofmannitol is not recommended. Mannitolmay be considered for use in euvolemicpatients who show signs of cerebral her-niation or acute neurologic deterioration.

Mild prophylactic hyperventilation isnot recommended. Hyperventilation maybe considered in patients who show signsof cerebral herniation or acute neuro-logic deterioration, after correcting hypo-tension or hypoxemia.

D. Indications from the Adult Guide-lines. In the adult guidelines for the pre-hospital management of severe TBI (3, 4),specific age-dependent ventilatory rateswere provided as shown in Table 1.

II. OVERVIEW

In TBI literature on both children andadults, there is a growing understandingof the extreme sensitivity of the injuredbrain to secondary insults, both systemicand intracranial (3, 5–15). Secondary sys-temic insults are common in pediatricsevere TBI. The systemic secondary in-sults that appear to have the most impacton outcome are hypoxia and hypotension.

The adult neurosurgical literature hastraditionally defined hypotension as sys-tolic blood pressure �90 mm Hg. In chil-dren, hypotension can be defined as lessthan the 5th percentile of normal systolicblood pressure for age. However, itshould be emphasized that hypotension isa late sign of shock in children. Pediatricpatients may maintain their blood pres-

sure despite significant hypovolemia andclinical signs of shock. Signs of decreasedperfusion include tachycardia, loss ofcentral pulses, decreased urine outputbelow 1 mL·kg�1·hr�1, or increased cap-illary filling time of �2 secs.

In children, fluid resuscitation is indi-cated for clinical signs of decreased per-fusion even when an adequate blood pres-sure reading is obtained. Shock is almostnever due to head injury alone; evalua-tion for internal or spinal cord injury isindicated (16). Fluid restriction to avoidexacerbating cerebral edema is contrain-dicated in the management of the head-injured child in shock (17). If peripheralvascular access is difficult in children,intraosseous infusion of fluids and medi-cations is indicated (18).

Apnea and hypoventilation are com-mon in pediatric severe TBI. As in adults,hypoxia may be defined by PaO2 �60–65torr or oxygen saturation �90%. How-ever, hypoxia develops more rapidly inthe child than in the adult during apneaor hypoventilation (19). Central cyanosisis neither an early nor a reliable indicatorof hypoxemia in children. Also, adequateoxygenation does not necessarily reflectadequate ventilation. Respiratory rateand effort should be monitored and cor-rected to age-appropriate parameters.

III. PROCESS

We searched Medline and Healthstarfrom 1966 to 2001 by using the searchstrategy for this question (see AppendixA) and supplemented the results with lit-erature recommended by peers or identi-fied from reference lists. Of 133 poten-tially relevant studies, eight were used asevidence for this question (Table 2).


The negative impact of hypoxia andhypotension on the outcome of severe


TBI has been demonstrated repeatedly instudies of mixed adult and pediatric pop-ulations (3, 6, 7, 12–14). In these studies,hypoxia, hypercarbia, and hypotensionwere common and correlated with in-creased morbidity and mortality rates.

Hypoxia

Pigula et al. (10) analyzed the influ-ence of hypoxia and hypotension on mor-tality from severe TBI (Glasgow ComaScale �8) in two prospectively collectedpediatric (age �16 yrs) databases. Hyp-oxia was defined as an emergency depart-ment admission PaO2 �60 mm Hg. Forboth the single-center database (n � 58)and the multiple-center database (n �509, including the 58 from the singlecenter), the presence of hypoxia alone didnot significantly alter mortality rate. Thecombination of hypotension and hypoxiaonly slightly (and not significantly) in-creased the mortality rate over hypoten-sion alone. It was concluded that hypo-tension is the most influential secondaryinsult determining short-term mortalityrate. Hypotension with or without hy-poxia was associated with mortality ratesapproaching those found in adults. Nei-ther the degree nor the duration of hyp-oxia was quantified. The participatingcenters were well-developed pediatrictrauma centers. As such, the apparentdiminution in the effect of hypoxia onoutcome might partially reflect the in-creased availability of effective airwaymanagement protocols for the prehospi-tal situation.

Michaud et al. (20) found that level ofoxygenation was associated with bothmortality rate and the severity of disabil-ity of survivors. Concurrent chest inju-ries were strongly associated with in-creased mortality and morbidity rates.Children with PaO2 levels between 105and 350 mm Hg had significantly worseoutcomes than those with PaO2 �350mm Hg.

In a prospective study of 200 children,Mayer and Walker (21) found that mor-

tality rate was 55% in the presence ofhypoxia, hypercarbia, or hypotension andonly 7.7% without any of these factorspresent (p � .01). In a prospective cohortstudy by Ong et al. (22) in Kuala Lumpur,the presence of hypoxia increased theprobability of a poor outcome by two- tofour-fold. In the setting of abusive headtrauma, Johnson et al. (23) found thatapnea was present in the majority of pa-tients and 50% were also hypotensive. Itwas concluded that cerebral hypoxiaand/or ischemia was more strongly asso-ciated with poor outcome than mecha-nism of injury.

Resuscitation of Blood Pressure

Five studies directly addressed the in-fluence of early hypotension on outcomefrom TBI. The impact of hypertension onsurvival was also addressed in two stud-ies.

In the aforementioned study by Pigulaet al. (10), they reported an 18% inci-dence of hypotension (defined as either asystolic blood pressure �90 mm Hg or asystolic blood pressure �5th percentilefor age) on arrival to the emergency de-partment. A mortality rate of 61% wasassociated with hypotension on admis-sion vs. 22% among patients without hy-potension. When hypotension was com-bined with hypoxia, the mortality ratewas 85%. Hypotension was a statisticallysignificant predictor of outcome with apositive predictive value of 61%. Earlyhypotension negated the improvement insurvival from severe TBI that is generallyafforded by youth.

Kokoska et al. (24) performed a retro-spective chart review of all pediatric pa-tients admitted to a single level 1 traumacenter over a 5-yr period. They limitedtheir patient populations to children withnonpenetrating TBI with postresuscita-tion age-adjusted Glasgow Coma Scalescores between 6 and 8 (n � 72). Theyindexed secondary insults occurring dur-ing transport to the emergency depart-ment up through the first 24 hrs in theintensive care unit. Hypotension was de-fined as �5 min at or below the 5thpercentile for age according to the TaskForce on Blood Pressure Control in Chil-dren (1). The majority of hypotensive ep-isodes occurred during resuscitation inthe emergency department (39%) and thepediatric intensive care unit (37%). Pa-tients left with moderate and severe dis-ability had significantly more hypotensiveepisodes than those with good outcomes.

Michaud et al. (20) found that hypo-tension in the field and emergency de-partment was significantly related tomortality rate in children. In a data bankstudy from four centers, Levin et al. (25)found that outcome was poorest in pa-tients 0–4 yrs old, which was the groupthat demonstrated high rates of hypoten-sion (32%).

In a prospective series of 6,908 adultsand 1,906 children �15 yrs of age at 41centers, Luerssen et al. (26) found thathypotension was significantly associatedwith higher mortality rates in children.They reported a greater deleterious effectof hypotension in children than adults.Notably, children with severe hyperten-sion had the lowest mortality rate.

In a recent retrospective study, Whiteet al. (27) found that odds of survival insevere pediatric TBI increased 19-foldwith maximum systolic blood pressure�135 mm Hg, also suggesting that su-pranormal blood pressures are associatedwith improved outcome. In contrast, pre-vious retrospective studies (28, 29) corre-lated early arterial hypertension with aworse neurologic outcome.

Brain Injury Specific Treatmentsin Prehospital Management

There is no evidence specifically deal-ing with the efficacy of any of the keybrain-directed prehospital therapies, in-cluding sedation and neuromuscularblockade, mannitol, hypertonic saline, orhyperventilation on the outcome from se-vere pediatric TBI. The scientific founda-tion for the in-hospital use of theseagents is discussed in separate sections ofthis document. Extrapolation of their useto the prehospital setting may be appro-priate and is provided by consensus at thelevel of options in the recommendationssection.


The evidence-based review of the lit-erature on prehospital airway, breathing,and ventilation management in the adultTBI population published as the Guide-lines for Pre-Hospital Management ofTraumatic Brain Injury (4) produced twoclass II studies (6, 7) demonstrating thatprehospital hypoxia has a statistically sig-nificant negative impact on outcome.These studies led to the following recom-mendation, “Hypoxemia (apnea, cyanosis,

Table 1. Age-dependent ventilatory rates(breaths/min) for eucapnea and hyperventilation

AgeRate for

EucapneaRate for

Hyperventilation

Adults 10 20Children 20 30Infants 25 35




Johnson et al. (23),1995

Retrospective medical record and imaging review of 28children with confirmed child abuse with significant headinjury (75% male, 50% age �3 mos, stratified GCS). Thosewith GCS 3–8 included five shaken, seven impact injuries.

Presence of fracture, GCS, SAH, SDH, contusion, DBS, IVH,apnea, intubation, early seizure, retinal hemorrhage, andoutcome were noted.

Analysis: Two-way tables with Fisher’s exact test forcategorical and Cochran-Mantel-Haenszel test for ordinal.Significance at p �.05.

III No patient with clinical evidence of cerebral hypoxiaand/or ischemia had a good outcome.

Trauma-induced apnea causes cerebral hypoxia,which is more fundamental to outcome thanmechanism of injury.

Kokoska et al. (24),1998

Retrospective chart review, 1990–95 level I, single center, 72children (3 mos–14 yrs 6 mos) GCS 6–8.

Measures: age, gender, mechanism, injury type, durationventilation and length of stay. Presence of hypoxia,hypotension, or hypercarbia during transport, ED, OR, andfirst 24 hrs in PICU.

Analysis: ANOVA on continuous data. �2 or Fisher’s exact testfor nominal data. p �.05 was significant.

Ages: 0–4 yrs, 5–9 yrs, and �10 yrsTransport time in 15-min intervals.

III 97% survival.Early hypotension linked to prolonged length of stay

and worse 3-month GOS.

Levin et al. (25),1992

Prospective databank cohort study of 103 children (�16 yrs)with severe TBI (GCS �9) at four centers. Patients receivedCT scan and ICP monitoring “treatment protocol.”

Hypotension was defined as below the age-dependent lowerlimit of normal systolic blood pressure.

Measures: age, race, gender, mechanism, time to center, worstGCS, median ICP, pupillary reactivity, hypoxia, shock, masslesion, skull fracture, GOS at 6 mos (86%) and 1 yr (73%).

Analysis: mean/SD, box plots, logistic and linear regression.

III Outcome was poorest in 0–4 yrs age group, whichhad an increased incidence of evacuated subduralhematomas (20%) and hypotension (32%).

14–21% in all age ranges were hypoxic.

Luerssen et al.(26), 1988

Prospective series of 8,814 adult and pediatric TBI patientsadmitted to 41 metropolitan hospitals in NY, TX, and CA in1980–81. 21.6% pediatric TBI patients (1,906 �15 yrs)compared with adult TBI patients (6,908 �15 yrs).

Measures: age, gender, admission vital signs, injurymechanism, GCS post resuscitation, pupillary response,associated injury/AIS, “major symptoms,” brain injury byimaging or at surgery, and mortality rate before hospitaldischarge. Hypoxia not studied.

Profound hypotension: systolic BP 30 mm Hg below medianfor age.

Analysis: Two-by-two tables by Pearson’s �2 test with Yatescorrection. Ordered contingency tables by Mantel-Haentzel.Logistic regression for age vs. survival.

II Only hypotension was associated with highermortality rate in children. Children with severehypertension had the lowest mortality rate.

Both hypotension and hypertension were associatedwith higher adult mortality rate.

Pediatric mortality rate was significantly lower thanadult mortality rate, with notable exceptions ofchildren with profound hypotension (33.3% �15yrs vs. 11.8% �15 yrs) or subdural hematoma(40.5% �15 yrs vs. 43.9% �15 yrs).

Age, even within the pediatric age group, is a majorindependent factor affecting TBI mortality.

Mayer and Walker(21), 1985

Prospective study (1978–1981) of 200 consecutive children (3wks–16 yrs, mean 5.6 yrs) with severe TBI (GCS �8).

124 male76 female43% IHI57% HI/MtMeasures: age, GCS, mass lesions, oculovestibular reflexes,

pupils, ICP, apnea, hypotension, hypoxia (PO2 �60),hypercarbic (�35 torr), multiple trauma, MISS score.

Interventions: ICP monitor for GCS �6, GCS 6, 7, andabnormal CT. ICP �20 (79%) received hyperventilation(PCO2 25–28 torr), diuretic, and barbiturate protocol.

Analysis: �2.

III Mortality rate 21.5%IHI 10.5%HI�MT 30%33% fixed dilated pupils29% hypotension, hypercarbia, or hypoxia28% altered OVR26% mass lesionsMortality rate 55% with any hypotension,

hypercarbia, or hypoxia vs. 7.7% without. 88% ofHI�MT group had hypotension, hypercarbia, orhypoxia.

GCS �4, increased ICP, MISS �25, hypotension,hypoxia, hypercarbia significant (p �.01) for pooroutcome.

Michaud et al. (20),1992

Retrospective study of prospectively collected Trauma Registrydata in 75 children presenting to Harborview Medical Centerwith severe TBI (GCS �8) between January 1, 1985, andDecember 31, 1986.

Mean 8.2 yrs; 67% male; 16% �2 yrs; 65% 3–14 yrs; 19%�14 yrs.

Assessed fatality rate in system with advanced EMS andregional trauma center (83% received EMS field care).

Identified factors predictive of survival and/or disability. GOSat discharge from acute care hospital measured.

EMS, ED, hospital, autopsy records analyzed.Analysis: SPSS; logistic regression with EGRET; �2 or Fisher’s

exact test statistical significance; p � .05.

II 33% fatality rate60% associated injuries86% intentional injuries fatalMortality rate increased if hypotension or abnormal

pupils noted in the field.ISS and pupillary reactivity predicted survival; 72-hr

motor GCS and ED PO2 predicted disability.ED PO2 � 350 better outcome; PO2 105–350 same

outcome as hypoxic group.


or arterial hemoglobin oxygen saturation[SaO2] �90 mm Hg) must be avoided, ifpossible, or corrected immediately. Whenavailable, oxygen saturation should bemonitored on all patients with severe TBIas frequently as possible or continuously.Hypoxemia should be corrected by ad-ministering supplemental oxygen.” Is-sues specifically pertaining to the man-agement of airway are addressed inChapter 3.

The “Guidelines for the Managementof [Adult] Severe Traumatic Brain Injury”(3) and the Guidelines for Pre-HospitalManagement of Traumatic Brain Injury(4) produced class II (6, 7, 27, 30, 31) andclass III (32–37) evidence from the adultliterature that early hypotension is a sta-tistically significant and independent fac-tor associated with worsening outcomefrom TBI. From the evidence report onprediction of outcome from TBI (38), hy-potension was one of the five factorsfound to have a �70% positive predictivevalue for mortality. Despite the solid ev-idence of the negative influence of earlyhypotension on outcome from TBI, thereis much less evidence that reducing or

preventing such secondary insults im-proves outcome.

Regarding the use of brain-specifictherapies in the prehospital setting, the“Guidelines for the Management of[Adult] Severe Traumatic Brain Injury”(3) and the Guidelines for Pre-HospitalManagement of Traumatic Brain Injury(4) collectively addressed the use of seda-tion, neuromuscular blockade, mannitol,and hyperventilation in managing severeTBI during the prehospital period. Withrespect to sedation and neuromuscularblockade, they found no studies dealingdirectly with the effects of prehospital useof these agents on outcome from severeTBI. It was recommended at the optionlevel that “sedation, analgesia, and neu-romuscular blockade can be useful to op-timize transport of the head-injured pa-tient. Because no outcome studiesprovide guidance on the use of these ad-juncts, the timing and choice of agentsare best left to local Emergency MedicalServices (EMS) protocols.” (4)

The adult guidelines suggested nosupport for the prehospital use of man-nitol. However, in two studies deleteri-

ous effects were not reported. Anequally acceptable alternative positionwould be that mannitol is an effectivebut potentially hazardous method oflowering intracranial pressure and thatits use during the prehospital periodshould be specifically limited to the eu-volemic patient with evidence of cere-bral herniation (a definite decrease inthe level of consciousness, motor pos-turing or flaccidity, or pupillarychanges such as anisocoria or bilateralpupillary dilation). Prophylactic usecannot be supported.

The Guidelines for Pre-Hospital Man-agement of Traumatic Brain Injury (4)found no studies of the effect on outcomeof the use of hyperventilation during theprehospital period. Recommendations,based on adult studies of the influence ofhyperventilation used during the in-hospital period on physiologic indexes andoutcome, stated; “hyperventilation (20 bpmin an adult, 25 bpm in a child, and 30 bpmin an infant) is the first line of interventionin the patient with suspected cerebral her-niation.” (4) Prophylactic hyperventilationwas not supported.

Table 2. (Continued)


Ong et al. (22),1996

Prospective cohort study of 151 consecutive children (�15 yrs)admitted within 24 hrs of head injury (GCS �15) from 1993to 1994 in Kuala Lumpur.

Age groups: 0–4 yrs (n � 51); 5–9 yrs (n � 55); 10–14 yrs (n� 45).

Stratified GCS 3–5; 6–8Measured; age, gender, GCS admit and 24-hr, pupils, motor

response, deficits, major extracranial injury, mass lesion,skull fracture, hypotension, and hypoxia.

Follow-up GOS at discharge and 6 mos.Analysis: �2 for categorical variables. Student’s t-test for

continuous variables, association clinical/radiological factors,and outcome. Logistic regression for combination of factorsto predict poor outcome.

II Poor outcome related to GCS �8, abnormal pupils,motor deficits, hypoxia, hypotension, and extracranialinjury.

Hypoxia increases poor outcome by two- to four-fold insevere TBI.

Five independent factors predict poor outcome:GCS at 24 hrs hypoxia on admissionSAHDAIbrain swelling on CT

Pigula et al. (10),1993

Five-yr prospective cohort study of 58 children (�17 yrs) and amatched set of 112 adults with severe TBI (GCS �8).

Group I—normal BP and PaO2.Group II—hypotension or hypoxia or both. Adults compared to

this subgroup.ABG, BP, GCS, ISS, PaO2, age on admission, and survival were

measured.Analysis: Outcome by two-tailed �2, Fisher’s exact test, and

ANOVA with Bonferroni’s adjusted t-test.Statistical significance p �.05.

II Hypotension with or without hypoxia causes significantmortality rate in children compared with levels foundin adults (p � .9). Adequate resuscitation probablythe single most critical factor for optimal survival.

Survival increased four-fold with neither hypoxia norhypotension compared with either hypoxia orhypotension (p �.001).

When added cohort to NPTR/509 children: Hypotensionincreased mortality rate even without hypoxia (p�.00001). If both hypoxia and hypotension present,only slightly increased mortality rate than withhypotension alone (p � .056).

GCS, Glasgow Coma Scale; SAH, subarachnoid hemorrhage; SDH, subdural hemorrhage; DBS, diffuse brain swelling; IVH, intraventricular hemorrhage;ED, emergency department; OR, operating room; PICU, pediatric intensive care unit; ANOVA, analysis of variance; GOS, Glasgow Outcome Scale; TBI,traumatic brain injury; CT, computed tomography; ICP, intracranial pressure; AIS, Abbreviated Injury Severity; BP, blood pressure; IHI, isolated headinjury; HI, head injury; MT, multiple trauma; MISS, Modified Injury Severity Scale; OVR, oculovestibular reflexes; ISS, Injury Severity Score; EMS,emergency medical service; DAI, difuse axonal injury; ABG, arterial blood gas; NPTR, National Pediatric Trauma Registry.


V. SUMMARY

The literature on the influence of hyp-oxia and hypotension on outcome fromsevere TBI in adults is fairly clear. Hypo-tension and hypoxia are serious, and po-tentially preventable, secondary insultsthat significantly increase the morbidityand mortality rates of TBI.

Unfortunately, there is minimal spe-cific evidence to indicate that prehospitalprotocols effective in preventing or min-imizing hypoxic and hypotensive insultsimprove outcome. Therefore, despite theuse of multivariate statistics to attempt tocontrol for such confounding, the possi-bility remains that some, most, or allsecondary insults occurring during theprehospital period that are associatedwith poor recovery are simply manifesta-tions of the severity of injury and are nottreatable entities.

A similar argument may be made forthe pediatric literature on hypotension.Decreases in systolic blood pressure be-low some threshold (vide supra) appearto be quantitatively associated with wors-ening of recovery. As such, despite theabsence of treatment efficacy data, maxi-mizing efforts directed at rapid and com-plete volume resuscitation, coupled withprotocols to minimize volume loss, aremost consistent with the present body ofliterature and should be strongly empha-sized components of prehospital care.

The situation with respect to prehos-pital hypoxia in pediatrics is less clear. Incontrast to the adult literature, the onlystudy that looked at prehospital hypoxiain any detail found that the presence ofhypoxia alone did not significantly altermortality rate. Such a finding, if not sim-ply an artifact, could reflect either anincreased resistance of the pediatric pop-ulation to hypoxic insults in the face ofsevere TBI or, alternatively, unquantifiedefficiency of the prehospital care provid-ers in preventing or minimizing hypoxicinsults in the setting of these studies. Ingeneral, it is believed that the pediatricbrain recovers better than an adult brainfrom a given traumatic insult. Pigula etal. (10), however, stated that they be-lieved the occurrence of a hypotensiveepisode eliminated the improvement insurvival from severe TBI that is generallyafforded by youth. If an improved resis-tance to hypoxia is responsible for thelack of a demonstrated adverse influenceon outcome, it is unlikely that such re-sistance is absolute. Therefore, althoughit is not proper to suggest altering treat-

ment in the absence of specific data, it iscertainly reasonable to recommend thatthe hypoxia avoidance/airway protectionprotocols afforded the patients studied byPigula et al. (10) be set as a favorableexample. Although these protocols werenot specified, they did use a populationfrom a pediatric trauma registry as alarge part of their study cohort. Sincesuch patients were treated by pediatrictrauma centers, this population samplecannot be assumed to represent routinepediatric prehospital trauma care. Assuch, the article by Pigula et al. (10)would seem strong, albeit indirect, sup-port for basic life support/advancedtrauma life support and pediatric ad-vanced life support protocols to be uni-versally applied as a minimum.

There is no contributing scientific lit-erature on the role of the prehospitaladministration of brain-specific therapiesin improving outcome from pediatricTBI. For the same period in adults, thereis no literature on neuromuscular block-ade or hyperventilation, one study on asingle sedative agent with very limitedapplicability to TBI, and two studies thatindirectly address the prehospital admin-istration of mannitol. As such, the Guide-lines for Pre-Hospital Management ofTraumatic Brain Injury base their rec-ommendations on data from the in-hospital period and consensus opinion. Inthe absence of evidence that their recom-mendations should be specifically alteredfor the pediatric population, we have sug-gested that the adult guidelines be con-sidered as the first line of approach. Theone area where we differ in our approachis that of mannitol; the adult guidelinesdispute its use, whereas we conclude thatthe absence of evidence for or against thisagent is more consistent with the stancethat mannitol is an effective but poten-tially hazardous method of lowering in-tracranial pressure and that its use dur-ing the prehospital period should bespecifically limited to the euvolemic pa-tient with evidence of cerebral hernia-tion. Prophylactic use cannot be sup-ported.

The presence of hypoxia or hypoten-sion after severe TBI in children increasesmorbidity and mortality rates. Specificthreshold values for ideal levels of oxy-genation and blood pressure support inthe pediatric age group have not beenclearly defined. Guidelines are warrantedto support avoidance or rapid correctionof systolic blood pressure less than thesecond standard deviation of normal for

age or of clinical signs of shock, apnea orhypoventilation, cyanosis, oxygen satura-tion �90%, or PaO2 �60 mm Hg in chil-dren with severe head injury.

Early control of the airway and recog-nition and treatment of associated ex-tracranial injuries are indicated. Despiteendotracheal intubation, head-injuredchildren remain at high risk for hypox-emia, hypercarbia, and major airwaycomplications (39). The frequency ofcomplications in airway procedures sup-ports the use of protocols including med-ications for cerebral protection, anesthe-sia, pain control, and paralysis (40).

The “golden hour” clearly begins atthe time of trauma. Although it is recog-nized that the field care of any traumapatient is encumbered both by the natureof the injury as well as by the often un-favorable and sometimes hostile environ-ment in which it is encountered, it isapparent that whatever function is com-promised by secondary insults duringthat period is generally not amenable tofull recovery. It is therefore critical tooptimize the prehospital care of the TBIpatient. Ideally, this would be realized bybringing hospital-type care to the acci-dent scene. Given the enormous variabil-ity of the early posttrauma period and thegenerally challenging environment inwhich care must be delivered, such a con-cept is not realistic. Indeed, the very con-cept itself continues to be the topic ofraging debate (the “scoop and run” versus“stay and play” controversy). As such, it isexpedient to simply select what seem tobe the most salient points of prehospitalTBI management and address them in anevidence-based fashion.

The goal of initial resuscitation inboth adults and children is to prevent

R esuscitation and

stabilization of

the cardiovascu-

lar and respiratory systems

in the field, during transfer,

and in the hospital need to

be emphasized in an effort to

optimize outcome from se-

vere pediatric brain injury.


secondary brain injury by restoring oxygen-ation, ventilation, and perfusion. Resuscita-tion and stabilization of the cardiovascularand respiratory systems in the field, duringtransfer, and in the hospital need to beemphasized in an effort to optimize out-come from severe pediatric brain injury.

VI. KEY ISSUES FOR FUTURERESEARCH

Unfortunately, there is a lack of pediat-ric studies on the ability of protocols di-rected at minimizing or preventing hypo-tensive episodes to improve outcome fromTBI. Therefore, the link between the pre-dictive value of hypotension in predictingoutcome and the treatment value of pre-venting hypotension in improving out-come, albeit logical, remains conjectural.

Hypotension

The determination of treatmentthresholds for hypotension is not amena-ble to randomized controlled trials forethical reasons. As such, it is necessary toaddress this issue by using large, prospec-tively collected observational databasesthat allow analysis of this variable whilecontrolling statistically for confoundingvariables. It has also been suggested thatsupranormal blood pressures may be ac-ceptable or even associated with im-proved outcome in children with severetraumatic brain injury. Further investi-gation in this area is also warranted.

Given the critical need to minimize oreliminate prehospital hypotensive epi-sodes, randomized controlled trials ad-dressing management protocols are nec-essary. Study of the timing, amount, andcomposition of resuscitation fluids to beused is warranted. Given the evidence onthe efficacy of in-hospital administrationof hypertonic saline plus the adult datasupporting its use in the prehospital careof the adult TBI patient, a formal study ofhypertonic prehospital resuscitation inpediatric TBI should be considered.

Hypoxia

Given the unclear nature of the pedi-atric literature on prehospital hypoxia,the first order of research should be tofurther define the nature of its occur-rence and influence on outcome. Studiesare needed with sufficient patient popu-lations that have the statistical power tomake definitive statements. The level ofoxygenation during this period needs to

be accurately and repeatedly measured(such as by serial monitoring of periph-eral oxygen saturation in the field) toaddress the influence of thresholds ofmagnitude and duration of hypoxia. Thiswould allow us to assess the role of pre-hospital hypoxia on outcome as well as toaccurately compare the efficacy of variousmanagement methods. Finally, since mor-bidity rate is generally believed to be morerelevant to measuring outcome from hy-poxic insults than mortality rate, we needto use functional recovery measures as ourdependent variables in such investigations.

The effect of hypercarbia, with orwithout hypoxemia, on outcome is alsonot clearly defined and deserves investi-gation. Potential use of more aggressiveoxygenation variables in the resuscitationperiod deserves further investigation.

Brain-Specific Treatments in thePrehospital Setting

The absence of pediatric literature inthis area is striking. Clearly, we need toaccomplish quantitative evaluation of vari-ous methods of managing the pediatric pa-tient with suspected TBI. Comparativestudies of different approaches to patientsedation are fundamental to every aspect ofmanaging such patients. Similar studies re-garding the use of hyperventilation andmannitol are also required. Althoughagent-specific, controlled studies would beoptimal, a large, multiple-center prospec-tive observational study might be a reason-able first-order approach.

REFERENCES

1. Horan MJ: Task Force on Blood PressureControl in Children, report of the SecondTask Force on Blood Pressure in Children.Pediatrics 1987; 79:1

2. American Heart Association: Pediatric Ad-vanced Life Support. American Heart Associ-ation, 2000

3. Bullock R, Chesnut RM, Clifton G, et al:Guidelines for the management of severetraumatic brain injury. J Neurotrauma 2000;17:451–553

4. Gabriel EJ, Ghajar J, Jagoda A, et al: Guide-lines for Pre-Hospital Management of Trau-matic Brain Injury. New York, Brain TraumaFoundation, 2000

5. Chesnut RM: Avoidance of hypotension: Con-ditio sine qua non of successful severe head-injury management. J Trauma 1997; 42:S4–S9

6. Chesnut RM, Marshall LF, Klauber MR, et al:The role of secondary brain injury in deter-mining outcome from severe head injury.J Trauma 1993; 34:216–222

7. Fearnside MR, Cook RJ, McDougall P, et al:The Westmead Head Injury Project outcomein severe head injury. A comparative analysisof pre-hospital, clinical and CT variables. Br JNeurosurg 1993; 7:267–279

8. Jones PA, Andrews PJ, Midgley S, et al: Mea-suring the burden of secondary insults inhead-injured patients during intensive care.J Neurosurg Anesthesiol 1994; 6:4–14

9. Pietropaoli JA, Rogers FB, Shackford SR, et al:The deleterious effects of intraoperative hypo-tension on outcome in patients with severehead injuries. J Trauma 1992; 33:403–407

10. Pigula FA, Wald SL, Shackford SR, et al: Theeffect of hypotension and hypoxia on chil-dren with severe head injuries. J PediatrSurg 1993; 28:310–314

11. Winchell RJ, Simons RK, Hoyt DB: Transientsystolic hypotension. A serious problem inthe management of head injury. Arch Surg1996; 131:533–539

12. Price DJ, Murray A: The influence of hypoxiaand hypotension on recovery from head in-jury. Br J Accid Surg 1972; 3:218–224

13. Gentleman, D: Causes and effects of systemiccomplications among severely head injuredpatients transferred to a neurosurgical unit.Int Surg 1992; 77:297–302

14. Stocchetti N, Furlan A, Volta F: Hypoxemia andarterial hypotension at the accident scene inhead injury. J Trauma 1996; 40:764–767

15. Celli P, Fruin A, Cervoni L: Severe head trau-ma: Review of the factors influencing theprognosis. Minerva Chir 1997; 52:1467–1480

16. Weinberg JA: Head trauma. Indian J Pediatr1988; 55:739–748

17. Armstrong, PF: Initial Management of theMultiply Injured Child: The ABC’s.

18. McNamara RM, Spivey W, Unger HD, et al:Emergency applications of intraosseous infu-sion. J Emerg Med 1987; 5:97–101

19. American Heart Association: Pediatric Ad-vanced Life Support. American Heart Associ-ation, 1997

20. Michaud LJ, Rivara FP, Grady MS, et al: Pre-dictors of survival and severity of disabilityafter severe brain injury in children. Neuro-surgery 1992; 31:254–264

21. Mayer TA, Walker ML: Pediatric head injury:The critical role of the emergency physician.Ann Emerg Med 1985; 14:1178–1184

22. Ong L, Selladurai BM, Dhillon MK, et al: Theprognostic value of the Glasgow Coma Scale,hypoxia and computerized tomography inoutcome prediction of pediatric head injury.Pediatr Neurosurg 1996; 24:285–291

23. Johnson DL, Boal D, Baule R: Role of apneain nonaccidental head injury. Pediatr Neuro-surg 1995; 23:305–310

24. Kokoska ER, Smith GS, Pittman T, et al: Earlyhypotension worsens neurological outcome inpediatric patients with moderately severe headtrauma. J Pediatr Surg 1998; 33:333–338

25. Levin HS, Aldrich ER, Saydjari C, et al: Se-vere head injury in children: Experience ofthe Traumatic Coma Data Bank. Neurosur-gery 1992; 31:435–444

26. Luerssen TG, Klauber MR, Marshall LF: Out-


come from head injury related to patient’sage. J Neurosurg 1988; 68:409–416

27. White JR, Farukhi Z, Bull C, et al: Predictorsof outcome in severely head-injured chil-dren. Crit Care Med 2001; 29:534–540

28. Kanter RK, Carroll JB, Post EM: Associationof arterial hypertension with poor outcomein children with acute brain injury. Clin Pe-diatr 1985; 24:320–323

29. Brink JD, Imbus C, Woo-Sam: Physical recov-ery after severe closed head trauma in childrenand adolescents. J Pediatr 1980; 97:721–727

30. Miller JD, Becker DP: Secondary insults tothe injured brain. J Royal Coll Surg Edinb1982; 27:292–298

31. Miller JD, Sweet RC, Narayan R, et al: Earlyinsults to the injured brain. JAMA 1978; 240:439–442

32. Carrel M, Moeschler O, Ravussin P, et al:

Medicalisation pre-hospitaliere heliportee etagressions cerebrales secondaires d’originesystemique chez les traumatises craniocere-braux graves. Ann Fr Anesth Reanim 1994;13:326–335

33. Hill DA, Abraham KJ, West RH: Factors affectingoutcome in the resuscitation of severely injuredpatients. Aust N Z J Surg 1993; 63:604–609

34. Jeffreys RV, Jones JJ: Avoidable factors con-tributing to the death of head injury patientsin general hospitals in Mersey Region. Lan-cet 1981; 2:459–461

35. Kohi YM, Mendelow AD, Teasdale GM, et al:Extracranial insults and outcome in pa-tients with acute head injury—Relation-ship to the Glasgow Coma Scale. Injury1984; 16:25–29

36. Rose J, Valtonen S, Jennett B: Avoidable fac-

tors contributing to death after head injury.BMJ 1977; 2:615–618

37. Seelig JM, Klauber MR, Toole BM, et al: In-creased ICP and systemic hypotension dur-ing the first 72 hours following severe headinjury. In: Intracranial Pressure VI. MillerJD, Teasdale GM, Rowan JO, et al (eds). Ber-lin: Springer-Verlag, 1986, pp 675–679

38. Chesnut RM, Ghajar J, Maas AIR, et al: Earlyindicators of prognosis in severe traumaticbrain injury. J Neurotrauma 2000; 17:554

39. Nakayama DK, Gardner MJ, Rowe MI: Emer-gency endotracheal intubation in pediatrictrauma. Ann Surg 1990; 211:218–223

40. Nakayama DK, Waggoner T, VenkataramanST, et al: The use of drugs in emergencyairway management in pediatric trauma.Ann Surg 1992; 216:205–211



Chapter 4. Resuscitation and Prehospital Brain-Specific Therapies: Strategy A—Resuscitation ofBlood Pressure and Oxygenation

1. exp craniocerebral trauma/2. head injur$.tw.3. brain injur$.tw.4. 1 or 2 or 35. exp anoxia/ or “hypoxia”.mp.6. exp hypotension/ or “hypotension”.mp.7. 5 or 68. 4 and 79. limit 8 to (newborn infant �birth to 1 month� or infant �1 to 23 months� or preschool child �2 to 5 years� or

child �6 to 12 years� or adolescence �13 to 18 years�)

Strategy B—Integration of Brain-Specific Treatments Into the Initial Resuscitation of the Severe HeadInjury Patient

1. exp craniocerebral trauma/2. head injur$.tw.3. brain injur$.tw.4. 1 or 2 or 35. exp analgesia/ or exp analgesics, opioid/ or exp “hypnotics and sedatives”/ or Midazolam/ or Propofol/ or “sedation”.mp.6. neuromuscular blockade/ or “neuromuscular blockade”.mp.7. exp resuscitation/ or “resuscitation”.mp.8. “ACUTE CARE”.mp.9. exp emergency medical services/ or “prehospital”.mp.

10. exp Ambulances/11. exp intensive care units/12. intensive care/ or “intensive care”.mp.13. 5 or 6 or 7 or 8 or 9 or 10 or 11 or 1214. 4 and 1315. limit 14 to (newborn infant �birth to 1 month� or infant �1 to 23 months� or preschool child �2 to 5 years� or

child �6 to 12 years� or adolescence �13 to 18 years�)16. limit 15 to english language


Chapter 5. Indications for intracranial pressure monitoring inpediatric patients with severe traumatic brain injury

I. RECOMMENDATIONS


B. Guidelines. There are insufficientdata to support a treatment guideline forthis topic.

C. Options. Intracranial pressuremonitoring (ICP) is appropriate in infantsand children with severe traumatic braininjury (TBI) (Glasgow Coma [GCS] score�8).

The presence of open fontanels and/orsutures in an infant with severe TBI doesnot preclude the development of intracra-nial hypertension or negate the utility ofICP monitoring.

Intracranial pressure monitoring isnot routinely indicated in infants andchildren with mild or moderate head in-jury. However, a physician may choose tomonitor ICP in certain conscious patientswith traumatic mass lesions or in pa-tients for whom serial neurologic exami-nation is precluded by sedation, neuro-muscular blockade, or anesthesia.

D. Indications from Adult Guidelines.In the adult guidelines for the manage-ment of severe TBI (1), there were suffi-cient data to support a treatment guide-line mandating ICP monitoring inpatients with severe TBI. Even thoughpediatric-specific data only support a rec-ommendation for ICP monitoring as atreatment option, no evidence exists tosuggest that monitoring is less importantin children than adults.

Monitoring ICP is appropriate in pa-tients with severe head injury with anabnormal admission CT scan. Severehead injury is defined as a GCS score of3–8 after cardiopulmonary resuscitation.An abnormal CT scan demonstrates he-matomas, contusions, cerebral edema,and/or compressed basal cisterns.

Intracranial pressure monitoring isappropriate for patients with severe headinjury and a normal CT if two or more of

the following features are noted on ad-mission: motor posturing, systemic hypo-tension, or age �40 yrs.

Intracranial pressure monitoring isnot routinely indicated in patients withmild or moderate head injury. However, aphysician may choose to monitor ICP incertain conscious patients with traumaticmass lesions or for whom serial neuro-logic examination is precluded by seda-tion or anesthesia.

II. OVERVIEW

Published data and consensus practicesince the late 1970s suggest that inten-sive management protocols may reducethe incidence of secondary brain injuryafter severe TBI and thus improve sur-vival and outcome (2–6). A central fea-ture of such protocols is the monitoringof ICP and medical and/or surgical treat-ment of intracranial hypertension. Con-trol of ICP within the normal range isintended to maintain adequate cerebralperfusion pressure, oxygenation, andmetabolic substrate delivery and to avoidcerebral herniation events. The relativeimportance of avoiding ICP elevation, perse, and reduced cerebral perfusion pres-sure, is currently uncertain. However, se-vere perturbations in either value arelikely to be adverse and are commonlyassociated with each other and with pooroutcome (2, 5, 7–12). Indications for thetreatment of variations in cerebral perfu-sion pressure are addressed in a separatechapter of these guidelines.

No randomized controlled trial toevaluate the effect on outcome of man-agement of severe TBI with or withoutICP monitoring has been conducted inany age group. The obstacles to perform-ing such a study include the ethical con-cern of not monitoring ICP in controlpatients and the widely accepted use ofICP monitoring in major pediatric cen-ters involved in TBI research. Althoughmodern, ICP-focused, intensive manage-

ment protocols have almost unquestion-ably improved outcomes, these protocolshave involved the simultaneous change ofa number of other major variables: im-proved prehospital care, cranial CT imag-ing for accurate diagnosis of mass le-sions, increased use of trachealintubation and controlled ventilation,more aggressive use of enteral and par-enteral nutrition, more active medicalmanagement of acute injury, and morewidespread availability of formal rehabil-itation programs. During this period, ICPmonitoring has emerged as an acceptedpractice without being subjected to a ran-domized trial. Thus, it is difficult to sep-arate the beneficial or deleterious effectsof ICP monitoring, or other efforts to“manage” ICP, from the many other de-velopments in treatment of severe TBI.

III. PROCESS

We searched Medline and Healthstarfrom 1966 to 2001 by using the searchstrategy for this question (see AppendixA) and supplemented the results with lit-erature recommended by peers or identi-fied from reference lists. Of 18 potentiallyrelevant studies, 14 were used as evidencefor this question (Table 1).

Of note, many studies in this literaturespecifically exclude children �3 yrs ofage and those with head injury resultingfrom nonaccidental trauma (13).


There are two lines of evidence to sup-port the use of ICP monitoring in severepediatric TBI:

1. Strong evidence supports the associa-tion of intracranial hypertension andpoor neurologic outcome.

2. ICP monitoring and aggressive treat-ment of intracranial hypertension areassociated with the best reported clin-ical outcomes.




Barzilay et al. (14),1988

In a single-center, case-controlled study, ICP and CPPwere monitored in 41 severely head-injured childrenby using a subarachnoid bolt. Intracranialhypertension was treated with ventilatory andmedical interventions.

III ICPmax was 16 � 3 in survivors and 54 � 11 innonsurvivors.

CPPmin was 66 � 9 in survivors and 6 � 4 innonsurvivors.

Sharples et al. (28),1995

In a single-center, prospective observational study,ICP, CBF, AJVDO2, and CMRO2 were measured in18 severely head-injured children.

III In 98% of measurements, raised ICP was associated withlow CBF. Patients with good outcome had higher CBFin the first 24 hrs after injury than patients with pooroutcome.

Chambers et al. (15),2000

In a single-center, observational study, 207 adults and84 children with severe head injury underwent ICPand CPP monitoring.

III ICPmax predictive of poor (GOS) outcome was �35 mmHg in adults and children, whereas CPPmin was 55 mmHg in adults and 45 mm Hg in children.

Michaud et al. (16),1992

In a single-center, observational study, 51 of 75children with severe CHI underwent ICPmonitoring.

III 94% of children with ICPmax �20 mm Hg survived,whereas only 59% with ICPmax �20 mm Hg survived(p � .02). 48% of children with ICP elevation �1 hrsurvived, compared with 89% of children with ICPelevated for �1 hr. Outcome was also better in childrenwith ICP elevation for �1 hr.

Alberico et al. (17),1987

In a single-center, prospective, observational study,330 severely head-injured patients (100 pediatric)underwent ICP monitoring and management.

III Despite similar ICPs, pediatric patients had betteroutcomes than adults. This difference was most obviousin patients with ICP �20 mm Hg, but impropersubgroup analysis limits the validity of the authors’conclusion that no advantage of young age is present inhigh ICP groups. Even within the pediatric age group,younger age was associated with significantly betteroutcome. However, a small group of very young (0–4yrs) children in this study had poor outcomes.

Kassof et al. (18), 1988 In a single-center, retrospective, observational study,25 severely head injured children underwent ICPmonitoring. Children with elevated ICP were treatedwith mannitol and, if refractory to mannitol,barbiturates.

III Children with elevated ICP had an absolute lower survivalrate than children with normal ICP, although nostatistical analysis is presented.

Eder et al. (19), 2000 In a single-center, retrospective study of 1,108children with severe head injury, 21 had clinicaland radiographic evidence of focal brainstem injury.ICP monitoring data and other factors werecompared with outcome.

III Children with brainstem injury and ICPmax �40 had asignificantly higher incidence of death/vegetative state(GOS 1–2) than children with lower ICP (statisticalreanalysis of data presented in Table 1 [19]).

Esparza et al. (20),1985

In a single-center, observational study of 56 severelyhead-injured children, ICP monitoring, evacuationof mass lesions, hyperventilation, and other medicaltherapies were used.

III Children with ICPmax �40 mm Hg had significantlyhigher rate of “good” outcomes (statistical reanalysis ofdata presented in Table 2. [20]).

Bruce et al. (27), 1979 In a single-center, observational study, 40 of 80children with severe TBI underwent ICP monitoringand medical management, emphasizinghyperventilation therapy to control intracranialhypertension.

III Intracranial hypertension (ICP �20 mm Hg) wasmarkedly more prevalent in children without (80%)than with (20%) spontaneous motor function. 87.5% ofchildren achieved “useful” recovery, and 9% died.

Peterson et al. (25),2000

In a single pediatric center, 68 children with closedhead injury, CT demonstration of diffuse injury and/or mass lesion, and ICP �20 mm Hg were studiedretrospectively. These children received intravenousinfusion of 3% hypertonic saline as needed toreduce ICP � 20 mm Hg.

III Treatment effectively lowered ICP in these patients. Threepatients died of uncontrolled intracranial hypertension.

Downard et al. (22),2000

In a retrospective study, 118 brain-injured childrenwho underwent ICP monitor placement within 24hrs of injury were studied at two urbanneurotrauma centers comprising a statewide level Itrauma system.

III In a stepwise logistic regression analysis, ICP �20 mmHg was significantly associated with an increased risk ofdeath.

Cho et al. (24), 1995 At a single institution, 23 children under 2 yrs of agewere treated for abusive head trauma causing severeTBI. Six children with ICP �30 mm Hg weretreated with medical therapy alone, and 17 childrenwith ICP �30 mm Hg were treated with eithermedical therapy or medical therapy plusdecompressive craniectomy, in nonrandomizedfashion. Decompressive craniectomy effectivelyreduced ICP.

III Children with ICP �30 mm Hg and children with ICP�30 mm Hg before treatment with decompressivecraniectomy experienced improved survival comparedwith children with ICP �30 mm Hg treated withmedical management alone.


Nine studies involving 518 pediatricpatients have demonstrated an associa-tion between intracranial hypertensionand poor neurologic outcome and/or in-creased mortality rate (14 –22). Thesestudies, and their individual ICP thresh-olds associated with poor outcome (gen-erally �20 mm Hg; see Chapter 6), arelisted in Table 1.

Monitoring and aggressive ICP-di-rected therapy have produced the besthistorical outcomes in the treatment ofsevere pediatric TBI. Studies that usedthree different intervention strategies forICP control have produced similar andconvincing reductions in the expectedmortality and/or neurologic morbidityrate. These strategies include decom-pressive craniectomy (23, 24), hyperos-molar therapy (25), and hyperventila-tion therapy (26). Each of thesestrategies is based and depends on ac-curate, continuous monitoring of intra-cranial pressure.

Two prospective studies of effectivecontrol of refractory intracranial hyper-tension that used decompressive craniec-tomy have demonstrated improved neu-rologic outcome (23, 24). Cho andcolleagues (24) performed decompressivecraniectomy for medically refractory in-tracranial hypertension due to abusivehead trauma in a cohort of infants andtoddlers. They showed improved Chil-dren’s Outcome Score in the surgical co-hort compared with a nearly contempo-raneous cohort of nonoperated childrenfrom the same institution. Taylor andcolleagues (23) performed a prospective,randomized clinical trial of decompres-sive craniectomy for severe TBI in chil-dren aged 1–18 yrs. They showed a sig-nificantly reduced ICP in the first 48 hrsafter randomization in those patients un-

dergoing surgery, compared with thosereceiving maximal medical managementand ventricular drainage alone. Surgicalpatients also experienced a strong trendtoward improvement in Glasgow Out-come Score 6 months after injury, in thisrelatively small, single-institution, pilotstudy.

Peterson et al. (25) reported a retro-spective study on the use of continuousinfusion of 3% saline in 68 infants andchildren with severe TBI and refractoryintracranial hypertension. Intracranialpressure was monitored and therapy wastitrated to control ICP at �20 mm Hg.Although there was no concurrent con-trol group, only three patients died ofuncontrolled ICP, and mortality rate(15%) was lower than expected based ontrauma severity score (40%).

Bruce et al. (27) studied 85 childrenwith severe TBI by using ICP monitoringand aggressive therapy with hyperventi-lation and barbiturates to control ICP�20 mm Hg. Again, although there wasno concurrent control group, a 9% mor-tality rate and 87.5% good outcome wereachieved.

Although none of these studiesachieve class II evidence for long-termoutcome, they strongly support the as-sociation of ICP monitoring and theeffective management of intracranialhypertension with improved outcome.In addition, raised ICP in children mayresult in secondary brain injury due toalterations in cerebral blood flow phys-iology (26, 28, 29). Therefore, given thelow risk of intracranial pressure moni-toring (see Chapter 7), these data sug-gest a strongly favorable risk/benefit ra-tio for ICP monitoring in severepediatric TBI.

The issue of who is at risk for intra-cranial hypertension is less clear for in-fants and young children than adults withTBI. Shapiro and Marmarou (21) re-ported a retrospective evaluation of ICPmonitoring in a case series of 22 childrenwith TBI. They found that 86% of chil-dren with a GCS �8 had ICP values �20mm Hg during the course of their mon-itoring.

Although GCS and neurologic exam-ination remain the standard for theclinical evaluation of patients with TBI,these may be less sensitive in infantsand young children (30, 31). Certainimaging correlates of intracranial hy-pertension (such as compressed basalcisterns) are also poorly defined and canbe misleading in children (21, 32).Combined with a paucity of pediatric-specific studies, this results in limitedinformation about which pediatric TBIpatients are at risk for intracranial hy-pertension and should be monitored.However, given that pediatric head in-jury patients may suffer late deteriora-tion and death (31), monitoring certainpatients with moderate TBI (e.g., thoseundergoing sedation, neuromuscularblockade, or anesthesia for the manage-ment of extracranial injuries) may beappropriate.

The clinical evaluation of an infantwith TBI may be difficult, and a normalinitial cranial CT does not rule out thepossibility of intracranial hypertension(30–32). The presence of open fontanelsand/or sutures does not preclude the de-velopment of intracranial hypertension(24). Thus, intracranial pressure moni-toring can be of significant use in infantssuffering from severe traumatic brain in-jury due to abusive head trauma or othermechanisms.

Table 1. (Continued)


Taylor et al. (23), 2001 This single-institution PRCT evaluated 27 childrenwith severe TBI and intracranial hypertensionrefractory to medical management and ventriculardrainage. Children were randomized to continuedmedical therapy vs. bitemporal decompressivecraniectomy.

II Decompressive craniectomy significantly lowered ICPcompared with medical management during the 48 hrsfollowing randomization. There was a strong trendtoward improved neurological outcome 6 mos afterinjury in children who underwent decompressivecraniectomy.

Shapiro and Marmarou(21), 1982

Twenty-two children with severe TBI (GCS �8) weremonitored with external ventricular drains.

III Eighty-six percent of children with severe TBI had ICPs�20 mm Hg. “Diffuse cerebral swelling” on CT scanwas 75% specific for the presence of intracranialhypertension.

ICP, intracranial pressure; CPP, cerebral perfusion pressure; CBF, cerebral blood flow; AJVDO2, arterial jugular venous oxidation; GOS, GlasgowOutcome Scale; CT, computed tomography; PRCT, prospective, randomized controlled trial.



Severely head-injured patients (GCS�8) are at high risk for intracranial hy-pertension (10, 33). The combination ofsevere head injury and an abnormal headCT scan suggests a high likelihood (53–63%) of raised ICP (12). However, evenwith a normal admission CT scan, intra-cranial hypertension may be present (3,34).

Patients with mild and moderate headinjury (GCS 9–15) are less likely to sufferfrom intracranial hypertension than se-verely head-injured patients, and there-fore the small risk and expense of ICPmonitoring may be relatively less justi-fied. Furthermore, serial neurologic ex-aminations are more eloquent in thesepatients and likely of greater accuracy inmonitoring clinically significant changesin neurologic status. However, consciouspatients with traumatic mass lesions sug-gestive of a risk of neurologic deteriora-tion, such as diffuse brain swelling on CTor temporal lobe contusion, may be mon-itored based on the opinion of the treat-ing physician (4, 32). Inability to performserial neurologic examinations, becauseof pharmacologic sedation or anesthesia,may also influence a clinician’s decisionto monitor ICP in an individual patient(30, 31).

In adults, the presence of two of threeadverse factors (systolic hypotension,

unilateral or bilateral motor posturing,age �40) predicted a significant rate ofintracranial hypertension despite a nor-mal CT scan. Data collected predomi-nantly in adult patients suggest that de-tection and treatment of intracranialhypertension may protect cerebral perfu-sion, avoid cerebral herniation, and im-prove neurologic outcome (5, 10, 32, 33,35, 36). Thus, the “Guidelines for theManagement of [Adult] Severe Head In-jury” (1) recommend ICP monitoring inadults with an abnormal CT scan or withtwo or more of these variables.

Clinical signs of raised ICP are gener-ally associated with a change in the levelof consciousness and/or cerebral hernia-tion. Level of consciousness is not mea-surable in severely head-injured (andthus by definition comatose) patients,and herniation comes after severe andoften irreversible brain injury has oc-curred (37, 38). Direct, physiologic mon-itoring is thus the most accurate methodof determining ICP (38, 39).

Intracranial pressure data allow themanagement of severe head injury by ob-jective criteria. This is particularly im-portant because many, and perhaps all,medical and surgical measures for thetreatment of intracranial hypertensionhave significant potential adverse conse-quences (40–42). Thus, ICP monitoringallows the judicious use of interventionssuch as hyperosmolar therapy, sedatives,paralytics, barbiturates, and ventilatormanagement, with a defined end pointthat is correlated with clinical outcome.This may avoid potentially harmful,overly aggressive treatment.

In adults, a number of importantstudies suggest that ICP monitoringand intervention for intracranial hyper-tension may have a significant salutaryeffect on survival and outcome aftersevere head injury. Intensive manage-ment protocols including ICP monitor-ing have lowered mortality rates com-pared with historical controls andcompared with centers in other coun-tries not using monitoring techniques(2, 32, 43, 44). Eisenberg et al. (45)reported that in severely head-injuredpatients, those in whom ICP could bewell controlled with barbiturates hadbetter outcomes than those with refrac-tory intracranial hypertension. Finally,in a small, single-institution study ofpatients triaged according to the at-tending neurosurgery call schedule,mortality rate was more than four timeshigher in nonmonitored than in moni-

tored patients with severe head injury(46).

V. SUMMARY

Two lines of evidence support the useof ICP monitoring as a treatment optionin severe pediatric TBI. In addition,guideline level support in the adult liter-ature mirrors the pediatric evidence thatICP monitoring is of clinical benefit insevere TBI. Intracranial hypertensionmay be difficult to diagnose in infants andyoung children and is associated with de-creased survival and poor neurologic out-come. The presence of an open fontaneland/or sutures in an infant with severeTBI does not preclude the development ofintracranial hypertension or negate theutility of ICP monitoring. ICP monitor-ing is not routinely indicated in infantsand children with mild or moderate headinjury. However, a physician may chooseto monitor ICP in certain conscious pa-tients with traumatic mass lesions or inwhom serial neurologic examination isprecluded by sedation, neuromuscularblockade, or anesthesia.


Definitive evidence of the value of ICPmonitoring would require the perfor-mance of a prospective, randomized clin-ical trial of appropriate power. However,it appears unlikely that such a study willever be carried out. Study of the efficacyof specific, ICP-directed therapies onlong-term, age-appropriate, neurologicoutcome in infants and children isneeded. Further study of the efficacy ofICP-directed therapies in infants andyoung children with an open fontaneland/or sutures is needed. Similarly, addi-tional studies of ICP monitoring and ICP-directed therapies in infants and youngchildren with abusive head traumashould be conducted.

REFERENCES


2. Becker DP, Miller JD, Ward JD, et al: Out-come from severe head injury with early di-agnosis and intensive management. J Neuro-surg 1977; 47:491–502

3. Marshall LF, Smith RW, Shapiro HM: Theoutcome with aggressive treatment in severehead injuries. Part I. The significance of in-

T wo lines of evi-

dence support the

use of intracranial

pressure (ICP) monitoring as

a treatment option in severe

pediatric traumatic brain in-

jury (TBI). In addition,

guideline level support in the

adult literature mirrors the

pediatric evidence that ICP

monitoring is of clinical ben-

efit in severe TBI.


tracranial pressure monitoring. J Neurosurg1979; 50:20–25

4. Marshall LF, Gautille T, Klauber MR, et al:The outcome of severe closed head injury.J Neurosurg 1991; 75:S28–S36

5. Miller JD, Butterworth JF, Gudeman SK, etal: Further experience in the management ofsevere head injury. J Neurosurg 1981; 54:289–299

6. Ritter AM, Muizelaar JP, Barnes T, et al:Brain stem blood flow, papillary response,and outcome in patients with severe headinjuries. Neurosurgery 1999; 44:941–948

7. Chesnut RM, Marshall LF, Klauber MR: Therole of secondary brain injury in determiningoutcome from severe head injury. J Trauma1993; 34:216–222

8. Johnston IH, Johnston JA, Jennett WB: In-tracranial pressure following head injury.Lancet 1970; 2:433–436

9. Lundberg N, Troupp H, Lorin H: Continuousrecording of the ventricular fluid pressure inpatients with severe acute traumatic braininjury. J Neurosurg 1965; 22:581–590

10. Narayan RK, Greenberg RP, Miller JD, et al:Improved confidence of outcome predictionin severe head injury: A comparative analysisof the clinical examination, multimodalityevoked potentials, CT scanning and intracra-nial pressure. J Neurosurg 1981; 54:751–762

11. Troupp H: Intraventricular pressure in pa-tients with severe brain injuries. Trauma1965; 5:373–378

12. Narayan RK, Kishore PR, Becker DP, et al:Intracranial pressure: To monitor or not tomonitor? A review of our experience withsevere head injury. J Neurosurg 1982; 56:650–659

13. Kumar R, West CGH, Quirke C, et al: Dochildren with severe head injury benefit fromintensive care? Childs Nerv Syst 1991;7:299–304

14. Barzilay Z, Augarten A, Sagy Y, et al: Vari-ables affecting outcome from severe braininjury in children. Intensive Care Med 1988;14:417–421

15. Chambers IR, Treadwell L, Mendelow AD:Determination of threshold levels of cerebralperfusion pressure and intracranial pressurein severe head injury by using receiver-operating characteristic curves: An observa-tional study in 291 patients. J Neurosurg2000; 94:412–416

16. Michaud LJ, Rivara FP, Grady MS, et al: Pre-dictors of survival and severity of disabilityafter severe brain injury in children. Neuro-surgery 1992; 31:254–264

17. Alberico AM, Ward JD, Choi SC, et al: Out-come after severe head injury: Relationshipto mass lesions, diffuse injury, and ICPcourse in pediatric and adult patients. J Neu-rosurg 1987; 67:648–656

18. Kasoff SS, Lansen TA, Holder D, et al: Ag-

gressive physiologic monitoring of pediatrichead trauma patients with elevated intracra-nial pressure. Pediatr Neurosci 1988; 14:241–249

19. Eder HG, Legat JA, Gruber W: Traumaticbrain stem lesions in children. Childs NervSyst 2000; 16:21–24

20. Esparza J, M-Portillo J, Sarabia M, et al:Outcome in children with severe head inju-ries. Childs Nerv Syst 1985; 1:109–114

21. Shapiro K, Marmarou A: Clinical applicationsof the pressure-volume index in treatment ofpediatric head injuries. J Neurosurg 1982;56:819–825

22. Downard C, Hulka F, Mullins R, et al: Rela-tionship of cerebral perfusion pressure andsurvival in pediatric brain-injured patients.J Trauma 2000; 49:654–659

23. Taylor A, Warwick B, Rosenfeld J, et al: Arandomized trial of very early decompressivecraniectomy in children with traumaticbrain injury and sustained intracranial hy-pertension. Childs Nerv Syst 2001; 17:154–162

24. Cho DY, Wang YC, Chi CS: Decompressivecraniotomy for acute shaken/impact syn-drome. Pediatr Neurosurg 1995; 23:192–198

25. Peterson B, Kanna S, Fisher B, et al: Pro-longed hypernatremia controls elevated in-tracranial pressure in head-injured pediatricpatients. Crit Care Med 2000; 28:1136–1143

26. Bruce DA, Alavi A, Bilaniuk L, et al: Diffusecerebral swelling following head injuries inchildren: The syndrome of “malignant brainedema.” J Neurosurg 1981; 54:170–178

27. Bruce DA, Raphely RC, Goldbert AI, et al:Pathophysiology, treatment and outcomefollowing severe head injury. Childs Brain1979; 5:174–191

28. Sharples PM, Stuart AG, Matthews DSF, et al:Cerebral blood flow and metabolism in chil-dren with severe head injury. Part 1: Relationto age, Glasgow coma score, outcome, intra-cranial pressure, and time after injury.J Neurol Neurosurg Psychiatry 1995; 58:145–152

29. Muizelaar JP, Marmarou A, DeSalles AAF, etal: Cerebral blood flow and metabolism inseverely head-injured children: Part 1: Rela-tionship with GCS score, outcome, ICP, andPVI. J Neurosurg 1989; 71:63–71

30. Humphreys RP, Hendrick EB, Hoffman HJ:The head-injured child who “talks and dies.”A report of 4 cases. Childs Nerv Syst 1990;6:139–142

31. Lobato RD, Rivas JJ, Gomez PA, et al: Headinjured patients who talk and deteriorateinto coma. Analysis of 211 cases studied withcomputerized tomography. J Neurosurg1991; 77:161–162

32. Levin HS, Eisenberg HM, Gary HE, et al:Intracranial hypertension in relation tomemory functioning during the first year

after head injury. Neurosurgery 1991; 28:196–199

33. Marmarou A, Anderson RL, Ward JD, et al:Impact of ICP instability and hypotension onoutcome in patients with severe headtrauma. J Neurosurg 1991; 75:S59–S66

34. O’Sullivan MG, Statham PF, Jones PA, et al:Role of intracranial pressure monitoring inseverely head injured patients without signsof intracranial hypertension on initial CT.J Neurosurg 1994; 80:46–50

35. Eisenberg HM, Gary HE, Aldrich EF, et al:Initial CT findings in 753 patients with se-vere head injury. A report from the NIHTraumatic Coma Data Bank. J Neurosurg1990; 73:688–698

36. Gopinath SP, Constant CF, Robertson CS, etal: Critical thresholds for physiological pa-rameters in patients with severe head injury.Presented at the Congress of NeurologicalSurgeons, Vancouver, BC, 1993

37. Browder J, Meyers R: Observations on thebehavior of the systemic blood pressure,pulse, and spinal fluid pressure followingcraniocerebral injury. Am J Surg 1936; 31:403–426

38. Lundberg N: Continuous recording and con-trol of ventricular fluid pressure in neuro-surgical practice. Acta Physiol Neurol ScandSuppl 1960; 36:1–193

39. Feldman Z, Kanter MJ, Robertson CS, et al:Effect of head elevation on intracranial pres-sure, cerebral perfusion pressure, and cere-bral blood flow in head-injured patients.J Neurosurg 1992; 76:207–211

40. Muizelaar JP, Marmarou A, Ward JD, et al:Adverse effects of prolonged hyperventilationin patients with severe head injury: A ran-domized clinical trial. J Neurosurg 1991; 75:731–739

41. Kaufmann AM, Cardosa ER: Aggravation ofvasogenic cerebral edema by multiple dosemannitol. J Neurosurg 1992; 77:584–589

42. Chesnut RM, Crisp CB, Klauber MR, et al:Early, routine paralysis for intracranial pres-sure control in severe head injury: Is it nec-essary? Crit Care Med 1994; 22:1471–1476

43. Jennett B, Teasdale G, Galbraith S, et al:Severe head injury in three countries. J Neu-rol Neurosurg Psychiatry 1977; 40:291–295

44. Colohan AR, Alves WM, Gross CR, et al: Headinjury mortality in two centers with differentemergency medical services and intensivecare. J Neurosurg 1989; 71:202–207

45. Eisenberg HM, Frankowski RF, Constant CF,et al: High dose barbiturates control elevatedintracranial pressure in patients with severehead injury. J Neurosurg 1988; 69:15–23

46. Ghajar JB, Hariri RJ, Patterson RH: Im-proved outcome from traumatic coma usingonly ventricular CSF drainage for ICP con-trol. Adv Neurosurg 1993; 21:173–177





Chapter 5. Indications for ICP Monitoring

1. exp craniocerebral trauma/2. head injur$.tw.3. brain injur$.tw.4. 1 or 2 or 35. intracranial pressure/ or “intracranial pressure”.mp.6. intracranial hypertension/ or “intracranial hypertension”.mp.7. 5 or 68. 4 and 79. limit 8 to (newborn infant �birth to 1 month� or infant �1 to 23 months� or preschool child �2 to 5 years� or



Chapter 6. Threshold for treatment of intracranial hypertension

I. RECOMMENDATIONS



C. Options. Treatment for intracranialhypertension, defined as a pathologic el-evation in intracranial pressure (ICP),should begin at an ICP �20 mm Hg.

Interpretation and treatment of intra-cranial hypertension based on any ICPthreshold should be corroborated by fre-quent clinical examination, monitoringof physiologic variables (e.g., cerebralperfusion pressure), and cranial imaging.

D. Indications From Adult Guidelines.There are insufficient data to support atreatment standard for this topic (1).

Treatment for intracranial hyperten-sion should be initiated at an ICP upperthreshold of 20–25 mm Hg.

Interpretation and treatment of intra-cranial hypertension based on any ICPthreshold should be corroborated by fre-quent clinical examination and cerebralperfusion pressure (CPP) data.

II. OVERVIEW

The effect of intracranial hypertension,or pathologically elevated ICP, on outcomeafter severe head injury in children appearsto be related to both the absolute peak andduration of elevated ICP and the inverserelation between ICP and cerebral physio-logic variables (e.g., cerebral perfusion andcompliance). Quantitative guidelines for in-tracranial hypertension threshold valuesare needed for management of elevated ICPin children.

III. PROCESS

We searched Medline and Healthstarfrom 1966 to 2001 by using the searchstrategy for this question (see AppendixA) and supplemented the results with lit-erature recommended by peers or identi-fied from reference lists. Of 62 potentially

relevant studies, five were used as evi-dence for this question (Table 1).


Specific thresholds of ICP for institu-tion of therapy in children with severetraumatic brain injury (TBI) have notbeen established. However, it is clear thatprolonged periods of intracranial hyper-tension or large increases in ICP are as-sociated with poor outcome as evidencedin the following studies. It should benoted that in none of the cited studies didthe authors prospectively address ICPtreatment thresholds.

Pfenninger et al. (2) retrospectively re-viewed the monitoring of ICP in 24 pa-tients with severe TBI. They used a defi-nition of ICP elevation as “persistently”�20–25 mm Hg. The goal of the treat-ment regimen they followed was to main-tain ICP �20 mm Hg and abolish ICPelevations that were �25–30 mm Hg thatlasted for �3 mins. They reported thatextremely high, sustained ICP (�40 mmHg) was associated with death (p � .001);ICP between 20–40 mm Hg was associ-ated with moderate outcome (one dead,two severely disabled, 13 moderate orgood); and ICP �20 mm Hg was associ-ated with good outcome (one severelydisabled, three moderate or good).

Esparza et al. (3) performed a retro-spective review of 56 pediatric patientswith severe TBI (defined as GlasgowComa Scale score �8 for �6 after inju-ry). They used a treatment threshold ofICP �20 mm Hg. Surgical evacuation ofmass lesions was performed as needed,but no decompressive craniotomy wasdone. They found that the group of pa-tients with ICP �20–40 mm Hg had amortality rate of 28%, whereas the groupwith an ICP �40 mm Hg had a mortalityrate of 100%.

Cho et al. (4) performed a retrospec-tive review of 23 infants (mean age � 5.8months) with TBI due to abusive headtrauma. They found that outcome wasworse with ICP �30 mm Hg comparedwith ICP �20 mm Hg or ICP �30 mm

Hg treated with surgical decompression.They suggested that patients with ICP�30 mm Hg may be treated successfullywith medical treatment only and thatthere is a role for decompressive craniot-omy in patients with ICP �30 mm Hg.

Two additional studies described physi-ologic derangements associated with anICP threshold �20 mm Hg. In a prospec-tive study of 21 pediatric patients (meanage � 8 yrs) with severe TBI (GlasgowComa Scale score �8), Sharples et al. (5)documented an inverse relation betweenelevations in ICP �20 mm Hg for �10mins and cerebral blood flow (CBF) in 18patients with ICP monitoring (r � �.24, p� .009). In only two cases was ICP �20mm Hg associated with CBF equal to orabove the normal range. In 66 simulta-neous measurements of ICP and CBF, theauthors found that the mean CBF � 0.57mL·g�1·min�1 when the ICP was �20 mmHg, whereas in 56 measurements the CBFwas 0.47 mL·g�1·min�1 when the ICP was�20 mm Hg (p � .037). Shapiro and Mar-marou (6) reported a retrospective, nonran-dom case series of 22 children with TBI andICP monitoring to determine a predefined“pressure-volume index” (PVI; i.e., a mea-sure of cerebral compliance) produced bybolus withdrawal or injection into a ven-triculostomy catheter. They defined intra-cranial hypertension as either an ICP �20mm Hg for �10 mins or the presence ofplateau waves or spot elevations �30 mmHg in the ICP waveform with noxious stim-ulation. They found that ICP �20 mm Hgwas associated with a PVI �80% of pre-dicted; an ICP 21–40 mm Hg was associ-ated with a PVI 60–80%; and ICP �40 mmHg correlated with a PVI �60%. They con-cluded that elevated ICP �20 mm Hg wasinversely correlated with PVI, supporting arelationship between intracranial hyperten-sion and impaired cerebral compliance.


Initiation of ICP treatment at an upperthreshold of 20–25 mm Hg was sup-


ported as a treatment guideline (1). Noprospective, randomized clinical trial di-rectly compared ICP treatment thresh-olds and outcome. The largest study us-ing prospectively collected, observationaldata, controlling for a large number of

confounding prognostic variables, associ-ated the mean ICP, in 5 mm Hg steps,with outcome in a logistic regressionmodel and found 20 mm Hg to be theoptimal threshold value predicting pooroutcome. Multiple, small, noncontrolled

reports suggested a range of 15–25 mmHg of ICP. Only one prospective, doubleblind, multiple-center, placebo-con-trolled study in 73 patients demonstratedimproved outcome when ICP could becontrolled by using a threshold of 20 mmHg (7). This study was class II with re-spect to outcome.

Patients may herniate at intracranialpressures �20–25 mm Hg. However, thelikelihood of herniation depends on thelocation of an intracranial mass lesion.Thus, the choice of any threshold mustbe closely and repeatedly corroboratedwith the clinical examination and com-puted tomography imaging in an individ-ual patient. The “Guidelines for the Man-agement of [Adult] Severe TraumaticBrain Injury” (1) concluded that adequateCPP may be maintained in adults withintracranial pressures of �20–25 mmHg. Thus, in select cases, a higher limit ofacceptable ICP may be chosen as long asan adequate CPP can be maintained.

V. SUMMARY

Current pediatric data support defin-ing intracranial hypertension as patho-logically elevated ICP �20 mm Hg and atreatment option setting an ICP of 20mm Hg as an upper threshold abovewhich treatment to lower ICP generallyshould be initiated. There have beensome suggestions that lower values foryounger children may be used, althoughthere are no data to support this. Intra-cranial hypertension with pathologicallyelevated ICP following severe TBI in chil-dren increases morbidity and mortality.


Specific threshold values of ICP forinstitution of therapy in pediatric agegroups need to be clearly defined. Defin-ing age-specific and injury-mechanism-specific ranges for ICP and CPP is vital fordetermining future treatment recom-mendations. For example, should a lowerICP treatment threshold of 15–20 mm Hgbe used for infants? The critical value ofICP and its interaction with other cere-bral physiologic variables are major un-answered questions.

As we recognize the importance ofCPP and improve our ability to safelymaintain an adequate CPP somewhat in-dependent of ICP, the issue of an absolutevalue for ICP appears to be most closelyrelated to the risk of herniation, which



Pfenninger et al.(2), 1983

Retrospective review of 24 patients.Treatment threshold set at ICPpersistently elevated �20–25 mm Hg.Severely sustained ICP �40 wasassociated with death. Moderatelysustained or acute ICP elevations werenot associated with outcome.

III Supports using ICP �20–25 mm Hg as treatmentthreshold.

Esparza et al.(3), 1985

Retrospective review of 56 patients withGCS �8. MVA (n � 40), fall (n � 14),child abuse (n � 2). Treatmentprotocol called for anti-intracranialhypertensive therapies at ICP �20.Mortality rate was 28% in ICP 20–40mm Hg group vs. 100% in ICP �40mm Hg group. Outcome was better inICP �20 group (27 good, two poor)compared with ICP 20–40 (10 goodand four poor) and ICP �40 (0 goodand 13 poor).

III Outcome was better if ICP�20 compared with ICP20–40.

Poor outcome related toICP �20–40 mm Hg.

Suggests that ICP �20mm Hg is a validtreatment threshold.

Cho et al. (4),1995

Retrospective, single-center study ofoutcome following shaken babysyndrome in patients �2 yrs old.Patient groups were as follows:

(A) ICP �30 with medical treatment only(n � 6)

(B) ICP �30 with medical treatmentonly (n � 7)

(C) ICP �30 with surgical decompressivecraniotomy (n � 10). Outcome wasworse with ICP �30 mm Hg comparedwith ICP �20 mm Hg or ICP �30 mmHg treated with surgicaldecompression.

III Outcome was worse withICP �30 mm Hgcompared with ICP �20mm Hg.

Shapiro andMarmarou(6), 1982

Prospective nonrandom case series of 22patients. ICP treatment thresholddefined as ICP �20 � 10 min, plateauwaves or spot elevations �30 mm Hgwith noxious stimuli, or progressiveincreases in ICP �20 mm Hg. ICP�20 mm Hg was associated with PVIof �80% of predicted; ICP 21–40 mmHg was associated with PVI 60–80%;and ICP �40 mm Hg correlated withPVI �60%.

III Elevated ICP �20 mm Hgis inversely correlatedwith PVI.

Clinical signs of increasedICP �20 mm Hg arenot always apparent.

Sharples et al.(5), 1995

Prospective, descriptive study of 18patients. Treatment threshold used wasICP �20 mm Hg for �10 min.Authors found an inverse relationshipbetween CBF and ICP. In only twocases of ICP �20 mm Hg was CBFequal to or greater than the normalrange.

III CBF inversely related toICP.

CBF data support use ofICP treatment thresholdof �20 mm Hg toprevent cerebralischemia.

ICP, intracranial pressure; GCS, Glasgow Coma Scale; MVA, motor vehicle accident; PVI, pressure-volume index; CBF, cerebral blood flow.


seems to vary between patients andwithin patients over the course of theirtherapy. A method to estimate this “her-niation pressure” needs to be developed,

and the range of values where CPP isindependent of mean arterial and intra-cranial pressures needs to be determined.

Large, coordinated, multiple-center,randomized clinical trials are the bestmeans of addressing many of these unan-swered issues. A national database for se-vere TBI in children would be useful andprovide important information.

REFERENCES


2. Pfenninger J, Kaiser G, Lutschg J, et al: Treat-ment and outcome of the severely head in-jured child. Intensive Care Med 1983; 9:13–16

3. Esparza J, Portillo JM, Sarabia M, et al: Out-come in children with severe head injuries.Childs Nerv Syst 1985; 1:109–114

4. Cho D, Wang Y, Chi C: Decompressive crani-otomy for acute shaken/impact baby syn-drome. Pediatr Neurosurg 1995; 23:192–198

5. Sharples PM, Stuart AG, Matthews DS, et al:Cerebral blood flow and metabolism in chil-dren with severe head injury. Part I: Relationto age, Glasgow coma score, outcome, intra-cranial pressure, and time after injury. J Neu-rol Neurosurg Psychiatry 1995; 58:145–152

6. Shapiro K, Marmarou A: Clinical applicationsof the pressure-volume index in treatment ofpediatric head injuries. J Neurosurg 1982; 56:819–825

7. Eisenberg H, Frankowski R, Contant C, et al:High-dose barbiturate control of elevated in-tracranial pressure in patients with severehead injury. J Neurosurg 1988; 69:15–23



Chapter 6. ICP Threshold



T he critical value of

intracranial pres-

sure and its inter-

action with other cerebral

physiologic variables are ma-

jor unanswered questions.


Chapter 7. Intracranial pressure monitoring technology

I. RECOMMENDATIONS

A. Standards. There are insufficientdata to support a technology standard forthis topic.

B. Guidelines. There are insufficientdata to support a technology guideline forthis topic.

C. Options. In pediatric patients whorequire intracranial pressure (ICP) moni-toring, a ventricular catheter or an externalstrain gauge transducer or catheter tippressure transducer device is an accurateand reliable method of monitoring ICP.

A ventriculostomy catheter device alsoenables therapeutic cerebrospinal fluid(CSF) drainage.

D. Indications from Adult Guidelines.Recommendations from the adult guide-lines (1) were not based on a level ofevidence.

A ventricular catheter connected to anexternal strain gauge is the most accu-rate, low-cost, and reliable method ofmonitoring ICP. It also allows therapeu-tic CSF drainage. ICP transduction viafiberoptic or strain gauge devices placedin ventricular catheters provides similarbenefits but at a higher cost.

Parenchymal ICP monitoring with fi-beroptic or strain gauge catheter tiptransduction is similar to ventricular ICPmonitoring but has the potential for mea-surement drift.

Subarachnoid, subdural, epiduralmonitors (fluid coupled or pneumatic)and externally placed anterior fontanelmonitors are less accurate.

The overall safety of ICP monitoringdevices is excellent, with clinically signif-icant complications (e.g., infection andhematoma) occurring infrequently.

II. OVERVIEW

In patients for whom ICP monitoringis indicated, a decision must be made asto what type of monitoring device to use.The optimal ICP monitoring device is onethat is accurate, reliable, and cost-effective and that causes minimal patient

morbidity. We reviewed the scientific lit-erature on ICP monitoring in childrenand adults and propose a ranking basedon the currently available technology.

III. PROCESS

We searched Medline and Healthstarfrom 1966 to 2001 by using the searchstrategy for this question (see AppendixA) and supplemented the results with lit-erature recommended by peers or identi-fied from reference lists. Of 41 potentiallyrelevant studies, two were used as evi-dence for this question (Table 1).


The scientific discussion of ICP mon-itoring technology is divided into the fol-lowing pediatric and adult sections: A.ICP monitoring device accuracy and sta-bility; B. optimal intracranial location ofmonitor; and C. complications.

A. Intracranial Pressure MonitoringDevice Accuracy and Stability. There areno pediatric studies on this topic. In in-fants, external placement on an open an-terior fontanel has been used, but thereare no corroborative data on accuracy orstability of the device.

B. Optimal Intracranial Location ofMonitor. There are no pediatric studies onthis topic.

C. Complications. In a retrospectivestudy of 49 pediatric patients with TBI be-tween 2–16 yrs of age, Gambardella et al.(2) compared the accuracy of Camino cath-eter measurements of ICP to ventriculos-tomy catheter measurements. There were12 ventriculostomy catheters used and 37intraparenchymal Camino cathetersplaced. The Glasgow Coma Scale (GCS)scores were as follows: 3 (19%), 4 (8.5%), 5(12%), 6 (27.5%), 7 (15%), and 8 (10%).The authors found that for patients withGCS between 3 and 4, the Camino cathetermeasurements averaged 3–4 mm Hg lessthan the ventriculostomy catheter; for GCSof 4, the Camino averaged 2–3 mm Hgmore than the ventriculostomy catheter;and in patients with GCS between 3 and 8,

the Camino averaged 1 mm Hg less thanthe ventriculostomy catheter. For patientsstudied on the same day and with the sameGCS score, there was good correlation be-tween ICP measurements with the Caminovs. the ventriculostomy catheter (r � .73–.89).

Most studies define infection as a posi-tive CSF culture in ventricular and sub-arachnoid bolt monitors or a positive cul-ture of the intracranial device. A betterdefinition is bacterial colonization of thedevice rather than infection since therehave been no reports in large prospectivestudies of clinically significant intracranialinfections associated with ICP monitoringdevices (1). In a prospective uncontrolledcase series, Jensen et al. (3) reported com-plications associated with ICP monitoringtechnology. They reported no clinically sig-nificant infections associated with ICPcatheters. However, they studied the inci-dence of positive bacterial cultures of thecatheter tip (i.e., colonization) following re-moval in 98 children with TBI who receivedICP monitoring. Initial placement occurredin the pediatric intensive care unit (54%),emergency department (34%), or operatingroom (12%). The positive catheter tip cul-ture rate was 7% (all positive for Staphylo-coccus aureus) and did not correlate withwhere the catheter was initially placed (in-tensive care unit, n � 3; emergency depart-ment, n � 4; operating room, n � 0). Themean duration of catheter placement was 7days (range, 3–40 days). The average dura-tion of catheters with negative tip cultureswas 7.3 days, whereas the duration of thosewith positive tip cultures was 12.1 days (p� .013). However, excluding the one out-lier of 40 days, the average duration ofthose with positive tip cultures was 7.5 days(p � .7 compared with those with negativetip cultures). Loss of waveform occurred in13% of catheters, occurring at a mean of9.5 days (range, 3–15 days). Finally, in cath-eters that suffered loss of waveform, theaverage ICP mean value was 11.1 mm Hg(range, 4–23 mm Hg) greater than mea-sured when the catheter was replaced.


Key Elements from the AdultGuidelines Relevant to PediatricTBI

A. Intracranial Pressure MonitoringDevice Accuracy and Stability. The fol-lowing information is quoted from the“Guidelines for the Management of [Adult]Severe Traumatic Brain Injury” (1).

The Association for the Advancement ofMedical Instrumentation has developed theAmerican National Standard for Intracra-nial Pressure Monitoring Devices in associ-ation with a neurosurgery committee (4).The purpose of this standard is to providelabeling, safety, and performance require-ments and to test methods that will helpensure a reasonable level of safety and ef-fectiveness of devices intended for use inthe measurement of ICP.

According to the Association for theAdvancement of Medical Instrumenta-tion’s standard, an ICP device shouldhave the following specifications:

Pressure range: 1–100 mm Hg

Accuracy: �2 mm Hg in range of 0–20mm Hg

Maximum error: 10% in range of 20–100 mm Hg

Current ICP monitors allow pressuretransduction by external strain gauge, cath-eter tip strain gauge, and catheter tip fiber-optic technology. External strain gaugetransducers are coupled to the patient’s in-tracranial space via fluid-filled lines,whereas catheter tip transducer technolo-gies are placed intracranially. Externalstrain gauge transducers are accurate andcan be recalibrated, but obstruction of thefluid couple can cause inaccuracy. In addi-tion, the external transducer must be con-sistently maintained at a fixed referencepoint relative to the patient’s head to avoidmeasurement error.

Catheter tip strain gauge or fiberopticdevices are calibrated before intracranial

insertion and cannot be recalibrated onceinserted (without an associated ventricu-lar catheter). Consequently, if the devicemeasurement drifts and is not recali-brated, there is potential for an inaccu-rate measurement especially if the ICPmonitor is used for several days.

There is potential for significant ICPmeasurement drift with fiberoptic pressuretransduction and strain gauge pressuretransduction in the parenchymal space.However, adult studies of catheter tip straingauge ICP devices have demonstrated lowor negligible drift over 5 days (1). The ac-curacy of a pressure transduction devicecan be assessed by placing the device withinthe lumen of a ventricular catheter andcomparing the fluid-coupled ventricularpressure reading to the device being tested.Catheter tip fiberoptic and strain gauge de-vices tested in this manner show differ-ences (��2 mm Hg) compared with ven-tricular ICP readings. This method ofpressure transduction comparison may beerroneous when the ventricular catheter ismisplaced or occluded.

B. Optimal Intracranial Location ofMonitor. The following information isquoted from the “Guidelines for the Man-agement of [Adult] Severe TraumaticBrain Injury” (1).

A pressure transduction device for ICPmonitoring can be placed in the epidural,subdural, subarachnoid, parenchymal, orventricular location.

Historically, ventricular ICP is used asthe reference standard in comparing theaccuracy of ICP monitors in other intra-cranial compartments (4). It also has thetherapeutic benefit of draining CSF in theevent of intracranial hypertension. Thepotential risks of catheter misplacement,infection, hemorrhage, and obstructionhave led to alternative intracranial sitesfor ICP monitoring.

The following statements ensue fromreview of the adult and pediatric literature:

● Ventricular pressure measurement isthe reference standard for ICP moni-toring.

● ICP measurement by parenchymalcatheter tip strain gauge pressuretransduction or a subdural catheterfluid-coupled device is similar to ven-tricular ICP. However, some investiga-tors have found that subdural and pa-renchymal fiberoptic catheter tippressure monitoring does not alwayscorrelate well with ventricular ICP.

● Fluid-coupled epidural devices or sub-arachnoid bolts and pneumatic epi-dural devices are less accurate thanventricular ICP monitors. Significantdifferences in readings have been dem-onstrated between catheter tip straingauge ICP devices that are placed inthe parenchyma vs. the subdural space.

C. Complications. The following para-graph is abstracted from the “Guidelinesfor the Management of [Adult] SevereTraumatic Brain Injury” (1).

The complication rate for ICP moni-toring is low. The most common compli-cations are infection and loss of wave-form. There are no pediatric reportsdocumenting the incidence of significantbrain injury, hemorrhage, or seizures as aresult of ICP monitoring. There are nopediatric data on the use of prophylacticantibiotics to prevent infectious compli-cations. In patients with ventriculostomycatheters who require continuous CSFdrainage, ICP cannot be measured simul-taneously. Although complications rarelyproduce long-term morbidity in patients,they can increase cost by requiring re-placement of the monitor, and they cangive inaccurate ICP readings. Each typeof pressure transduction system and in-tracranial location of the monitor has aprofile of potential complications. Cali-bration, monitoring for infection, andchecking fluid coupled devices for ob-



Gambardella et al.(2), 1993

Retrospective study of 49 patients that evaluatedcorrelation between intraparenchymal Caminoand ventriculostomy ICP catheters.

III Good correlation between ICP measurements with the Camino vs.ventriculostomy catheter (r � .73–.89), with differences rangingfrom 1 to 4 mm Hg.

Jensen et al. (3),1997

Prospective uncontrolled case series of 98 patientswith 12 ventriculostomy catheters and 37intraparenchymal Camino catheters placed.

III Infectious complication rate was 7% (all positive forStaphylococcus aureus). Loss of waveform occurred in 13% ofcatheters occurring at a mean of 9.5 days (range, 4–7 days).Data support low incidence of infection and mechanical failure.

ICP, intracranial pressure.


struction are necessary tasks in maintain-ing an optimal ICP monitoring system.

V. SUMMARY

In pediatric patients who require ICPmonitoring, a ventricular catheter and/oran external strain gauge transducer orcatheter tip pressure transducer device isan accurate and reliable method of mon-itoring ICP. A ventriculostomy catheterdevice also enables therapeutic CSFdrainage. Clinically significant infectionsassociated with ICP devices causing pa-tient morbidity are rare and should notdeter the decision to monitor ICP. Theincidence of other complications, such ashemorrhage or seizures, is unknown, butthe absence of reported incidents in the

pediatric literature suggests that the in-cidence is probably low.

Parenchymal catheter tip pressuretransducer devices measure ICP similarto ventricular ICP pressure but have thepotential for measurement differencesand drift due to the inability to recali-brate. These devices are advantageouswhen ventricular access is limited or un-available or if there is obstruction in thefluid couple. There are no credible data(class III or better) on the accuracy ofsubarachnoid or subdural-coupled de-vices, epidural ICP devices, or externallyplaced anterior fontanel devices.


Prospective clinical studies in pediat-ric patients of the accuracy and compli-cation rate of ventricular and intraparen-chymal ICP measuring devices need to beperformed. An industry or Food and DrugAdministration supported national pedi-atric registry should be established to col-lect information on this and other issuesin pediatric medicine.

The specification standard for pediatricICP monitoring should include in vivo clin-ical ICP drift measurement. In vitro testingdevices do not necessarily reflect clinicalperformance. Specifications for ICP devicesshould be reviewed in the context of whatdata are useful in the management of pa-tients who require ICP monitoring.

A study of simultaneous parenchymaland ventricular ICP measurements usingan accurate catheter tip transducer device

in children would be useful. We must an-swer the question: Does parenchymal mon-itoring in or near a contusion site provideICP data that improve intracranial pressuremanagement and outcome compared withother sites (including contralateral sites) ofICP monitoring in children?

Recommendations for the use of pro-phylactic antibiotics, surgical techniques,ICP data collection, monitoring for com-plications, and timing for removal of ICPmonitoring devices in children need to bedeveloped. Further improvement in ICPmonitoring technology should focus ondeveloping an ICP device that can provideventricular CSF drainage and parenchy-mal ICP measurement simultaneously.This would allow in situ recalibration andgive accurate ICP measurements in caseof fluid obstruction or when CSF is ac-tively drained. Noninvasive measure-ments of ICP need to be developed.

REFERENCES


2. Gambardella G, Zaccone C, Cardia E, et al:Intracranial pressure monitoring in children:Comparison of external ventricular devicewith the fiberoptic system. Childs Nerv Syst1993; 9:470–473

3. Jensen RL, Hahn YS, Ciro E: Risk factors ofintracranial pressure monitoring in childrenwith fiberoptic devices: A critical review. SurgNeurol 1997; 47:16–22

4. Brown E: Intracranial Pressure MonitoringDevices. Arlington, VA, Association for the Ad-vancement of Medical Instrumentation, 1988



Chapter 7. ICP Monitoring Technology



I n pediatric patients who

require intracranial pres-

sure monitoring, a ven-

tricular catheter and/or an ex-

ternal strain gauge transducer

or catheter tip pressure trans-

ducer device is an accurate and

reliable method of monitoring

intracranial pressure.


Chapter 8. Cerebral perfusion pressure

I. RECOMMENDATIONS

A. Standards. There are insufficientdata to support treatment standards forthis topic.

B. Guidelines. A cerebral perfusionpressure (CPP) �40 mm Hg in childrenwith traumatic brain injury (TBI) shouldbe maintained.

C. Options. A CPP between 40 and 65mm Hg probably represents an age-related continuum for the optimal treat-ment threshold. There may be exceptionsto this range in some infants and neo-nates.

Advanced cerebral physiologic moni-toring may be useful to define the opti-mal CPP in individual instances.

Hypotension should be avoided.D. Indications from the Adult Guide-

lines. The adult guidelines (1) statedthere were insufficient data to supporteither a treatment standard or guideline.Under Options, it stated, “Cerebral perfu-sion pressure (CPP) should be main-tained at a minimum of 70 mm Hg.” (1)

II. OVERVIEW

Global or regional cerebral ischemia isan important secondary insult to theacutely injured brain. Grossly, the CPP—defined as the mean arterial pressure mi-nus the intracranial pressure (ICP)—defines the pressure gradient drivingcerebral blood flow (CBF), which, in turn,is related to metabolic delivery of essen-tial substrates. The posttraumatic brainhas a significant incidence of vasospasmthat may increase the cerebral vascularresistance and decrease the CPP, produc-ing ischemia. With the use of continuousmonitoring capabilities including inva-sive blood pressure and ICP equipment,the CPP could be manipulated in an at-tempt to avoid both regional and globalischemia.

III. PROCESS

We searched Medline and Healthstarfrom 1966 to 2001 by using the search

strategy for this question (see AppendixA) and supplemented the results with lit-erature recommended by peers or identi-fied from reference lists. Of 53 potentiallyrelevant studies, five were used as evi-dence for this question (Table 1).


There is abundant evidence that CBFdeclines following TBI and may fre-quently reach the ischemic threshold forbrain tissue (2–6). Regional CBF may beeven more reduced in the vicinity of in-tracranial hematomas and contusions (7,8). There is much debate on how best tomeasure CBF and at what threshold thereis actual tissue ischemia. Cerebral perfu-sion pressure is relatively easy to measureand appears to correlate well with CBFwhen measured. A low CPP is highly cor-related with poor outcome, but there isless evidence that manipulating the CPPcan change eventual neurologic outcomein both adults and children.

There is little quality evidence for therole of CPP in pediatric patients. Thecomprehensive literature search for thisguideline only found one class II studyand no class I studies.

A retrospective cohort, class II studycollected data on all pediatric TBI pa-tients presenting to both level I pediatrictrauma centers in Oregon who receivedan ICP monitor (118 patients, GlasgowComa Scale [GCS] 6 � 3, age 7.4 � 4.6yrs). By logistic regression methods, theauthors found mortality rate significantlyassociated with a mean CPP �40 mm Hg(p � .01) and mean ICP �20 mm Hg (p� .001). Mean arterial pressure �70 mmHg (their definition of “hypotension”)was not statistically independently asso-ciated with death. They also found noincremental reduction of mortality rateor improved 3-month Glasgow OutcomeScale score associated with mean in-creases of CPP �40 mm Hg. However,only 60% of patients had documentedfollow-up at 3 months (9).

Barzilay et al. (10) studied 56 consec-utive admissions to their pediatric inten-

sive care unit (PICU) with coma fromTBI, central nervous system infections(five cases), and miscellaneous etiologies(ten cases) for at least 6 hrs before admis-sion. Mean arterial pressure and CPPwere treated in an uncontrolled fashion,and patients were followed until hospitaldischarge only. Comparison of survivorsand nonsurvivors showed results in Table2. The difference in minimum CPP issignificant at p � .001.

Elias-Jones et al. (11) studied 39 con-secutive PICU admissions for TBI with aninitial GCS ranging from 3 to 11 and anage range of 2 months to 13 yrs (averageage 7.8 yrs) who had multiple interven-tions for ICP �20 mm Hg or CPP �50mm Hg. They showed that all but onesurvivor (of total 30) had CPP �40 mmHg, and CPP was �40 mm Hg in seven ofnine fatalities (p � .00002, Fisher’s exacttest). The interventions included hypo-thermia to 32°C and hyperventilation to aPaCO2 of 3.0–3.5 kPa for all patients. Se-vere hypocapnia was associated with aworse outcome.

Another retrospective case series of 24consecutive admissions to a PICU of pa-tients with a GCS �8, average age 6.3 yrs,with ten patients between 1 and 5 yrs ofage, showed that all survivors had CPP �50mm Hg (p � .005, Fisher’s exact test) (12).

Sharples et al. (13, 14) studied a con-venience sample of 17 pediatric patients(2–16 yrs old, average age 7 yrs) with TBIand GCS 3–8 by measuring continuousmean arterial pressures and ICPs and cal-culating CBF and cerebral vascular resis-tance by using the nitrous oxide method.The authors found cerebral vascular re-sistance was proportional to CPP (Pear-son r � .32, p � .0003) and the relation-ship was even closer in those patientsdeemed to have a “good” outcome (i.e.,near-normal neurologic function).


A large randomized controlled trial of189 patients with severe TBI (15% with a


gunshot wound) compared ICP-targetedtherapy with CBF-targeted therapy thatresulted in a CPP �50 mm Hg in the firstgroup and a CPP �70 mm Hg in thesecond. It showed no difference in 3- or6-month Glasgow Outcome Scale score,but there was an increased risk of adultrespiratory distress syndrome in the sec-ond group (15). Another large prospec-tive study of 353 severe TBI patients, with178 receiving continuous monitoring andmanagement of cerebral oxygen extrac-tion in addition to management of CPP,showed that the additional interventionwas associated with better neurologicoutcome at 6 months (16). A retrospec-tive study by Changaris et al. (17) in 136adult patients showed that all patientswith a CPP �60 mm Hg on the secondday of injury died and more patients witha CPP �80 mm Hg had a better outcome.Another retrospective study of 221 pa-tients followed for a year after TBI alsoshowed better outcomes including mor-tality rate (p � .001) if CPP �80 mm Hg(18). Multiple prospective studies evalu-ating interventions such as hypothermia,barbiturate coma, and controlling jugu-

lar venous oxygen content used CPP �70mm Hg as a target without directly eval-uating this variable alone (19–23). Twostudies looked at the relationship of arti-ficially raising the mean arterial pressurein TBI patients with inotropes, and thesestudies found that this intervention hadlittle effect on the ICP, thereby raisingCPP and CBF. The studies were too smallto find an effect on ultimate outcome (24,25). Another small study (n � 21) showedthat an increase of CPP from an averageof 32 to 67 mm Hg significantly improvedbrain tissue PO2 but that further in-creases did not improve this variable (26).

V. SUMMARY

A CPP �40 mm Hg is consistentlyassociated with increased mortality, inde-pendent of age. It is unclear whether thisvalue represents a minimal threshold orwhether the optimal CPP may be abovethis in children (e.g., 50–65 mm Hg),based on available data. There are likelyto be age-related differences in optimalCPP goals that are indiscernible due tosmall numbers in these pediatric studies.No study demonstrates that active main-tenance of CPP above any target thresh-old in pediatric TBI is responsible forimproved mortality or morbidity.


Controlled, prospective, randomizedstudies in children are needed to deter-

mine optimal CPP levels in various pedi-atric age groups and mechanisms of in-jury. The relative importance of ICP andCPP-targeted therapies needs to be as-sessed by using age-appropriate long-term (�1 yr) functional outcomes.

REFERENCES


2. Bouma GJ, Muizelaar JP, Choi SC, et al:Cerebral circulation and metabolism after se-vere traumatic brain injury: The elusive roleof ischemia. J Neurosurg 1991; 75:685–693

3. Bouma GJ, Muizelaar JP, Stringer WA, et al:Ultra-early evaluation of regional cerebralblood flow in severely head-injured patientsusing xenon-enhanced computerized tomog-raphy. J Neurosurg 1992; 77:360–368

4. Jaggi JL, Obrist WD, Gennarelli TA, et al:Relationship of early cerebral blood flow andmetabolism to outcome in acute head injury.J Neurosurg 1990; 72:176–182

5. Marion DW, Darby J, Yonas H: Acute regionalcerebral blood flow changes caused by severehead injuries. J Neurosurg 1991; 74:407–414

6. Robertson CS, Contant CF, Narayan RK, etal: Cerebral blood flow, AVDO2, and neuro-logic outcome in head-injured patients.J Neurotrauma 1992; 9(Suppl 1):S349–S358

7. Salvant JB Jr, Muizelaar JP: Changes in ce-rebral blood flow and metabolism related tothe presence of subdural hematoma. Neuro-surgery 1993; 33:387–393

8. McLaughlin MR, Marion DW: Cerebral bloodflow and vasoresponsivity within and aroundcerebral contusions. J Neurosurg 1996; 85:871–876



Barzilay et al. (10),1988

Retrospective analysis of survival at hospital discharge withminimum CPP in 56 patients (41 with severe TBI, fivewith CNS infections, and ten miscellaneous).

III CPP 65.5 � 8.5 mm Hg for survivors vs. 6.0� 3.9 mm Hg for nonsurvivors (p �.001).

Elias-Jones et al.(11), 1992

Retrospective analysis of neurologic outcome at 2.5 yrs (0.5–5 yrs) post-TBI (GCS 5.5, 3–11) associated with extremesof CPP in 39 patients (average age 7.8 yrs, 2 mos–13 yrs).

III CPP �40 mm Hg in all but one survivor,�40 mm Hg in seven of nine fatalities (p�.00002).

Kaiser andPfenninger (12),1984

Retrospective analysis of 24 patients (average age 6.3 yrs)with severe TBI (all GCS �8), 21.5% with ICH, GOSfollow-up 2.5 yrs (1.5–4.4 yrs).

III All survivors with minimum CPP �50 mmHg (p � .005).

Sharples et al.(13), 1995

Retrospective analysis of 17 patients (convenience sample),aged 2–16 yrs (average 7 yrs), neurologic outcome (“goodand moderate” vs. “severe and died”). CPP, CBF, andCMRO2 calculated.

III CVR proportional to CPP (r � .32, p �.0003). CVR not related to age, GCS, ortime from injury.

Downard et al. (9),2000

Retrospective cohort study of 118 patients (age 7.4 � 4.6yrs) with ICP monitors established within 24 hrs ofadmission, GCS 6 � 3, 50% with space occupying lesions,hourly CPP calculations. GOS last recorded in chart.

II No survivors with mean CPP �40 mm Hg (p�.01); no relationship for increments ofCPP �40 mm Hg.

CPP, cerebral perfusion pressure; TBI, traumatic brain injury; CNS, central nervous system; ICH, intracranial hemorrhage; GOS, Glasgow OutcomeScale; CBF, cerebral blood flow; CMRO2, cerebral metabolic rate of oxygen; CVR, cerebral vascular resistance; GCS, Glasgow Coma Scale.

Table 2. Comparison of cerebral perfusion pres-sure (CPP) and intracranial pressure (ICP) insurvivors vs. nonsurvivors (10)

Minimum CPP,mm Hg

Maximum ICP,mm Hg

Survivors 65.5 � 8.5 16.9 � 3.1Nonsurvivors 6.0 � 3.9 53.7 � 10.8


9. Downard C, Hulka F, Mullins RJ, et al: Rela-tionship of cerebral perfusion pressure andsurvival in pediatric brain-injured patients.J Trauma 2000; 49:654–658

10. Barzilay Z, Augarten A, Sagy M, et al: Vari-ables affecting outcome from severe braininjury in children. Intensive Care Med 1988;14:417–421

11. Elias-Jones AC, Punt JA, Turnbull AE, et al:Management and outcome of severe headinjuries in the Trent region 1985–90. ArchDis Child 1992; 67:1430–1435

12. Kaiser G, Pfenninger J: Effect of neurointen-sive care upon outcome following severe

head injuries in childhood—A preliminaryreport. Neuropediatrics 1984; 15:68–75

13. Sharples PM, Matthews DS, Eyre JA: Cerebralblood flow and metabolism in children withsevere head injuries. Part 2: Cerebrovascularresistance and its determinants. J NeurolNeurosurg Psychiatry 1995; 58:153–159

14. Sharples PM, Stuart AG, Matthews DS, et al:Cerebral blood flow and metabolism in chil-dren with severe head injury. Part 1: Relationto age, Glasgow coma score, outcome, intra-cranial pressure, and time after injury.J Neurol Neurosurg Psychiatry 1995; 58:145–152

15. Robertson CS, Valadka AB, Hannay HJ, et al:Prevention of secondary ischemic insults af-ter severe head injury. Crit Care Med 1999;27:2086–2095

16. Cruz J: The first decade of continuous mon-itoring of jugular bulb oxyhemoglobin satu-ration: Management strategies and clinicaloutcome. Crit Care Med 1998; 26:344–351

17. Changaris DG, McGraw CP, Richardson JD,et al: Correlation of cerebral perfusion pres-sure and Glasgow Coma Scale to outcome.J Trauma 1987; 27:1007–1013

18. McGraw CP: A cerebral perfusion pressuregreater than 80 mmHg is more beneficial. In:Intracranial Pressure VII. Hoff JT, Betz AL(Eds). Berlin, Springer-Verlag, 1989, pp839–841

19. Clifton GL, Allen S, Barrodale P, et al: Aphase II study of moderate hypothermia in

severe brain injury. J Neurotrauma 1993;10:263–271

20. Fortune JB, Feustel PJ, Weigle CG, et al:Continuous measurement of jugular ve-nous oxygen saturation in response to tran-sient elevations of blood pressure in head-injured patients. J Neurosurg 1994; 80:461–468

21. Marion DW, Obrist WD, Carlier PM, et al: Theuse of moderate therapeutic hypothermia forpatients with severe head injuries: A prelim-inary report. J Neurosurg 1993; 79:354–362

22. Rosner MJ, Daughton S: Cerebral perfusionpressure management in head injury.J Trauma 1990; 30:933–940

23. Yoshida A, Shima T, Okada Y, et al: Outcomeof patients with severe head injury evaluationby cerebral perfusion pressure. In: RecentAdvances in Neurotrauma. Nakamura N,Hashimoto T, Yasue M (Eds). Hong Kong,Springer-Verlag, 1993, pp 309–312

24. Bouma GJ, Muizelaar JP: Relationship be-tween cardiac output and cerebral blood flowin patients with intact and with impairedautoregulation. J Neurosurg 1990; 73:368–374

25. Bruce DA, Langfitt TW, Miller JD, et al: Re-gional cerebral blood flow, intracranial pres-sure, and brain metabolism in comatose pa-tients. J Neurosurg 1973; 38:131–144

26. Kiening KL, Hartl R, Unterberg AW, et al:Brain tissue pO2-monitoring in comatose pa-tients: Implications for therapy. Neurol Res1997; 19:233–240



Chapter 8. Cerebral Perfusion Pressure

1. exp craniocerebral trauma/2. head injur$.tw.3. brain injur$.tw.4. 1 or 2 or 35. cerebral perfusion pressure.tw.6. cerebrovascular circulation/ and blood pressure/7. 5 or 68. 4 and 79. limit 8 to (newborn infant �birth to 1 month� or infant �1 to 23 months� or preschool child �2 to 5 years� or

child �6 to 12 years� or adolescence �13 to 18 years�)10. limit 9 to english language

C ontrolled, pro-

spective, random-

ized studies in

children are needed to deter-

mine optimal cerebral perfu-

sion pressure levels in vari-

ous pediatric age groups and

mechanisms of injury.


Chapter 9. Use of sedation and neuromuscular blockade in thetreatment of severe pediatric traumatic brain injury

I. RECOMMENDATIONS

A. Standards. There are insufficientdata to support a treatment standard forthis topic.*

B. Guidelines. There are insufficientdata to support a treatment guideline forthis topic.*

C. Options. In the absence of outcomedata, the choice and dosing of sedatives,analgesics, and neuromuscular blockingagents used in the management of infantsand children with severe traumatic braininjury (TBI) should be left to the treatingphysician. However, the effect of individ-ual sedatives and analgesics on intracra-nial pressure (ICP) in infants and chil-dren with severe TBI can be variable andunpredictable.

D. Indications from Adult Guidelines.The guidelines on the management ofadults with severe TBI (1) did not includea specific chapter on the use of sedation,analgesia, or neuromuscular blockade.

In the chapter on initial management(2), it was stated that neuromuscularblocking agents can facilitate mechanicalventilation and management of raisedICP, but their use should be reserved forspecific indications. The depth and dura-tion of neuromuscular blockade shouldbe monitored and optimized, respec-tively.

II. OVERVIEW

Sedatives, analgesics, and neuromus-cular blocking agents are commonly usedin the management of infants and chil-dren with severe TBI. Use of these agentscan be divided into two major categories:a) for emergency intubation, and b) formanagement including control of ICP inthe intensive care unit (ICU). The use of

sedatives, analgesics, and neuromuscularblocking agents for emergency intuba-tion is addressed in chapter 3. This sec-tion evaluates use of sedation, analgesia,and neuromuscular blockade during ICUtreatment.

Despite their common use in the man-agement of severe TBI in infants and chil-dren, sedatives, analgesics, and neuro-muscular blocking agents have beensubjected to very limited clinical investi-gation. Most of the medical literature onthese agents in pediatric TBI consists ofeither descriptions of small numbers ofchildren included in adult studies (butnot fully described) or case reports—often describing an unanticipated re-sponse to administration of a given agent.The lack of high-quality pediatric studiesseverely limits any conclusions that canbe made.

III. PROCESS

We searched Medline and Healthstarfrom 1966 to 2001 by using the searchstrategy for this question (see AppendixA) and supplemented the results with lit-erature recommended by peers or identi-fied from reference lists. Of 40 potentiallyrelevant studies, one was used as evi-dence for this question (Table 1).


A. Sedation and Analgesia. The rec-ommendations on the use of sedatives,analgesics, and neuromuscular blockingagents in this chapter are for patientswith a secure airway who are receivingmechanical ventilatory support yieldingthe desired arterial blood gas values. Sed-atives and analgesics are believed to fa-vorably treat a number of importantpathophysiologic derangements in severeTBI. They can facilitate necessary generalaspects of patient care such as the abilityto maintain the airway, vascular cathe-ters, and other monitors. Sedatives andanalgesics also can facilitate patient

transport for diagnostic procedures. Sed-atives and analgesics also are believed tobe useful by mitigating aspects of second-ary damage. Pain and stress markedly in-crease cerebral metabolic demands andcan pathologically increase cerebralblood volume and raise ICP.

Studies in experimental modelsshowed that a two- to three-fold increasein cerebral metabolic rate for oxygen ac-companies painful or stressful stimuli (3,4). Noxious stimuli such as suctioningalso can increase ICP (5–8). Painful andnoxious stimuli and stress also can con-tribute to increases in sympathetic tone,with hypertension, and bleeding from op-erative sites (9). However, sedative-induced reductions in arterial blood pres-sure can lead to cerebral vasodilation andexacerbate increases in cerebral bloodvolume and ICP. In the absence of ad-vanced monitoring, care must be taken toavoid this complication.

Sedatives and analgesics are used totreat painful and noxious stimuli. Theyalso facilitate mechanical ventilatory sup-port. Other proposed benefits of sedativesafter severe TBI include anticonvulsantand anti-emetic actions, the preventionof shivering, and mitigation of the long-term psychological trauma of pain andstress. Prielipp and Coursin (10) de-scribed the ideal sedative for patientswith severe TBI as one that is rapid inonset and offset, is easily titrated to effect,has well-defined metabolism (preferablyindependent of end-organ function), nei-ther accumulates nor has active metabo-lites, exhibits anticonvulsant actions, hasno adverse cardiovascular or immune ac-tions, and lacks drug-drug interactions,while preserving the neurologic examina-tion.

Eight studies were identified that ad-dressed the use of sedatives and/or anal-gesics in severe pediatric TBI. However,none of these reports reached the level ofclass III data. All either were studies inadults that included a small unstratified

*As stated by the Food and Drug Administration,continuous infusion of propofol for either sedation orthe management of refractory intracranial hyperten-sion in infants and children with severe traumatic braininjury is not recommended.


number of children or were case reports.The sedatives and analgesics in thesestudies included narcotics, benzodiaz-epines, ketamine, and propofol.

Tobias (11) reported that bolus fenta-nyl (5 �g/kg body weight) produced aspike in ICP in an 11-yr-old child withsevere TBI. ICP responded to barbiturateand mannitol administration. Remark-ably, this is the only identified report oneither fentanyl or morphine use in themanagement of ICP in pediatric TBI. Al-banese et al. (12) studied the effect ofsufentanil (1 �g/kg intravenous bolusplus infusion) on ICP in ten comatosepatients with severe TBI, including threeadolescents. Sufentanil increased ICP 9� 7 mm Hg and decreased cerebral per-fusion pressure (CPP) 38% after admin-istration. Infusion of the ultra-short-acting narcotic remifentanil controlledrefractory ICP in a 16-yr-old child withsevere TBI, when hypotension limitedpropofol use (13).

Cotev and Shalit (14) studied the ef-fect of diazepam in eight patients withsevere TBI, including one adolescent. An�25% reduction in cerebral metabolicrate for oxygen and cerebral blood flowwas seen without an effect on blood pres-sure. Studies of other commonly usedbenzodiazepines (midazolam, lorazepam)in pediatric TBI are lacking. Albanese etal. (15) studied the effect of ketamine(1.5, 3, and 5 mg/kg intravenous boluses)on ICP and electroencephalogram ineight patients (including three teenagers)with severe TBI. Surprisingly, bolus dosesof ketamine, a sedative agent that hasbeen contraindicated for use in the set-ting of increased ICP, was associated witha 2–5 mm Hg reduction in ICP.

Spitzfaden et al. (16) reported success-ful treatment of refractory intracranialhypertension in a 7-yr-old with TBI usingcontinuous infusion of propofol (3–5mg·kg�1·hr�1 for 4 days). Similarly,Farling et al. (17) reported a study on theeffect of propofol (intravenous infusion of

1.04–4.97 mg·kg�1·hr�1) in ten coma-tose patients (including two teenagers)with severe TBI. Propofol infusion (for 24hrs) produced adequate sedation and nomajor changes in ICP or CPP. However, anumber of reports (in cases not restrictedto TBI) suggest that administration ofpropofol by continuous infusion is asso-ciated with an unexplained increase inmortality risk. A syndrome of lethal met-abolic acidosis can occur (18 –22).“Propofol syndrome” also has been re-ported in an adult with severe TBI (23).In light of these risks, and with alterna-tive therapies available, continuous infu-sion of propofol for either sedation ormanagement of refractory intracranialhypertension in severe pediatric TBI isnot recommended. The Center for DrugEvaluation and Research (24) of the FDAstates, “Propofol is not indicated for pe-diatric ICU sedation as safety has notbeen established.”

Although there is one report of seda-tion with infusion of etomidate in TBIthat included children (25), lack of agestratification made it impossible to defineits effect in the pediatric subgroup. Noarticles were located that evaluated theuse of lidocaine to blunt the response toairway stimulation in children with se-vere TBI. Finally, barbiturates can begiven as sedatives by using doses lowerthan those required to induce barbituratecoma. The use of high-dose barbituratesin the management of infants and chil-dren with severe TBI will be addressed inChapter 13.

B. Neuromuscular Blockade. Neuro-muscular blocking agents have been sug-gested to reduce ICP by a variety of mech-anisms including a reduction in airwayand intrathoracic pressure with facilita-tion of cerebral venous outflow and byprevention of shivering, posturing, orbreathing against the ventilator (26). Re-duction in metabolic demands by elimi-nation of skeletal muscle contraction also

has been suggested to represent a bene-ficial effect of neuromuscular blockade.

Risks of neuromuscular blockade in-clude the potential devastating effect ofhypoxemia secondary to inadvertent ex-tubation, risks of masking seizures, in-creased incidence of nosocomial pneu-monia (shown in adults with severe TBI)(26), cardiovascular side effects, immobi-lization stress (if neuromuscular block-ade is used without adequate sedationand analgesia), and increased ICU lengthof stay (26, 27). Myopathy is most com-monly seen with the combined use ofnondepolarizing agents and corticoste-roids. Incidence of this complication var-ies greatly between studies and rangesbetween 1% and �30% of cases (28–30).Monitoring of the depth of neuromuscu-lar blockade can shorten duration of neu-romuscular blockade in the ICU (31).

Two pediatric studies of neuromuscu-lar blocking agents, which were not re-stricted to children with TBI, suggestthat these agents are more commonlyused in the management of critically illinfants and children than in adults—asmuch as five times more common (28,32). However, only two studies were iden-tified that addressed the use of neuro-muscular blocking agents in the settingof severe pediatric TBI (33, 34). One ofthese reports reached the level of class IIdata for the effect of neuromuscularblocking agents on systemic oxygen con-sumption. Vernon et al. (33) performed aprospective, unblinded crossover study ofthe effect of neuromuscular blockadewith vecuronium or pancuronium on to-tal body oxygen consumption in 20 me-chanically ventilated children, six ofwhom had severe TBI. Neuromuscularblockade reduced oxygen consumptionand energy expenditure 8.7 � 1.7% and10.3 � 1.8%, respectively. The authorsconcluded that although neuromuscularblockade reduces oxygen consumption,the degree of reduction is small. No studyof the efficacy of specific therapeutic ap-



Vernon and Witte(33), 2000

Prospective, unblinded crossover study of the effect ofneuromuscular blockade on oxygen consumption in20 mechanically ventilated children, six of whomhad severe TBI.

IIa Neuromuscular blockade reduced oxygen consumptionand energy expenditure 8.7 � 1.7% and 10.3 �1.8%, respectively. Although the effect wassignificant, the magnitude was modest.

TBI, traumatic brain injury.aClass II evidence only for the effect of neuromuscular blockade on oxygen consumption—not on long-term outcome.


proaches to neuromuscular blockade inthe treatment of pediatric TBI was iden-tified. Finally, in a study of eight patients(including two adolescents), continuousinfusion of doxacurium provided stableneuromuscular blockade without alteringICP or CPP and was less expensive thanother commonly used agents (34).


In the chapter on initial management(2) in the adult guidelines, it was statedthat approaches to sedation and neuro-muscular blockade vary widely. Therehave been no studies on the influence ofsedation on outcome from severe TBI;therefore, the decisions on the use ofsedation and the choice of sedative agentswere left up to the treating physician.Adult guidelines also were not written forthe use of neuromuscular blockingagents. However, the initial managementsection cited one class II study by Hsianget al. (26) that examined 514 entries inthe Traumatic Coma Data Bank and re-ported an increased incidence of nosoco-mial pneumonia and prolonged ICU stayassociated with early prophylactic use ofneuromuscular blockade. It was sug-gested that use of neuromuscular block-ing agents be reserved for specific indica-tions (intracranial hypertension,transport).

V. SUMMARY

There were no studies with sedativesor analgesics providing acceptable evi-dence for the present report. There wasonly one study of the use of neuromus-cular blockade that qualified as class II,

and that involved the effect of neuromus-cular blockade on oxygen consumptiononly. Until experimental comparisonsamong specific regimens of these seda-tive, analgesic, and neuromuscular block-ing agents are carried out, the choice anddosing of sedatives and analgesic agentsused in the management of infants andchildren with severe TBI should be left tothe treating physician.

Based on recommendations of theFDA, continuous infusion of propofol isnot recommended in the treatment ofpediatric TBI.


Additional study is needed comparingthe various sedatives and analgesics inpediatric patients with severe TBI. Assess-ments are needed of optimal agents, dos-ing, duration, and interaction effects withother concurrent therapies. Study of theeffect of various sedation strategies onthe development and therapeutic inten-sity level of intracranial hypertension alsois needed. Although multiple-center tri-als assessing the effect of these agents onoutcome would be optimal, based on thecurrent dearth of investigation on the useof sedatives and analgesics in pediatricTBI, even case series or small cohortstudies would advance the literature.Similarly lacking are studies addressingthe important issue of age-related differ-ences and the unique subgroup of infantswho are victims of abusive head trauma.The issue of age-related differences maybe of particular importance in the area ofsedation, since studies in experimentalanimal models of TBI suggest that somelevel of synaptic activation is essential tonormal development in infancy and thatanti-excitotoxic agents may trigger apo-ptosis in the injured brain (35, 36). Thus,optimal sedation after severe TBI maydiffer between infants and older childrenand deserves specific investigation. Fi-nally, the specific role of neuromuscularblocking agents in infants and childrenwith severe TBI also remains to be stud-ied.

REFERENCES


2. The Brain Trauma Foundation, The Ameri-can Association of Neurological Surgeonsand The Joint Section on Neurotrauma and

Critical Care. Initial management. J Neuro-trauma 2000; 17:463–469

3. Nilsson B, Rehncrona S, Siesjo BK: Couplingof cerebral metabolism and blood flow inepileptic seizures, hypoxia and hypoglyce-mia. In: Cerebral Vascular Smooth Muscleand Its Control. Purves M (Ed). Amsterdam,Excerpta Medica, Elsevier, 1978, pp 199–218

4. Rehncrona S, Siesjo BK: Metabolic and phys-iologic changes in acute brain failure. In:Brain Failure and Resuscitation. Grenvik A,Safar P (Eds). New York, Churchill Living-stone, 1981, pp 11–33

5. White PF, Schlobohm RM, Pitts LH, et al: Arandomized study of drugs for preventingincreases in intracranial pressure during en-dotracheal suctioning. Anesthesiology 1982;57:242–244

6. Raju TNK, Vidyasagar D, Torres C, et al:Intracranial pressure during intubation andanesthesia in infants. J Pediatr 1980; 96:860–862

7. Kerr ME, Weber BB, Sereika SM, et al: Effectof endotracheal suctioning on cerebral oxy-genation in traumatic brain-injured patients.Crit Care Med 1999; 27:2776–2781

8. Fortune JB, Feustel PJ, Weigle C, et al: Con-tinuous measurement of jugular venous ox-ygen saturation in response to transient ele-vations of blood pressure in head-injuredpatients. J Neurosurg 1994; 80:461–468

9. Todres ID: Post anesthesia recovery in thepediatric intensive care unit. In: PediatricCritical Care. Fuhrman BP, Zimmerman JJ(Eds). St. Louis, Mosby, 1998, pp 1391–1398

10. Prielipp RC, Coursin DB: Sedative and neu-romuscular blocking drug use in critically illpatients with head injuries. New Horiz 1995;3:456–468

11. Tobias JD: Increased intracranial pressureafter fentanyl administration in a child withclosed head trauma. Pediatr Emerg Care1994; 10:89–90

12. Albanese J, Durbec O, Viviand X, et al: Sufen-tanil increases intracranial pressure in pa-tients with head trauma. Anesthesiology1993; 79:493–497

13. Tipps LB, Coplin WM, Murry KR, et al: Safetyand feasibility of continuous infusion ofremifentanil in the neurosurgical intensivecare unit. Neurosurgery 2000; 46:596–601

14. Cotev S, Shalit MN: Effects of diazepam oncerebral blood flow and oxygen uptake afterhead injury. Anesthesiology 1975; 43:117–122

15. Albanese J, Arnaud S, Rey M, et al: Ketaminedecreases intracranial pressure and electro-encephalographic activity in traumatic braininjury patients during propofol sedation. An-esthesiology 1997; 87:1328–1334

16. Spitzfaden AC, Jimenez DF, Tobias JD:Propofol for sedation and control of intracra-nial pressure in children. Pediatr Neurosurg1999; 31:194–200

17. Farling PA, Johnston JR, Coppel DL: Propo-fol infusion for sedation of patients with headinjury in intensive care. Anaesthesia 1989;44:222–226

B ased on recom-

mendations of the

Food and Drug

Administration, continuous

infusion of propofol is not

recommended in the treat-

ment of pediatric traumatic

brain injury.


18. Bray RJ: Propofol infusion syndrome in chil-dren. Paediatr Anaesth 1998; 8:491–499

19. Hanna JP, Ramundo ML: Rhabdomyolysisand hypoxia associated with prolongedpropofol infusion in children. Neurology1998; 50:301–303

20. Cray SH, Robinson BH, Cox PN: Lactic aci-demia and bradyarrhythmia in a child se-dated with propofol. Crit Care Med 1998;26:2087–2092

21. Parke TJ, Stevens JE, Rice AS, et al: Meta-bolic acidosis and fatal myocardial failureafter propofol infusion in children: Five casereports. BMJ 1992; 305:613–616

22. Canivet JL, Gustad K, Leclercq P, et al: Mas-sive ketonuria during sedation with propofolin a 12 year old girl with severe head trauma.Acta Anaesthesiol Belg 1994; 45:19–22

23. Cremer OL, Moons KGM, Bouman EAC, et al:Long-term propofol infusion and cardiac fail-ure in adult head-injured patients. Lancet2001; 357:117–118

24. Center for Drug Evaluation and Research.Available at: http://www.fda.gov/cder/pediatric/labelchange.htm, accessed on May 5, 2003

25. Prior JGL, Hinds CJ, Williams J, et al: Theuse of etomidate in the management of se-

vere head injury. Intensive Care Med 1983;9:313–320

26. Hsiang JK, Chesnut RM, Crisp CB, et al:Early, routine paralysis for intracranial pres-sure control in severe head injury: Is it nec-essary? Crit Care Med 1994; 22:1471–1476

27. Durbin CG: Neuromuscular blocking agentsand sedative drugs. Clinical uses and toxiceffects in the critical care unit. Crit Care Clin1981; 7:480–506

28. Martin LD, Bratton SL, Quint P, et al: Pro-spective documentation of sedative, analge-sic, and neuromuscular blocking agent usein infants and children in the intensive careunit: A multicenter perspective. Pediatr CritCare Med 2001; 2:205–210

29. Douglass JA, Tuxen DV, Horne M, et al: My-opathy in severe asthma. Am Rev Respir Dis1992; 146:517–519

30. Rudis MI, Guslits BJ, Peterson EL, et al:Economic impact of prolonged motor weak-ness complicating neuromuscular blockadein the intensive care unit. Crit Care Med1996; 24:1749–1756

31. Rudis MI, Sikora CA, Angus E, et al: A pro-spective, randomized, controlled evaluationof peripheral nerve stimulation versus stan-

dard clinical dosing of neuromuscular block-ing agents in critically ill patients. Crit CareMed 1997; 25:575–583

32. Murray MJ, Strickland RA, Weiler C: Theuse of neuromuscular blocking drugs inthe intensive care unit: A US perspective.Intensive Care Med 1993; 19(Suppl 2):S40 –S44

33. Vernon DD, Witte MK: Effect of neuromus-cular blockade on oxygen consumption andenergy expenditure in sedated, mechanicallyventilated children. Crit Care Med 2000; 28:1569–1571

34. Prielipp RC, Robinson JC, Wilson JA, et al:Dose response, recovery, and cost of doxa-curium as a continuous infusion in neuro-surgical intensive care unit patients. CritCare Med 1997; 25:1236–1241

35. Ikonomidou C, Bittigau P, Koch C, et al:Neurotransmitters and apoptosis in the de-veloping brain (1). Biochem Pharmacol 200;62:401–405

36. Bittigau P, Sifringer M, Pohl D, et al: Apo-ptotic neurodegeneration following traumais markedly enhanced in the immature brain.Ann Neurol 1999; 45:724–735



Chapter 9. Sedation and Neuromuscular Blockade

1. exp craniocerebral trauma/2. head injur$.tw.3. brain injur$.tw.4. 1 or 2 or 35. brain ischemia/ or “cerebral ischemia”.mp.6. 4 and 57. limit 6 to (newborn infant �birth to 1 month� or infant �1 to 23 months� or preschool child �2 to 5 years� or



Chapter 10. The role of cerebrospinal fluid drainage in thetreatment of severe pediatric traumatic brain injury

I. RECOMMENDATIONS



C. Options. Cerebrospinal fluid (CSF)drainage can be considered as an optionin the management of elevated intracra-nial pressure (ICP) in children with se-vere closed head injury.

Drainage can be accomplished via aventriculostomy catheter alone or incombination with a lumbar drain. Theaddition of lumbar drainage should beconsidered as an option only in the caseof refractory intracranial hypertensionwith a functioning ventriculostomy, openbasal cisterns, and no evidence of a majormass lesion or shift on imaging studies.

II. OVERVIEW

In children with severe traumatic braininjury (TBI) and intracranial hypertension,ventricular CSF drainage is a commonlyemployed therapeutic modality in conjunc-tion with ICP monitoring. The role of CSFdrainage is to reduce intracranial fluid vol-ume and thereby lower ICP. The scientificliterature pertaining to CSF drainage intrauma, and in pediatric trauma in partic-ular, was reviewed.

III. PROCESS

We searched Medline and Healthstarfrom 1966 to 2001 by using the searchstrategy for this question (see AppendixA) and supplemented the results with lit-erature recommended by peers or identi-fied from reference lists. Of 68 potentiallyrelevant studies, three were used as evi-dence for this question (Table 1).


With the use of the ventriculostomy asa common means of measuring ICP of

patients with TBI (see Chapter 7), thepotential therapeutic benefits of CSFdrainage became of interest. Before theuse of the ventriculostomy in TBI, theprincipal use of CSF drainage was in pa-tients with hydrocephalus, but the abilityof this procedure to affect ICP led to itsincreased use as a therapeutic device.

We found one class III study in childrenevaluating the use of ventricular drainagein TBI. Shapiro and Marmarou (1) retro-spectively studied 22 children with severeTBI—defined as a score �8 on the Glas-gow Coma Scale (GCS), all of whom weretreated with ventricular drainage. Variablesmeasured included ICP, pressure-volumeindex, and mortality rate. In addition to thefinding that draining CSF increased pres-sure-volume index and decreased ICP, onlytwo neurologic deaths occurred in patientswith refractory intracranial hypertension.

Drainage of CSF is not limited to theventricular route. In response to observa-tions that the ventricles are often small inTBI and that up to 30% of the total com-pliance of the CSF system is in the spinalaxis, a series of articles have addressedthe feasibility of using lumbar drains inaddition to ventricular drainage. Baldwinand Rekate (2) reported on a series of fivechildren with severe TBI, in whom lum-bar drains were placed after failure tocontrol ICP with both ventricular drain-age and barbiturate coma. Three childrenhad quick and lasting resolution of raisedICP, two of them with good outcome andone with moderate remaining disability.In the other two cases, there was no effecton ICP and both children died.

Levy et al. (3) reported on the effect ofcontrolled lumbar drainage, with simul-taneous ventricular drainage, on out-come in 16 pediatric patients with severeTBI. In two patients ICP was unaffected,and both died. The remaining 14 sur-vived, eight having good outcome, threehaving moderate disability, and three hav-ing severe disability. Although there was nodirect outcome study on the use of barbi-

turates in this series, the authors proposedthat barbiturate coma, and its associatedmorbidity, could be avoided by the use oflumbar drainage. This statement was basedon the fact that in this series not all patientswere given barbiturates (five of 16 receivingno barbiturates and six of 16 receiving onlyintermittent dosing).


Following earlier reports of an effect onICP by drainage of CSF (4), Ghajar et al. (5)performed a prospective study, withoutrandomization, of the effect of CSF drain-age in adults with TBI. Treatment was se-lected by the admitting neurosurgeon and,after evacuation of mass lesions, patientsreceived either ventriculostomies withdrainage if ICP exceeded 15 mm Hg alongwith medical management (group 1) ormedical management only (group 2). Themedical management consisted of mild hy-perventilation to PCO2 � 35 mm Hg, headof bed elevation, normovolemia, and man-nitol (although only on admission). Pa-tients in group 2 had no ICP monitor of anykind. The outcome measurements weremortality rate and degree of disability. Mor-tality ate was 12% in group 1 vs. 53% ingroup 2. Of the patients in group 1, 59%were living independently at follow-up vs.20% of group 2.

Fortune et al. (6) studied the effects ofhyperventilation, mannitol, and CSFdrainage on cerebral blood flow (CBF) inTBI. Twenty-two patients were studied,with a mean age of 24 yrs (range, 14–48).Children were not reported separately.Although patient outcome was not re-ported, this study established that CSFdrainage, hyperventilation, and intermit-tent mannitol were all effective in reduc-ing ICP. The authors also found thatmannitol use increased CBF, CSF drain-age had negligible impact on CBF, andhyperventilation decreased CBF.


V. SUMMARY

Ventricular CSF drainage in severe pe-diatric TBI is supported as a treatmentoption in the setting of refractory intra-cranial hypertension; the addition oflumbar drainage in patients showingopen cisterns on imaging and withoutmajor mass lesions or shift also is sup-ported as a treatment option.


Future studies in this area should in-clude the following:

Prospective data collection on the out-come benefits of CSF drainage.

Studies to compare CSF drainage withother therapeutic modalities used inTBI management, such as osmolartherapy or barbiturates.

Work on technical aspects of drain us-age, such as continuous vs. intermit-tent drainage, age-specific use, and userelated to mechanism of injury.

Comparison of lumbar drainage withother second-tier therapies, such as de-compressive craniotomy.

Study of the potential role of subgalealdrainage in infants.

REFERENCES

1. Shapiro K, Marmarou A: Clinical applications ofthe pressure-volume index in treatment of pediat-ric head injuries. J Neurosurg 1982; 56:819–825

2. Baldwin HZ, Rekate HL: Preliminary experi-ence with controlled external lumbar drainagein diffuse pediatric head injury. Pediatr Neu-rosurg 1991–2; 17:115–120

3. Levy DI, Rekate HL, Cherny WB, et al: Con-trolled lumbar drainage in pediatric head in-jury. J Neurosurg 1995; 83:452–460

4. Bullock R, Chesnut RM, Clifton G, et al: Guide-lines for the management of severe traumaticbrain injury. J Neurotrauma 2000; 17:451–553

5. Ghajar JBG, Hariri RJ, Patterson RH: Im-proved outcome from traumatic coma usingonly ventricular cerebrospinal fluid drainagefor intracranial pressure control. Adv Neuro-surg 1993; 21:173–177

6. Fortune JB, Feustel PJ, Graca L, et al: Effect ofhyperventilation, mannitol, ventriculostomydrainage on cerebral blood flow after headinjury. J Trauma 1995; 39:1091–1099



Chapter 10. CSF Drainage

1. exp craniocerebral trauma/2. head injur$.tw.3. brain injur$.tw.4. 1 or 2 or 35. lumbar drain$.mp.6. lumbar shunt$.mp.7. exp cerebrospinal fluid shunts/8. *drainage/9. 5 or 6 or 7 or 8

10. 4 and 911. limit 10 to (newborn infant �birth to 1 month� or infant �1 to 23 months� or preschool child �2 to 5 years� or



References Description of StudyDataClass Conclusions

Shapiro and Marmaron(1), 1982

Retrospective series, 22patients with EVD, ICP/PVImeasured.

III Drainage increased PVI,decreased ICP, deaths only inpatients with uncontrolled ICP.

Baldwin and Rekate(2), 1991–1992

Clinical series, five patientswith lumbar drain.

III Three of five survived afterlowering ICP.

Levy et al. (3), 1995 Retrospective study, 16patients with lumbar drain.

III ICP lowered in 14/16, deaths intwo patients with uncontrolledICP.

EVD, external ventricular drain; ICP, intracranial pressure; PVI, pressure-volume index.

V entricular cerebro-

spinal fluid drain-

age in severe pedi-

atric traumatic brain injury is

supported as a treatment op-

tion in the setting of refrac-

tory intracranial hyperten-

sion; the addition of lumbar

drainage in patients showing

open cisterns on imaging and

without major mass lesions or

shift also is supported as a

treatment option.


Chapter 11. Use of hyperosmolar therapy in the management ofsevere pediatric traumatic brain injury

I. RECOMMENDATIONS



C. Options. Hypertonic saline is effec-tive for control of increased intracranialpressure (ICP) after severe head injury.Effective doses as a continuous infusionof 3% saline range between 0.1 and 1.0mL/kg of body weight per hour, adminis-tered on a sliding scale. The minimumdose needed to maintain ICP �20 mmHg should be used. Pending multiple-center confirmation of effectiveness andlack of toxicity, caution should be exer-cised in widespread adoption of this ther-apy.

Mannitol is effective for control of in-creased ICP after severe traumatic braininjury (TBI). Effective bolus doses rangefrom 0.25 g/kg of body weight to 1 g/kg ofbody weight.

Euvolemia should be maintained byfluid replacement. A Foley catheter is rec-ommended in these patients to avoidbladder rupture.

Serum osmolarity should be main-tained below 320 mOsm/L with mannitoluse, whereas a level of 360 mOsm/L ap-pears to be tolerated with hypertonic sa-line, even when used in combination withmannitol.

The choice of mannitol or hypertonicsaline as a first-line hyperosmolar agentshould be left to the treating physician.

D. Indications from Adult Guidelines.Most of the pediatric options regardingmannitol, listed previously, mirror thosestated in the adult guidelines (1). Theadult guidelines only addressed the use ofmannitol and not hypertonic saline. Man-nitol administration achieved guidelinestatus for the control of intracranial hy-pertension in the adult document.

II. OVERVIEW

Mannitol is a cornerstone in the man-agement of raised ICP in pediatric andadult TBI. In a recent survey that included70% of the pediatric intensive care units inthe United Kingdom (2), mannitol was usedin pediatric TBI in all of the units. Despitethis fact, mannitol has not been subjectedto controlled clinical trials vs. placebo,other osmolar agents, or other mechanism-based therapies in children. Most of theearly and recent study on the use of man-nitol focused on the treatment of adults(3–15). Either children were excluded orthe composition or outcome of the pediat-ric tail was not defined (3–18). In a keystudy, low mean ages were reported, indi-cating the inclusion of many adolescentsand/or children (3). Studies in which thepediatric composition is clearly defined arediscussed subsequently. The use of hyper-osmolar therapy in the management of in-fants and children with severe TBI, how-ever, is an area in which there has beenmuch contemporary study. This work, dis-cussed subsequently, has reported on thesuccessful use of hypertonic saline to pre-vent or treat increased ICP in infants andchildren with severe TBI.

In constructing an evidence-based doc-ument on the use of hyperosmolar therapyin pediatric TBI, one must recognize thatthe guideline level evidence supporting theuse of mannitol in adults relies on studiesthat often included but did not define theproportion of children. There is a largebody of clinical experience using mannitolin infants and children but a limited num-ber of pediatric studies (class III only) thatdocument efficacy of mannitol. In contrast,several recent studies support the use ofhypertonic saline in infants and childrenwith severe TBI. However, the use of hyper-tonic saline has been limited to a smallnumber of centers, and clinical experiencewith the use of hypertonic saline is limitedcompared with clinical experience withmannitol.

III. PROCESS

We searched Medline and Healthstarfrom 1966 to 2001 by using the searchstrategy for this question (see AppendixA) and supplemented the results with lit-erature recommended by peers or identi-fied from reference lists. Of 46 potentiallyrelevant studies, six were used as evi-dence for this question (Table 1).


Intravenous administration of hyper-osmosal agents was shown to reduce ICPearly in the 20th century (19). Wise andChater (20) introduced mannitol intoclinical use in 1961. Despite widespreaduse of a number of osmolar agents (man-nitol, urea, glycerol) up until the late1970s (20), mannitol gradually replacedother hyperosmolar agents in the man-agement of intracranial hypertension.

Mannitol can reduce ICP by two dis-tinct mechanisms. Mannitol rapidly re-duces ICP by reducing blood viscositywith a resultant decrease in blood vesseldiameter (21–24). This occurs as a resultof cerebral blood flow (CBF) autoregula-tion. The level of CBF is maintained, de-spite a reduction in blood viscosity,through reflex vasoconstriction. Thus,cerebral blood volume and ICP decrease.This mechanism is dependent on intactviscosity autoregulation of CBF, which islinked to blood pressure autoregulationof CBF (21, 23, 24). The effect of manni-tol administration on blood viscosity israpid but transient (�75 mins) (22).Mannitol administration also reduces ICPby an osmotic effect, which developsmore slowly (over 15–30 mins), due tothe gradual movement of water fromparenchyma into the circulation. The effectpersists up to 6 hrs and requires an intactblood-brain barrier (25, 26). Mannitol mayaccumulate in injured brain regions (27),where a reverse osmotic shift may occur—with fluid moving from the intravascularcompartment into the brain parenchyma—


possibly increasing ICP. This phenomenonhas been suggested to be most markedwhen mannitol is present in the circulationfor extended periods of time, supportingthe use of intermittent boluses (28). Man-nitol possesses antioxidant effects (29), butthe contribution of this mechanism to itsoverall efficacy remains unclear.

Mannitol is excreted unchanged inurine, and a risk of the development ofacute tubular necrosis and renal failurehas been suggested with mannitol admin-istration with serum osmolarity levels�320 mOsm in adults (30–32). However,the literature supporting this finding islimited in scope and was generated at atime when dehydration therapy was com-mon. A euvolemic hyperosmolar stategenerally is targeted with contemporarycare. Much higher levels of serum osmo-larity (365 mOsm) appear to be well tol-erated in children when induced with hy-pertonic saline (33, 34). It is unclear ifthis threshold for complications withmannitol results from concomitant dehy-dration, the use of mannitol rather thanhypertonic saline, or differences between

adults and children in their susceptibilityto nephrotoxicity. As stated in the adultguidelines, few data exist supporting theconcomitant use of diuretics and manni-tol to reduce ICP (30).

James (26) carried out a retrospectivestudy of 60 patients (1–73 yrs of age)treated with intravenous mannitol (0.18–2.5 g/kg per dose) for increased ICP (�25mm Hg). Although cited as class III evi-dence in the adult guidelines (30), thisstudy included a large number of chil-dren. After bolus dosing, ICP decreasedby �10% for 116 of the 120 doses. Bolusdoses of 0.25 g/kg, 0.5 g/kg, and �1.0g/kg reduced ICP in 25%, 78%, and 98%of cases, respectively. This contrasts thework of Marshall et al. (4), who reportedequivalence for doses between 0.25 and1.0 g/kg in adults. The mean time for thereturn of ICP to baseline was 196 mins,with the shortest reduction being 40mins. Other concomitant therapies usedfor patient management in this study in-cluded dexamethasone, neuromuscularblockade and hyperventilation, barbitu-

rates, and/or hypothermia, in refractorycases.

Miller et al. (35) reported a compari-son of mannitol (0.5 g/kg) vs. a hypnotic(thiopentone 5 mg/kg and/or gamma hy-droxyl butyrate 60 mg/kg) for refractoryICP (�25–30 mm Hg in 17 patients, in-cluding six children, 3–17 yrs of age).Response to therapy was defined in eachpatient in the study, allowing individualassessment of children. Mannitol wasfound to be superior to the hypnotic infive cases, the hypnotic was superior tomannitol in three cases, and both wereeffective in five cases. All of the childrenresponded to either or both agents. Hyp-notics were more effective in cases ofdiffuse TBI, whereas mannitol was effec-tive in focal TBI, but the sample size inthis study was limited.

In other studies with exclusively pedi-atric patients (36, 37), mannitol repre-sented a key component of therapy oreven defined a specific subgroup; how-ever, the specific effect of mannitol onICP or outcome was not reported. In con-trast, Bruce and coworkers (38) sug-


References Description of Study Data Class Conclusion

James (26), 1980 Retrospective study of 60 patients (1–73 yrs of age)treated with mannitol (0.18–2.5 g/kg per dose) forincreased ICP (�25 mm Hg). In 18 patients (12 withTBI, mean age 14 yrs), bolus mannitol was followedby intravenous continuous infusion (6–100 hrs).

III ICP decreased by �10% after 116 of the 120 doses. Bolus doses�0.5 g/kg produced an ICP reduction 97% of the time. Otherconcomitant therapies included dexamethasone, neuromuscularblockade and hyperventilation, barbiturates, and hypothermia, inrefractory cases.

Miller et al. (35),1993

Paired comparison of mannitol (0.5 g/kg) hypnotic(thiopentone 5 mg/kg and/or GABA 60 mg/kg) forrefractory ICP �25 mm Hg or �30 mm Hg in 17patients, including six children (3–17 yrs).

III Mannitol was superior to hypnotic in five cases; hypnotic wassuperior to mannitol in three cases; both were effective in fivecases; and neither was effective in four cases. Hypnotics weremore effective in cases of diffuse TBI; mannitol was effective infocal TBI. Other concomitant therapies included neuromuscularblockade and sedation.

Fisher et al. (52),1992

Double-blind crossover study comparing 3% saline(1025 mOsm/L) and 0.9% saline (308 mOsm/L) in18 children with severe TBI. Doses of each agentwere equal and ranged between 6.5 and 10 mL/kg ineach patient.

III (class II for ICP) During the 2-hr trial, hypertonic saline was associated with a lowerICP and reduced need for additional interventions (thiopental andhyperventilation) to control ICP pressure. Serum sodiumconcentration increased �7 mEq/L after 3% saline.

Khanna et al. (34),2000

Prospective study of administration of 3% saline (1025mOsm/L) on a sliding scale to maintain ICP �20mm Hg in ten children with raised ICP resistant toconventional therapy.

III (class II for ICP) A significant reduction in ICP spikes and an increase in CPP wereobserved during treatment with 3% saline. The mean duration oftreatment was 7.6 days, and the mean highest serum sodiumconcentration and osmolarity were 170.7 mEq/L and 364.8mOsm/L, respectively. Reversible renal failure developed in twopatients. Sustained hypernatremia and hyperosmolarity weresafely tolerated in pediatric patients.

Simma et al. (53),1998

Open-randomized prospective study of hypertonicsaline (598 mOsm/L) vs. lactated Ringer’sadministered over the initial 3 days in 35consecutive children with severe TBI.

III (class II for ICP) Patients treated with hypertonic saline required fewer interventionsthan those treated with lactated Ringer’s to maintain ICP control.The hypertonic saline treatment group also had shorter length ofICU stay, shorter duration of mechanical ventilation, and fewercomplications than the lactated Ringer’s-treated group.

Peterson et al. (33),2000

Retrospect study of the use of a continuous infusion ofhypertonic saline (3%) titrated to reduce ICP �20mm Hg in 68 infants and children with closed headinjury. Doses of 0.1–1.0 mL�kg�1�hr�1 resulting inmean daily dosages between �11 and 27mL�kg�1�day�1 were used. There was no controlgroup.

III Three patients died of uncontrolled ICP, and mortality rate waslower than expected based on trauma and injury severity score.No patients developed renal failure. Concomitant therapy includedneuromuscular blockade, fentanyl, sedation, hyperventilation, andbarbiturates. CSF drainage was rarely used. Hypertonic saline(3%) appeared safe. Central pontine myelinolysis, subarachnoidhemorrhage, or rebound increases in ICP were not observed.

ICP, intracranial pressure; TBI, traumatic brain injury; GABA, �-aminobutyric acid; CPP, cerebral perfusion pressure; ICU, intensive care unit.


gested restricted mannitol use in the1980s and 1990s. Based in part on thesuggested hyperemic response to TBI inchildren (39) and the possible increase incerebral blood volume with mannitol ad-ministration (when pressure and viscos-ity autoregulation are defective), it wasproposed that mannitol administrationcarried special risk in pediatric patientswith diffuse cerebral swelling early afterTBI. The authors recommended againstthe use of mannitol in the absence of ahigh probability of mass lesion. Since re-cent studies showed that early posttrau-matic CBF generally is reduced, ratherthan increased, in infants and children(40), this hypothetical risk of mannitoladministration in pediatric patientsshould not a priori dissuade cliniciansfrom administration in the initial 48 hrs.

In the initial description in 1919 of thereduction in ICP by intravenous admin-istration of hyperosmosal agents, hyper-tonic saline was the agent used (19). Theuse of hypertonic saline in the treatmentof increased ICP, however, failed to gainclinical acceptance. Resurgence in inter-est in this treatment for raised ICP re-sulted from the report of Worthley et al.(41), who described two cases in whichhypertonic saline (small volumes of anextremely hypertonic solution, �29% sa-line) reduced refractory ICP elevations.One of those cases involved treatment ofa 17-yr-old boy with TBI. In the last de-cade, numerous laboratories have studiedthe use of small volume hypertonic salinein resuscitation of hemorrhagic shockwith or without TBI in experimentalmodels and in humans (42–45). Thesestudies are summarized in several recentreviews (46, 47).

Like mannitol, the penetration of so-dium across the blood-brain barrier islow (46). Sodium thus shares both thefavorable rheologic and osmolar gradienteffects involved in the reduction in ICP bymannitol. Hypertonic saline also exhibitsseveral theoretical beneficial effects in-cluding restoration of normal cellularresting membrane potential and cell vol-ume (48, 49), stimulation of atrial natri-uretic peptide release (50), inhibition ofinflammation (reviewed in Ref. 46), andenhancement of cardiac output (51). Pos-sible side effects of hypertonic saline in-clude rebound in ICP, central pontinemyelinolysis, and subarachnoid hemor-rhage (reviewed in Ref. 46).

Hypertonic saline has been the subjectof considerable investigation with threeclass II studies (for ICP) and one class III

study in �130 pediatric patients with se-vere TBI. It should be pointed out thatnone of these studies produced class IIdata demonstrating a beneficial effect onlong-term outcome.

Fisher et al. (52) carried out a double-blind crossover study comparing 3% sa-line and 0.9% saline in 18 children withsevere TBI. Bolus doses of each agentwere equal and ranged between 6.5 and10 mL/kg. During the 2-hr trial, serumsodium concentration increased about 7mEq/L, and hypertonic saline was associ-ated with a lower ICP and reduced needfor additional interventions. Concomi-tant therapies used for patient manage-ment in this study included thiopental,dopamine, mannitol, and hyperventila-tion. Cerebrospinal fluid drainage wasnot used.

Khanna et al. (34) reported a prospec-tive study with administration of 3% sa-line (514 mEq/L) on a sliding scale tomaintain ICP �20 mm Hg in ten chil-dren with increased ICP resistant to con-ventional therapy. The maximal rate ofincrease in serum sodium was 15mEq·L�1·day�1, and the maximal rate ofdecrease in serum sodium was 10mEq·L�1·day�1. A reduction in ICPspikes and an increase in cerebral perfu-sion pressure were seen during treatmentwith 3% saline. The mean duration oftreatment was 7.6 days, and the meanhighest serum sodium concentration andosmolarity were 170.7 mEq/L and 364.8mOsm/L, respectively. The maximum se-rum osmolarity in an individual patientwas 431 mOsm/L. Sustained hypernatre-mia and hyperosmolarity were generallywell tolerated in the children. Two pa-tients, both with sepsis and/or multipleorgan failure, developed acute renal fail-ure. Both received continuous veno-venous hemofiltration and recovered re-nal function.

Simma et al. (53) carried out an openrandomized prospective study of hyper-tonic saline (598 mOsm/L) vs. lactatedRinger’s solution administered over theinitial 3 days in 35 children with severeTBI. Patients treated with hypertonic sa-line required fewer interventions (includ-ing mannitol use) to control ICP thanthose treated with lactated Ringer’s solu-tion. Patients in the hypertonic salinetreatment group also had shorter lengthof intensive care unit stay, shorter dura-tion of mechanical ventilation, and fewercomplications than the lactated Ringer’s-treated group.

Peterson et al. (33), reported a retro-spective study on the use of a continuousinfusion of 3% saline titrated to reduce ICPto �20 mm Hg in 68 infants and childrenwith TBI. The mean daily doses of hyper-tonic saline over a 7-day period ranged be-tween 11 and 27 mL·kg�1·day�1. There wasno control group, but only three patientsdied of uncontrolled ICP, and mortality ratewas lower than expected based on Traumaand Injury Severity Score categorization.No patient with a serum sodium concen-tration �180 mEq/L had a good outcome.No patients developed renal failure. Con-comitant therapies included sedation, neu-romuscular blockade, mannitol, hyperven-tilation, and barbiturates, but cerebrospinalfluid drainage was used in only three chil-dren. The mean daily dose of mannitol was1–2 g·kg�1·day�1. Hypertonic saline ap-peared to be safe. Rebound in ICP, centralpontine myelinolysis, and subarachnoidhemorrhage were not seen.


Based on an evidence table in theadult guidelines (30) (two class I and fiveclass II studies), mannitol was deemed tobe effective for controlling increased ICPafter severe TBI, with effective dosesranging from 0.25 g/kg to 1 g/kg of bodyweight. Limited data in adults suggestthat intermittent boluses may be moreeffective than a continuous infusion. Sev-eral key studies were cited. Schwartz etal. (3) carried out a randomized compar-ison of mannitol vs. barbiturates in 59adults with severe TBI. Cerebral perfu-sion pressure was better maintained inthe mannitol-treated group. Gabb et al.(54) and Rosner and Coley (11) reportedsimilar effects. Fortune et al. (14) com-pared mannitol, ventriculostomy drain-age, and hyperventilation to control ICPin 22 adults. Mannitol was the most ef-fective. Use of mannitol for TBI recentlywas subjected to Cochrane review, and noconclusion could be reached regardingefficacy vs. placebo or any other therapy(55).

V. SUMMARY

Two class III studies support the use ofmannitol in pediatric TBI. Neither ofthese studies included exclusively pediat-ric patients. One must thus weigh thevalue of long-standing clinical acceptanceand safety of a therapy (mannitol) that


has limited evidentiary support (two classIII studies) of its efficacy against a newertherapy (hypertonic saline) with a limitedclinical experience but reasonably goodperformance in contemporary clinical tri-als (three class II studies for ICP and oneclass III study). Bolus administration ofmannitol or continuous infusion of 3%saline is supported. Thus, in pediatricTBI, there is guideline-level support forhypertonic saline to treat increased ICPbut limited clinical experience. In con-trast, there is only class III evidence formannitol, despite long-standing clinicalacceptance. Until one or more directcomparisons between these two therapiesare carried out in infants and childrenwith severe TBI, the choice of eithermannitol or hypertonic saline in themanagement of pediatric TBI is a matterof physician preference.


Additional investigation is neededcomparing mannitol administration withhypertonic saline, particularly studiesevaluating long-term neurologic out-come. Similarly, study of the use of more

aggressive hyperosmolar therapy withother second-tier therapies is needed, in-cluding investigation of the prevention ofintracranial hypertension by continuousinfusion of hypertonic saline vs. treat-ment in response to spikes. Documenta-tion of the effect of mannitol in studiesrestricted to infants and children isneeded. Similarly lacking are studies invictims of child abuse. Despite the overallquality of the investigations assessing theeffect on ICP, the use of hypertonic salinehas been limited to a small number ofpediatric centers, and a number of factorsinvolved in patient management, such asthe use of concomitant therapies like ce-rebrospinal fluid drainage and the extentof use of specific second-tier therapies,varies greatly between centers. Additionalstudy is needed. Optimal dosing and bet-ter definitions of treatment threshold areneeded for the development of nephro-toxicity, rebound intracranial hyperten-sion, central pontine myelinolysis, andother complications with mannitol andhypertonic saline.

REFERENCES


2. Segal S, Gallagher AC, Shefler AG, et al:Neonatal and pediatric intensive care: Surveyof the use of intracranial pressure monitor-ing in children in the United Kingdom. In-tensive Care Med 2001; 27:236–239

3. Schwartz ML, Tator CH, Rowed DW, et al:The University of Toronto Head Injury Treat-ment Study: A prospective, randomized com-parison of pentobarbital and mannitol. CanJ Neurol Sci 1984; 11:434–440

4. Marshall LF, Smith RW, Rauscher LA, et al:Mannitol dose requirements in brain-injuredpatients. J Neurosurg 1978; 48:169–172

5. Mendelow AD, Teasdale GM, Russell T, et al:Effect of mannitol on cerebral blood flow andcerebral perfusion pressure in human headinjury. J Neurosurg 1985; 63:43–48

6. McGraw CP, Alexander E, Howard G: Effectof dose and dose schedule on the response ofintracranial pressure to mannitol. Surg Neu-rol 1978; 10:127–130

7. McGraw CP, Howard G: Effect of mannitol onincreased intracranial pressure. Neurosur-gery 1983; 13:269–271

8. Muizelaar JP, Lutz HA III, Becker DP: Effectof mannitol on ICP and CBF and correlationwith pressure autoregulation in severelyhead-injured patients. J Neurosurg 1984; 61:700–706

9. Miller JD, Leech P: Effects of mannitol andsteroid therapy on intracranial volume-

pressure relationships in patients. J Neuro-surg 1975; 42:274–281

10. Node Y, Nakazawa S: Clinical study of man-nitol and glycerol on raised intracranial pres-sure and on their rebound phenomenon. AdvNeurol 1990; 52:359—363

11. Rosner MJ, Coley I: Cerebral perfusion pres-sure: A hemodynamic mechanism of manni-tol and the postmannitol hemogram. Neuro-surgery 1987; 21:147–156

12. Biestro A, Alberti R, Galli R, et al: Osmoth-erapy for increased intracranial pressure:Comparison between mannitol and glycerol.Acta Neurochir (Wien) 1997; 139:725–733

13. Bingham WF: The limits of cerebral dehydra-tion in the treatment of head injury. SurgNeurol 1986; 25:340–345

14. Fortune JB, Feustel PJ, Graca L, et al: Effectof hyperventilation, mannitol, and ventricu-lostomy drainage on cerebral blood flow afterhead injury. J Trauma 1995; 39:1091–1099

15. Procaccio F, Menasce G, Sacchi L, et al: Ef-fects of thiopentone and mannitol on cere-bral perfusion pressure and E. E. G. in headinjured patients with intracranial hyperten-sion. Agressologie 1991; 32:381–385

16. Rosner MJ, Rosner SD, Johnson AH: Cerebralperfusion pressure: Management protocoland clinical results. J Neurosurg 1995; 83:949–962

17. Smith HP, Kelly DL, McWhorter JM, et al: Com-parison of mannitol regimens in patients withsevere head injury undergoing intracranial mon-itoring. J Neurosurg 1986; 65:820–824

18. Nara I, Shiogai T, Hara M, et al: Comparativeeffects of hypothermia, barbiturate, and os-motherapy for cerebral oxygen metabolism,intracranial pressure, and cerebral perfusionpressure in patients with severe head injury.Acta Neurochir 1998; 71(Suppl):22–26

19. Weed LH, McKibben PS: Pressure changes inthe cerebro-spinal fluid following intrave-nous injection of solutions of various con-centrations. Am J Physiol 1919; 48:512–530

20. Wise BL, Chater N: Use of hypertonic man-nitol solutions to lower cerebrospinal fluidpressure and decrease brain bulk in man.Surg Forum 1961; 12:398–399

21. Levin AB, Duff TA, Javid MJ: Treatment of in-creased intracranial pressure: A comparison ofdifferent hyperosmotic agents and the use of thio-pental. Neurosurgery 1979; 5:570–575

22. Muizelaar JP, Lutz HA, Becker DP: Effect ofmannitol on ICP and CBF and correlation withpressure autoregulation in severely head in-jured patients. J Neurosurg 1984; 61:700–706

23. Muizelaar JP, Wei EP, Kontos HA, et al: Man-nitol causes compensatory vasoconstrictionand vasodilation in response to blood viscos-ity changes. J Neurosurg 1983; 59:822–828

24. Muizelaar JP, Wei EP, Kontos HA, et al: Ce-rebral blood flow is regulated by changes inblood pressure and in blood viscosity alike.Stroke 1986; 17:44–48

25. Bouma GJ, Muizelaar JP: Cerebral bloodflow, cerebral blood volume, and cerebrovas-cular reactivity after severe head injury.J Neurotrauma 1992; 9:S333–S348

O ne must thus

weigh the value

of long-standing

clinical acceptance and

safety of a therapy (manni-

tol) that has limited eviden-

tiary support (two class III

studies) of its efficacy

against a newer therapy (hy-

pertonic saline) with a lim-

ited clinical experience but

reasonably good perfor-

mance in contemporary

clinical trials (three class II

studies for intracranial pres-

sure and one class III study).


26. James HE: Methodology for the control ofintracranial pressure with hypertonic man-nitol. Acta Neurochir 1980; 51:161–172

27. Kaieda R, Todd MM, Cook LN, et al: Acuteeffects of changing plasma osmolality andcolloid oncotic pressure on the formation ofbrain edema after cryogenic injury. Neuro-surgery 1989; 24:671–677

28. Kaufmann AM, Cardoso ER: Aggravation ofvasogenic cerebral edema by multiple-dosemannitol. J Neurosurg 1992; 77:584–589

29. Kontos HA, Hess ML: Oxygen radicals andvascular damage. Adv Exp Med Biol 1983;161:365–375

30. Management and Prognosis of Severe TraumaticBrain Injury. Part 1: Guidelines for the manage-ment of severe traumatic brain injury. Use ofmannitol. J Neurotrauma 2000; 17:521–525

31. Becker DP, Vries JK: The alleviation of in-creased intracranial pressure by the chronicadministration of osmotic agents. In: Intra-cranial Pressure. Brock M, Dietz H (Eds).Berlin, Springer, 1972, pp 309–315

32. Feig PU, McCurdy DK: The hypertonic state.N Engl J Med 1977; 297:1444–1454

33. Peterson B, Khanna S, Fisher B, et al: Pro-longed hypernatremia controls elevated in-tracranial pressure in head-injured pediatricpatients. Crit Care Med 2000; 28:1136–1143

34. Khanna S, Davis D, Peterson B, et al: Use ofhypertonic saline in the treatment of severerefractory posttraumatic intracranial hyper-tension in pediatric traumatic brain injury.Crit Care Med 2000; 28:1144–1151

35. Miller JD, Piper IR, Dearden NM: Manage-ment of intracranial hypertension in headinjury: Matching treatment with cause. ActaNeurochir 1993; 57:152–159

36. Kasoff SS, Lansen TA, Holder D, et al: Aggressive

physiologic monitoring of pediatric head traumapatients with elevated intracranial pressure. Pedi-atr Neurosci 1988; 14:241–249

37. Eder HG, Legat JA, Gruber W: Traumaticbrain stem lesions in children. Childs NervSyst 2000; 16:21–24

38. Raphaely RC, Swedlow DB, Downes JJ, et al: Man-agement of severe pediatric head trauma. PediatrClin North Am 1980; 27:715–727

39. Bruce DA, Alavi A, Bilaniuk L, et al: Diffusecerebral swelling following head injuries inchildren: The syndrome of “malignant brainedema.” J Neurosurg 1981; 54:170–178

40. Adelson PD, Clyde B, Kochanek PM, et al:Cerebrovascular response in infants andyoung children following severe traumaticbrain injury: A preliminary report. PediatrNeurosurg 1997; 26:200–207

41. Worthley LI, Cooper DJ, Jones N: Treatmentof resistant intracranial hypertension withhypertonic saline. Report of two cases. J Neu-rosurg 1988; 68:478–481

42. Vassar M, Fischer RP, O’Brien P, et al: A mul-ticenter trial for resuscitation of injured pa-tients with 7.5% sodium chloride. The effect ofadded dextran 70. The Multicenter Group forthe Study of Hypertonic Saline in Trauma Pa-tients. Arch Surg 1993; 128:1003–1011

43. Prough DS, Whitley JM, Taylor CL, et al:Regional cerebral blood flow following resus-citation from hemorrhagic shock with hyper-tonic saline. Influence of a subdural mass.Anesthesiology 1991; 75:319–327

44. Shackford SR, Bourguignon PR, Wald SL, etal: Hypertonic saline resuscitation of patientswith head injury: A prospective, randomizedclinical trial. J Trauma 1998; 44:50–58

45. Walsh JC, Zhuang J, Shackford SR: A com-parison of hypertonic to isotonic fluid in the

resuscitation of brain injury and hemor-rhagic shock. J Surg Res 1991; 50:284–292

46. Qureshi AI, Suarez JI: Use of hypertonic sa-line solutions in treatment of cerebral edemaand intracranial hypertension. Crit Care Med2000; 28:3301–3313

47. Zornow MH, Prough DS: Fluid managementin patients with traumatic brain injury. NewHoriz 1995; 3:488–498

48. Nakayama S, Kramer GC, Carlsen RC, et al:Infusion of very hypertonic saline to bledrats: Membrane potentials and fluid shifts.J Surg Res 1985; 38:180–186

49. McManus ML, Soriano SG: Rebound swellingof astroglial cells exposed to hypertonic man-nitol. Anesthesiology 1998; 88:1586–1591

50. Arjamaa O, Karlqvist K, Kanervo A, et al:Plasma ANP during hypertonic NaCl infusionin man. Acta Physiol Scand 1992; 144:113–119

51. Moss GS, Gould SA: Plasma expanders. Anupdate. Am J Surg 1988; 155:425–434

52. Fisher B, Thomas D, Peterson B: Hypertonicsaline lowers raised intracranial pressure inchildren after head trauma. J Neurosurg An-esthesiol 1992; 4:4–10

53. Simma B, Burger R, Falk M, et al: A prospec-tive, randomized, and controlled study offluid management in children with severehead injury: Lactated Ringer’s solution ver-sus hypertonic saline. Crit Care Med 1998;26:1265–1270

54. Gaab MR, Seegers K, Smedema RJ, et al: Acomparative analysis of THAM in traumaticbrain edema. Acta Neurochir 1990;51(Suppl):320–323

55. Schierhout G, Roberts I: Mannitol for acutetraumatic brain injury (Cochrane Review).Cochrane Database Syst Rev 2000; 2

APPENDIX: LITERATURE SEARCH STRATEGIESSEARCHED MEDLINE AND HEALTHSTAR FROM 1966 TO 2001

Chapter 11. Hyperosmolar Therapy

1. exp craniocerebral trauma/2. head injur$.tw.3. brain injur$.tw.4. 1 or 2 or 35. hyperosmolar therapy.mp.6. hyperosmolar treatment.mp.7. Fluid therapy/ or “fluid therapy”.mp.8. saline solution, hypertonic/9. osmolar concentration/

10. 5 or 6 or 7 or 8 or 911. 4 and 1012. limit 11 to (human and english language)13. limit 12 to (infant �1 to 23 months� or preschool child �2 to 5 years� or child �6 to 12 years� or adolescence �13

to 18 years�)14. from 13 keep 1–13


Chapter 12. Use of hyperventilation in the acute management ofsevere pediatric traumatic brain injury

I. RECOMMENDATIONS



C. Options. Mild or prophylactic hyper-ventilation (PaCO2 �35 mm Hg) in childrenshould be avoided.

Mild hyperventilation (PaCO2 30–35mm Hg) may be considered for longerperiods for intracranial hypertension re-fractory to sedation and analgesia, neuro-muscular blockade, cerebrospinal fluiddrainage, and hyperosmolar therapy.

Aggressive hyperventilation (PaCO2

�30 mm Hg) may be considered as asecond tier option in the setting of refrac-tory hypertension. Cerebral blood flow(CBF), jugular venous oxygen saturation,or brain tissue oxygen monitoring is sug-gested to help identify cerebral ischemiain this setting.

Aggressive hyperventilation therapytitrated to clinical effect may be necessaryfor brief periods in cases of cerebral her-niation or acute neurologic deterioration.

D. Indications from the Adult Guide-lines. The adult guidelines recommended(1) at the level of a treatment standardthat in the absence of increased intracra-nial pressure (ICP), chronic prolongedhyperventilation therapy (PaCO2 of �25mm Hg) should be avoided after severeTBI. At the level of a treatment guideline,it was recommended that prophylactichyperventilation (PaCO2 �35 mm Hg)therapy during the first 24 hrs after se-vere TBI should be avoided because it cancompromise cerebral perfusion during atime when CBF is reduced.

It was recommended as a treatmentoption that hyperventilation therapy maybe necessary for brief periods when thereis acute neurologic deterioration or forlonger periods if there is intracranial hy-pertension refractory to sedation, paraly-

sis, cerebrospinal fluid drainage, and os-motic diuretics. Jugular venous oxygensaturation, arterial jugular venous oxy-gen content differences, brain tissue ox-ygen monitoring, and CBF monitoringmay help to identify cerebral ischemia ifhyperventilation, resulting in PaCO2 val-ues �30 mm Hg, is necessary.

II. OVERVIEW

Aggressive hyperventilation therapyhas been used in the management of se-vere pediatric TBI for the rapid reductionof ICP since the 1970s. In an uncon-trolled study, Bruce et al. (2) used a pro-tocol that included aggressive hyperven-tilation and reported very goodoutcomes. This approach was based onthe assumption that hyperemia was com-mon after pediatric TBI. Hyperventilationtherapy also was thought to benefit theinjured brain through a variety of mech-anisms including reduction of brain aci-dosis (3), improvement of cerebral me-tabolism (4), restoration of bloodpressure autoregulation of cerebral bloodflow (5), and increasing perfusion to isch-emic brain regions (local inverse steal)(6).

More recent pediatric studies haveshown that hyperemia is uncommon andalso have raised concerns about the safetyof hyperventilation therapy. Study of theeffect of hyperventilation in children hasfocused on assessment of cerebral physi-ologic variables. The effect of hyperventi-lation therapy on outcome in infants andchildren with severe TBI has not beendirectly compared with other therapiessuch as hyperosmolar agents, barbitu-rates, hypothermia, or early decompres-sive craniectomy.

Hyperventilation reduces ICP by in-ducing hypocapnia. This leads to cerebralvasoconstriction and a reduction in CBF.This is accompanied by a reduction incerebral blood volume, resulting in a de-crease in ICP. However, hyperventilation

is associated with a risk of iatrogenicischemia. In an experimental model, Mui-zelaar et al. (7) reported that the vaso-constrictor effect of hyperventilation wassustained for a period of �24 hrs.Chronic hyperventilation depletes braintissue interstitial bicarbonate bufferingand causes cerebral circulation to be-come hyper-responsive to subsequent in-creases in PaCO2. In addition, the respi-ratory alkalosis that accompanieshyperventilation causes a left shift of thehemoglobin-oxygen dissociation curve,which may impair delivery of oxygen totissue.

The assumption of benefit from hyper-ventilation recently has been challenged.Recent clinical studies in mixed adult andpediatric populations also have demon-strated that hyperventilation may de-crease cerebral oxygenation and may in-duce brain ischemia (8–11). After TBI,the CBF response to changes in PaCO2 canbe unpredictable and should be specifi-cally monitored.

III. PROCESS



Diffuse cerebral swelling is a commonfinding in pediatric patients with severeTBI (12, 13). Increased cerebral bloodvolume and CBF had been considered tobe the unique cause of this diffuse swell-ing, and raised ICP in children and ag-gressive hyperventilation was advocated(14). In the classic study by Bruce et al.(2), 36 of 76 children with severe TBIwere found to have diffuse cerebral swell-ing on CT scan. Six patients, ages 14–21


yrs, were found to have CBF that wasnormal or above normal. In three pa-tients, CBF decreased back to control lev-els after diffuse swelling had resolved.Consequently, aggressive hyperventila-tion (PaCO2 23–25 mm Hg) was advocatedand mannitol was discouraged.

There are now data to suggest thathyperemia is not as common as previ-ously thought (15). In a series of 80 nor-mal, unanesthetized children, CBFranged from 40 mL·100 g�1·min�1 in thefirst 6 months of life to a peak of 108mL·100 g�1·min�1 at age 3–4 yrs, declin-ing to 71 mL·100 g�1·min�1 after age 9yrs. Similarly, Chiron et al. (16) demon-strated that CBF ranged from about 50mL·100 g�1·min�1 in normal neonates toa peak of 71 mL·100 g�1·min�1 at 5 yrs.After age 19, CBF gradually decreased toadult levels. Thus, posttraumatic CBFmay not be greater than normal in chil-dren. However, caution should be exer-cised in interpreting these studies be-cause techniques used to measure CBFdiffered between reports.

Adelson et al. (17) studied 30 childrenwith severe TBI, all �8 yrs of age. Seven-ty-seven percent had CBF �20 mL·100g�1·min�1 on admission. Children weretreated with a protocol including mild(PCO2 32–35 mm Hg) hyperventilationand barbiturate coma (60%). CBF washighest at 24–48 hrs (59.6 � 4.5 mL·100g�1·min�1) and decreased (�50 mL·100g�1·min�1) after 3 days. Any child withglobal CBF of 20 mL·100 g�1·min�1 orless at any time had a poor outcome. CBFof �55 mL·100 g�1·min�1 was associatedwith a higher proportion of children witha good outcome.

Muizelaar et al. (18) studied 32 chil-dren with severe TBI (age 3–18 yrs). Theaverage CBF was only 44 � 22 mL·100g�1·min�1, which is considerably lowerthan the average of 68 � 4 mL·100g�1·min�1 found in four normal unanes-thetized children. No correlation wasfound between CBF and ICP.

Although the effect of hyperventila-tion on long-term outcome has not beendirectly addressed in pediatric TBI, sev-eral reports have described the effects ofhyperventilation on CBF and brain phys-iology. Stringer et al. (19) studied localCBF and vascular reactivity before andafter hyperventilation in 12 patients in-cluding three children with severe TBI.Hyperventilation-induced blood flow re-ductions affected both injured and appar-ently intact areas of the brain and werenot reflected by ICP measurement.

Sharples et al. (20) investigated CBF,arterial jugular venous oxygen difference,and cerebral metabolic rate in 21 chil-dren with TBI. No fundamental differencebetween adults and children in the patho-physiologic response of CBF to severe TBIwas found. Absolute cerebral hyperemiawas uncommon. Raised ICP was associ-ated with low, rather than increased,CBF. Cerebral metabolic rate was initiallynormal in 81% of children with TBI.These data do not support the hypothesisthat ICP increases as a result of excessiveCBF in children with TBI. Based on thisstudy, and on a subsequent study of ce-rebral vascular reactivity (21), the au-thors recommended maintaining a nor-mal PaCO2.

Skippen et al. (22) found that hyper-emia was uncommon in children with

severe TBI. However, CBF rates remainedabove the metabolic requirements ofmost children studied. A modest decreasein CBF and a much larger decrease incerebral oxygen consumption were foundat baseline. As PaCO2 was reduced withhyperventilation, CBF was decreased inalmost all patients despite decreased ICPand increased cerebral perfusion pres-sure. A clear relationship between hypo-carbia and frequency of cerebral ischemiawas seen. The frequency of regional isch-emia (CBF �18 mL·100 g�1·min�1) was28.9% during normocapnia and in-creased to 73.1% for PaCO2 �25 mm Hg.

The effect of hyperventilation therapyon outcome of infants and children withsevere TBI has not been directly com-pared with other therapies such as hyper-osmolar agents, barbiturates, hypother-mia, or early decompressive craniectomy.Surprisingly, outcome data reported byBruce et al. (2) in the late 1970s, whenaggressive hyperventilation (PaCO2 20–25mm Hg) represented the cornerstone oftherapy for pediatric TBI (5), has notbeen surpassed by contemporary proto-cols and only rarely equaled (23). Hyper-ventilation may have unique advantagesas a therapy in severe pediatric TBI; how-ever, it can only be supported as a secondtier therapy based on the current evi-dence.


The adult guidelines (1) conclude thatprophylactic hyperventilation (PaCO2 �35mm Hg) therapy during the first 24 hrsafter severe TBI should be avoided be-


References Description of StudyDataClass Conclusion

Skippen et al.(22), 1997

Prospective cohort, 23 children with isolated severe TBI, GCS �8.Ages 3 mos to 16 yrs, mean 11 yrs.PaCO2 was adjusted by minute ventilation to �35, 25–35, and �25

torr.Measured CBF, C(a–j)O2, CMRO2 � C(a–j)O2 � CBF.Follow-up GOS 6 mos post-ICU discharge.

II Severe TBI produced modest decrease in CBF, largerdecrease in cerebral oxygen consumption.

Hyperemia was uncommon, but measured CBF rateswere above metabolic requirements of most.

As PaCO2 reduced, ICP decreased and CPP increased.However, in almost all patients, CBF decreased.

Stringer et al.(19), 1993

Nonrandomized selected series of case studies.Twelve patients referred for CBF measurement. Three were children

with head trauma and coma, ages 1 mo, 6 yrs, and 8 yrs.Xenon-enhanced CT scans.Measured ICP, CPP, MAP, ETCO2, XeCT, CBF.

II Hyperventilation-induced ischemia occurs and affectsboth injured and apparently intact areas of braintissue.

TBI, traumatic brain injury; GCS, Glasgow Coma Scale; CBF, cerebral blood flow; C(a–j) O2, cerebral arteriojugular venous oxygen content difference;CMRO, cerebral metabolic rate; ICU, intensive care unit; CT, computed tomography; ICP, intracranial pressure; CPP, coronary perfusion pressure; MAP,mean arterial pressure; ETCO2, end-tidal CO2; XeCT, xenon-enhanced computed tomography.


cause it can compromise cerebral perfu-sion during a time when CBF is alreadyreduced. The guidelines (1) strongly con-tend that chronic prolonged hyperventi-lation therapy (PaCO2 �25 mm Hg)should be avoided after severe TBI in theabsence of increased ICP. It was empha-sized that the preponderance of the phys-iologic literature concludes that hyper-ventilation during the first few daysfollowing severe TBI, whatever thethreshold, is potentially deleterious inthat it can promote cerebral ischemia.

Specifically, CBF during the first dayafter injury is less than half that of nor-mal individuals (18, 24–31). During thefirst 24 hrs after injury, there is a directcorrelation between CBF and GlasgowComa Scale score or outcome (24, 29).Hyperventilation reduces CBF (22, 32–34) even further but does not consistentlyreduce ICP (35, 36) and may cause loss ofautoregulation (37). Aggressive hyper-ventilation may cause arterial jugular ve-nous oxygen and CBF to approach isch-emic levels.

In a prospective randomized clinicaltrial by Muizelaar et al. (8), 77 severe TBIpatients were randomized to a grouptreated with chronic prophylactic hyper-ventilation (PaCO2 of 25 � 2 mm Hg) orto a group kept relatively normocapneic(PaCO2 of 35 � 2 mm Hg). At 3 and 6months after injury, patients with initialGlasgow Coma Scale motor scores of 4–5in the hyperventilation group had a sig-nificantly worse outcome than did pa-tients in the normocapneic group. Statis-tically significant differences between thetwo groups were not found at 1 yr afterinjury; this was attributed to a type IIstatistical error since substantially fewerpatients were available for 1-yr follow-up.

V. SUMMARY

Hyperemia may not be as common insevere pediatric TBI as previously re-ported. Hyperventilation can reduce CBFto potentially ischemic levels. Addition-ally, the cerebrovascular response to hy-perventilation can be extremely variablefollowing TBI. Studies in children withsevere TBI raise the concern that thetoxicity of hyperventilation may be simi-lar to the toxicity that has been demon-strated in adults and related to adverseoutcome. Unfortunately, the precise rela-tionship between hyperventilation andoutcome has not been studied in childrenwith severe TBI.


● Studies to identify subgroups of pa-tients who might benefit from hyper-ventilation are needed.

● Studies are needed to address the tim-ing and duration of the optimal use ofhyperventilation.

● Studies to determine the optimal mon-itoring technique of patients undergo-ing hyperventilation are lacking.

● Studies are needed to address the in-fluence of age on the response to hy-perventilation

● The effects on long-term outcomeshould be addressed in all aspects ofresearch on hyperventilation. Hyper-emia may not be as common in severepediatric traumatic brain injury as pre-viously reported.

REFERENCES


2. Bruce D, Raphaely R, Goldberg A, et al:Pathophysiology, treatment and outcomefollowing severe head injury in children.Childs Brain 1979; 5:174–191

3. Gordon E, Rossanda M: The importance ofcerebrospinal fluid acid base status in thetreatment of unconscious patients with brainlesions. Acta Anesthesiol Scand 1968; 12:51–73

4. Obrist WD, Clifton GL, Robertson CS, et al:Cerebral metabolic changes induced by hy-perventilation in acute head injury. In: Cere-bral Vascular Disease 6. Meyer JS, et al.(Eds). Elsevier Science, 1987, pp 251–255,241–253

5. Raphaely RC, Swedlow DB, Downes JJ, et al:Management of severe pediatric headtrauma. Pediatr Clin North Am 1980; 27:715–727

6. Darby JM, Yonas H, Marion DW, et al: Local“inverse steal” induced by hyperventilationin head injury. Neurosurgery 1988; 23:84–88

7. Muizelaar JP, Vanderpoel HG, Li Z, et al: Pialarteriolar diameter and CO2 reactivity duringprolonged hyperventilation in the rabbit.J Neurosurg 1988; 69:923–927

8. Muizelaar JP, Marmarou A, Ward JD, et al:Adverse effects of prolonged hyperventilationin patients with severe head injury: A ran-domized clinical trial. J Neurosurg 1991; 75:731–739

9. Schneider GH, von Helden A, Lanksch WR etal: Continuous monitoring of jugular bulboxygen saturation in comatose patients—Therapeutic implications. Acta Neurochir1995; 134:71–75

10. von Helden A, Schneider GH, Unterberg A, et

al: Monitoring of jugular venous oxygen sat-uration in comatose patients with subarach-noid haemorrhage and intracerebral hema-tomas. Acta Neurochir 1993; 59:102–106

11. Kiening KL, Hartl R, Unterberg AW, et al:Brain tissue pO2 monitoring in comatose pa-tients: Implications for therapy. Neurol Res1997; 19:233–240

12. Aldrich EF, Eisenberg HM, Saydjari C, et al:Diffuse brain swelling in severely head-injured children. Neurosurgery 1992; 76:450–454

13. Lang DA, Teasdale GM, MacPherson P, et al:Diffuse brain swelling after head injury: Moreoften malignant in adults than children?J Neurosurg 1994; 80:675–680

14. Schut L, Bruce DA: Recent advances in thetreatment of head injuries. Pediatr Ann 1976;10:81–104

15. Zwienenberg M, Muizelaar JP: Severe pediat-ric head injury: The role of hyperemia revis-ited. J Neurotrauma 1999; 16:937–943

16. Chiron C, Raynaud C, Maziere B, et al:Changes in regional cerebral blood flow dur-ing brain maturation in children and adoles-cents. J Nucl Med 1992; 33:696–703

17. Adelson PD, Clyde B, Kochanek P, et al: Ce-rebrovascular response in infants and youngchildren following severe traumatic brain in-jury: A preliminary report. Pediatr Neuro-surg 1997; 26:200–207

18. Muizelaar JP, Marmarou A, DeSalles AA, etal: Cerebral blood flow and metabolism inseverely head injured children: Part 1: Rela-tionship with GCS score, outcome, ICP andPVI. J Neurosurg 1989; 71:63–71

19. Stringer WA, Hasso AN, Thompson JR, et al:Hyperventilation-induced cerebral ischemiain patients with acute brain lesions: Demon-stration by Xenon-enhanced CT. AJNR 1993;14:475–484

20. Sharples PM, Stuart AG, Matthews DSF, et al:Cerebral blood flow and metabolism in chil-dren with severe head injury. Part I: Relation toage, Glasgow Coma Score, outcome, intracra-nial pressure, and time after injury. J NeurolNeurosurg Psychiatry 1995; 58:145–152

21. Sharples PM, Matthews DSF, Eyre JA: Cere-bral blood flow and metabolism in childrenwith severe head injuries. Part 2: Cerebrovas-cular resistance and its determinants. J Neu-rol Neurosurg Psychiatry 1995; 58:153–159

22. Skippen P, Seear M, Poskitt K, et al: Effect ofhyperventilation on regional cerebral bloodflow in head-injured children. Crit Care Med1997; 25:1402–1409

23. Peterson B, Khanna S, Fisher B, et al: Pro-longed hypernatremia controls elevated in-tracranial pressure in head-injured pediatricpatients. Crit Care Med 2000; 28:1136–1143

24. Bouma GJ, Muizelaar JP, Choi SC, et al:Cerebral circulation and metabolism after se-vere traumatic brain injury: The elusive roleof ischemia. J Neurosurg 1991; 75:685–693

25. Bouma GJ, Muizelaar JP, Stringer WA, et al:Ultra early evaluation of regional cerebralblood flow in severely head injured patients


using xenon enhanced computed tomogra-phy. J Neurosurg 1992; 77:360–368

26. Cruz J: Low clinical ischemic threshold for cere-bral blood flow in severe acute brain trauma. Casereport. J Neurosurg 1994; 80:143–147

27. Fieschi C, Battistini N, Beduschi A, et al: Regionalcerebral blood flow and intraventricular pressurein acute head injuries. J Neurol Neurosurg Psy-chiatry 1974; 37:1378–1388

28. Jaggi JL, Obrist WD, Gennarelli TA, et al:Relationship of early cerebral blood flow andmetabolism to outcome in acute head injury.J Neurosurg 1990; 72:176–182

29. Marion DW, Darby J, Yonas H: Acute regionalcerebral blood flow changes caused by severehead injuries. J Neurosurg 1991; 74:407–414

30. Robertson CS, Clifton GL, Groosman RG, etal: Alterations in cerebral availability of met-abolic substrates after severe head injury.J Trauma 1988; 28:1523–1532

31. Schroder ML, Muizelaar JP, Kuta AJ: Docu-mented reversal of global ischemia immedi-ately after removal of an acute subdural he-matoma. Neurosurgery 1994; 80:324–327

32. Dahl B, Bergholt B, Cold GE, et al: CO2 andindomethacin vasoreactivity in patients withhead injury. Acta Neurochir 1996; 138:265–273

33. Fortune JB, Feustel PJ, Graca L, et al: Effectof hyperventilation, mannitol, and ventricu-lostomy drainage on cerebral blood flow afterhead injury. J Trauma 1995; 39:1091–1099

34. Sioutos PJ, Orozco JA, Carter LP, et al: Con-tinuous regional cerebral cortical blood flowmonitoring in head injured patients. Neuro-surgery 1995; 36:943–950

35. Crockard HA, Coppel DL, Morrow WF: Eval-uation of hyperventilation in treatment ofhead injuries. BMJ 1973; 4:634–640

36. Obrist WD, Langfitt TW, Jaggi JL, et al: Ce-rebral blood flow and metabolism in coma-tose patients with acute head injury. J Neu-rosurg 1984; 61:241–253

37. Cold GE, Christensen MS, Schmidt K: Effectsof two levels of induced hypocapnia on cere-bral autoregulation in the acute phase ofhead injury coma. Acta Anaesthesiol Scand1981; 25:397–401



Chapter 12. Hyperventilation

1. exp craniocerebral trauma/2. head injur$.tw.3. brain injur$.tw.4. 1 or 2 or 35. brain ischemia/ or “cerebral ischemia”.mp.6. 4 and 57. limit 6 to (newborn infant �birth to 1 month� or infant �1 to 23 months� or preschool child �2 to 5 years� or



Chapter 13. The use of barbiturates in the control of intracranialhypertension in severe pediatric traumatic brain injury

I. RECOMMENDATIONS



C. Options. High-dose barbituratetherapy may be considered in hemody-namically stable patients with salvageablesevere head injury and refractory intra-cranial hypertension.

If high-dose barbiturate therapy isused to treat refractory intracranial hy-pertension, then appropriate hemody-namic monitoring and cardiovascularsupport are essential.

D. Indications from Adult Guidelines.The adult guidelines (1) recommend con-sideration of high-dose barbiturate ther-apy in “hemodynamically stable patientswith salvageable severe TBI and intracra-nial hypertension refractory to maximalmedical and surgical intracranial pres-sure-lowering therapy” as a guideline.

II. OVERVIEW

It is estimated that 21–42% of chil-dren with severe traumatic brain injury(TBI) will develop intractable elevated in-tracranial pressure (ICP) despite medicaland surgical management (3–8). The re-ported mortality rates are 29–100% whenthe ICP is �40 mm Hg despite therapy tolower ICP.

The ICP-reducing and direct neuro-protective properties of barbiturates haveprompted the investigation of two ap-proaches for their use in the manage-ment of patients with severe traumaticbrain injury: a) prophylactic administra-tion early after injury, and b) use in thetreatment of refractory ICP.

III. PROCESS

We searched Medline and Healthstarfrom 1966 to 2001 by using the search

strategy for this question (see AppendixA) and supplemented the results with lit-erature recommended by peers or identi-fied from reference lists. Of 19 potentiallyrelevant studies, two were used as evi-dence for this question (Table 1).


High-dose barbiturates are known toreduce ICP; however, side effects havelimited their use to cases refractory tofirst-line therapies (4, 9, 10). Barbituratesappear to exert their ICP-lowering effectsthrough two distinct mechanisms: sup-pression of metabolism and alterations invascular tone (11–13). Barbiturates canlower resting cerebral metabolic rate foroxygen by about 50% (11). When cerebralblood flow and cerebral blood volume arecoupled to regional metabolic rate, theyare also decreased. This mechanism me-diates the observed beneficial effects ofbarbiturates on ICP and cerebral perfu-sion pressure. However, Cruz (14) re-ported that some patients treated for in-tractable ICP with barbiturate comadeveloped jugular venous oxygen satura-tion levels �45%, which was associatedwith a significantly worse outcome com-pared with patients with higher jugularvenous oxygen saturations. This sug-gested that in some patients, barbituratecoma induced oligemic hypoxia. Cruz in-cluded teenagers and adults; however, theresults of the teenagers were not sepa-rately reported. Barbiturates confer addi-tional direct neuroprotective effects inde-pendent of their ICP-lowering properties,such as inhibition of free radical-medi-ated lipid peroxidation or membrane sta-bilization (11).

Few studies have evaluated barbitu-rate pharmacokinetics and pharmacody-namics in children with head injury (15–18). Clearance appears to vary widely andmay be increased with duration of ther-apy (18). Barbiturate serum levels arepoorly correlated with electrical activity

(16, 17). Monitoring of electroencephalo-graphic patterns for burst suppression isthought to be more reflective of thera-peutic effect than measuring serum druglevels (11, 13). Near-maximum reductionin cerebral metabolism and cerebralblood flow occurs when burst suppres-sion is induced (9, 13).

Barbiturates suppress metabolism; how-ever, there is insufficient information aboutcomparative efficacy to recommend onebarbiturate over another, except in relationto their particular pharmacologic proper-ties. The use of both pentobarbital and thio-pental has been reported.

Prophylactic Use of Barbiturates

There are no published studies of pro-phylactic barbiturate use in children withsevere TBI. The “Guidelines for the Man-agement of [Adult] Severe Traumatic BrainInjury” (1) reported on two randomizedclinical trials that examined early prophy-lactic administration of barbiturates. Nei-ther study demonstrated clinical benefit(19, 20). Schwartz et al. (20) did not definethe lower age limits in their study, but themean patient age suggests that childrenwere included. Ward et al. (21) includedadolescents over the age of 12 yrs; however,they did not separately report the effects ofthe barbiturate therapy among the chil-dren. Ward et al. (21) reported that 54% ofbarbiturate-treated subjects developed hy-potension—defined as a systolic bloodpressure �80 mm Hg—compared with 7%of controls.

Refractory IntracranialHypertension

Use of barbiturates to treat elevated ICPin children with severe head injury hasbeen reported since the 1970s (22). Mar-shall et al. (22a) were the first to report thatboth control of ICP and outcome were im-proved with the use of barbiturates. Patientage was not specified in this report. In thiscase series, 25 patients with ICP �40 mm


Hg were treated with high-dose pentobar-bital. When ICP was controlled, mortalityrate was significantly reduced comparedwith patients with persistently elevated ICP(21% vs. 83%).

Kasoff et al. (4) reported a case series of25 children with severe TBI. ICP was mon-itored in all patients, and surgically correct-able lesions were treated with immediateoperation. Pentobarbital was administeredif ICP remained �20 mm Hg despite hy-perventilation to PaCO2 25–30 torr and ad-ministration of dexamethasone and manni-tol. Each patient (n � 11) who receivedpentobarbital was monitored with a pulmo-nary artery catheter. The clinical goals wereto maintain ICP �20 mm Hg, cerebral per-fusion pressure (CPP) �40 mm Hg, andhemodynamic stability. Ten of 11 children(91%) who received barbiturates requireddopamine to maintain CPP compared with11% of children who did not receive barbi-turates. The authors state that all childrenwho received barbiturates had diminishedcardiac output and systemic vascular resis-tance. Nine of the children experienced hy-potensive episodes despite intensive moni-toring, fluid resuscitation, and dopamineinfusions. Thirty-seven percent of the pa-tients died. The specific effects of barbitu-rates on ICP and CPP were not reported.

Pittman et al. (23) reported a caseseries of 27 children with severe TBI whoreceived pentobarbital for ICP �30 mmHg if the intracranial hypertension failedto respond to other treatment modalities.Fourteen (52%) achieved ICP �20 mmHg after addition of barbiturates. Six(22%) died within 48 hrs despite therapy,and seven had sustained elevation of ICP�35 mm Hg and reduction of CPP (�50mm Hg) for several hours. No conclu-sions can be drawn from this study re-garding the effect of pentobarbital-relatedreduction of intracranial hypertension onneurologic outcome. Three of seven pa-tients with sustained elevation of ICPmade good recoveries. The authors sug-gested that barbiturates may have a ben-

eficial effect on outcome even when re-fractory ICP is not controlled.


Eisenberg et al. (24) reported a multi-ple-center randomized clinical trial ofhigh-dose barbiturates in severely head-injured patients with intractable ICP ele-vations. Patients were between the agesof 15 and 50 yrs. Information on childrenwas not reported separately. This study isconsidered the best evidence for the useof high-dose barbiturates in adults withuncontrolled ICP. It is the primary studyon which the adult guidelines for use ofhigh-dose barbiturates are based (1).

Patients were randomly assigned tobarbiturate therapy, whereas the controlsubjects continued to receive conven-tional therapies of hyperventilation, mus-cle relaxation, sedation, mannitol, andventricular drainage (when possible).Successful control of ICP was the primaryoutcome variable. Patients in the controlgroup could cross over to the barbituratetreatment group. Thirty-two percent ofpatients randomized to barbiturate ther-apy had control of ICP. ICP control wasalmost twice as likely to be achieved inbarbiturate-treated patients comparedwith the conventional treatment group.The likelihood of survival among barbi-turate responders at 1 month after injurywas 92% compared with 17% amongnonresponders. The primary cardiovascu-lar complication was hypotension.

Therapeutic Regimens

A number of therapeutic regimens havebeen reported. Eisenberg et al. (24) usedthe following protocol for pentobarbital.

Loading dose: 10 mg/kg over 30 mins

Then 5 mg/kg every hour for three doses

Maintenance: 1 mg·kg�1·hr�1

Nordby and Nesbakken (25) reportedon the use of thiopental in children andadults with severe TBI and used the fol-lowing dosing regimen.

Loading dose 10–20 mg/kg

Maintenance: 3–5 ·kg�1·hr�1

Doses of thiopental were reduced ifblood pressure decreased or ICP was�25 mm Hg.

Although the duration and optimalmethod to discontinue high-dose barbitu-rate administration have not been studied,often clinicians seek a period of 24 hrsduring which there is good ICP control andno dangerous elevations before beginningto taper off the barbiturate infusion (26).

V. SUMMARY

Small studies of high-dose barbituratetherapy suggest that barbiturates are ef-fective in lowering ICP in selected casesof refractory intracranial hypertension inchildren with severe TBI. However, stud-ies on the effect of barbiturate therapy foruncontrolled ICP have not evaluated neu-rologic outcome. Use of barbiturates isassociated with myocardial depression,increased risk of hypotension, and needfor blood pressure support with intravas-cular fluids and inotropic infusions.Studies have not evaluated the effect ofage on the risk of hemodynamic compro-mise during high-dose barbiturate ther-apy. The potential complications of high-dose barbiturate therapy in infants andchildren with severe TBI mandate that itsuse be limited to critical care providersand that appropriate systemic monitoringbe used to avoid and rapidly treat hemo-dynamic instability.

There is no evidence to support use ofbarbiturates for the prophylactic neuro-protective effects or prevention of the de-velopment of intracranial hypertension inchildren with severe TBI.



Kasoff et al. (4), 1988 Case series of 21 children with severe TBI; 11 treatedwith pentobarbital for intractable ICP. Invasivehemodynamic monitoring used.

III Children receiving high-dose barbiturates had decreasedcardiac index and lower systemic vascular resistance; 91%required dopamine to maintain hemodynamic stability.

Pittman et al. (23), 1989 Case series of 27 children who received pentobarbitalfor ICP �20 mm Hg despite conventional care.

III Fourteen of 27 achieved ICP �20 mm Hg with addition ofpentobarbital. Seven of 27 experienced persistently elevatedICP, and three of those seven made good ultimate recovery.

TBI, traumatic brain injury; ICP, intracranial pressure.



● Clearly, a study of high-dose barbitu-rates for intractable ICP after severeTBI in children is needed to deter-mine whether they improve out-come.

● The effect of barbiturate therapy incases of diffuse cerebral swelling inchildren should be evaluated. Fur-thermore, no studies have reportedthe efficacy of high-dose barbituratesfor intractable ICP in infants or afterinjury due to abusive head trauma.

● Age dependence of the deleterioushemodynamic effects of barbituratesdeserves further study. Prolonged in-hibition of synaptic activity in thedeveloping brain during infancy hasbeen shown to have deleterious ef-fects in recent studies in laboratorymodels of brain injury (27).

● The effects of barbiturates in infantswith severe TBI requires study.

REFERENCES

1. Bullock R, Chesnut RM, Clifton G, et al:Guidelines for the management of severe

traumatic brain injury. J Neurotrauma 2000;17:451–553

2. Deleted in proof3. Cordobes F, Lobato RD, Rivas JJ, et al: Post-

traumatic diffuse brain swelling: Isolated orassociated with cerebral axonal injury: Clin-ical course and intracranial pressure in 18children. Childs Nerv Syst 1987; 3:235–238

4. Kasoff SS, Lansen TA, Holder D, et al: Ag-gressive physiologic monitoring of pediatrictrauma patients with elevated intracranialpressure. Pediatr Neurosci 1988; 14:241–249

5. Bruce DA, Schut L, Bruno LA, et al: Outcomefollowing severe head injuries in children.J Neurosurg 1978; 48:679–688

6. Edger HG, Legat JA, Gruber W: Traumaticbrain stem lesions in children. Childs NervSyst 2000; 16:21–24

7. Esparza J, Portillo JM, Sarabia M, et al: Out-come in children with severe head injuries.Childs Nerv Syst 1985; 1:109–114

8. Berger MS, Pitts LH, Lovely M, et al: Out-come from severe head injury in childrenand adolescents. J Neurosurg 1985; 62:194–199

9. Wilberger JE, Cantella D: High-dose barbitu-rates for intracranial pressure control. NewHoriz 1995; 3:469–473

10. Pfenninger J: Neurological intensive care inchildren. Intensive Care Med 1993; 19:243–250

11. Piatt JH, Schiff SJ: High dose barbituratetherapy in neurosurgery and intensive care.Neurosurgery 1984; 15:427–444

12. Demopoulous HB, Flamm ES, PietronigroDD, et al: The free radical pathology and themicrocirculation in the major central ner-vous system. Acta Physiol Scand 1980;492(Suppl):91–119

13. Kassell NF, Hitchon PW, Gerk MK et al:Alterations in cerebral blood flow, oxygenmetabolism, and electrical activity producedby high-dose thiopental. Neurosurgery 1980;7:598–603

14. Cruz J: Adverse effects of pentobarbital oncerebral oxygenation of comatose patientswith acute traumatic brain swelling: Rela-tionship to outcome. J Neurosurg 1996; 85:758–761

15. Russo H, Bressolle F, Duboin MP: Pharma-cokinetics of high-dose thiopental in pediat-ric patients with increased intracranial pres-sure. Ther Drug Monit 1997; 19:63–70

16. Schikendantz J, Funk W, Ittner KP, et al:

Elimination of methohexitone after long-term, high-dose infusion in patients withcritically elevated intracranial pressure. CritCare Med 1999; 27:1570–1576

17. Turcant A, Delhumeau A, Premel-Cabic A, etal: Thiopental pharmacokinetics under con-dition of long-term infusion. Anesthesiology1985; 63:50–54

18. Russo H, Bres J, Duboin MP, et al: Variabilityof thiopental clearance in routine criticalcare patients. Eur J Clin Pharmacol 1995;48:479–487

19. Deleted in proof20. Schwartz ML, Tator CH, Rowed DW, et al:

The University of Toronto head injury treat-ment study: A prospective randomized com-parison of pentobarbital and mannitol. CanJ Neurol Sci 1984; 11:434–440

21. Ward JD, Becker DP, Miller JD, et al: Failureof prophylactic barbiturate coma in the treat-ment of severe head injury. J Neurosurg1985; 62:383–388

22a. Marshall LF, Smith RW, Shapiro HM: Theoutcome with aggressive treatment insevere head injuries: Acute and chronicbarbiturate administration in the manage-ment of head injury. J Neurosurg 1979;50:26 –30

22. Bruce DA, Raphaely RC, Goldberg AI, et al:Pathophysiology, treatment and outcomefollowing severe head injury in children.Childs Brain 1979; 5:174–191

23. Pittman T, Bucholz R, Williams D: Efficacy ofbarbiturates in the treatment of resistant in-tracranial hypertension in severely head-injured children. Pediatr Neurosci 1989; 15:13–17

24. Eisenberg HM, Frankowski RF, Contant CF,et al: High-dose barbiturate control of ele-vated intracranial pressure in patients withsevere head injury. J Neurosurg 1988; 69:15–23

25. Nordby HK, Nesbakken R: The effect of highdose barbiturate decompression after severehead injury. A controlled clinical trial. ActaNeurochir 1984; 72:157–166

26. Raphaely RC, Swedlow DB, Downes JJ, etal: Management of severe pediatric headtrauma. Pediatr Clin N Am 1980; 27:715–727

27. Ikonomidou C, Bittigau P, Koch C, et al:Neurotransmitters and apoptosis in the de-veloping brain. Biochem Pharmacol 2001;62:401–405

S mall studies of

high-dose barbitu-

rate therapy suggest

that barbiturates are effective

in lowering intracranial pres-

sure in selected cases of re-

fractory intracranial hyper-

tension in children with

severe traumatic brain injury.





Chapter 13. Barbituates

1. exp craniocerebral trauma/2. head injur$.tw.3. brain injur$.tw.4. 1 or 2 or 35. intracranial pressure/or “intracranial pressure” .mp.6. intracranial hypertension/or “intracranial hypertension” .mp.7. 5 or 68. 4 and 79. limit 8 to (newborn infant �birth to 1 month� or infant �1 to 23 months� or preschool child �2 to 5 years� or



Chapter 14. The role of temperature control following severepediatric traumatic brain injury

I. RECOMMENDATIONS



C. Options. Extrapolated from theadult data, hyperthermia should beavoided in children with severe traumaticbrain injury (TBI).

Despite the lack of clinical data inchildren, hypothermia may be consideredin the setting of refractory intracranialhypertension.

II. OVERVIEW

Posttraumatic hyperthermia is classi-fied as a core body temperature �38.5°C,whereas hypothermia is classified as tem-perature �35°C. At present, the data inthe basic science literature on adult ani-mal models indicate that hyperthermiacontributes to greater posttraumaticdamage by increasing the acute patho-physiologic response following injury,through a multitude of mechanisms. Therationale for avoidance of hyperthermiaand for use of therapeutic hypothermia isto lessen the effect that temperature mayhave on these mechanisms of secondaryinjury by decreasing cerebral metabo-lism, inflammation, lipid peroxidation,excitotoxicity, cell death, and acute sei-zures. Based on experimental studies inanimal models and clinical studies inadults (1) in which hyperthermia wascorrelated with poor outcome, it has beenrecommended that hyperthermia follow-ing TBI in children should be avoided.There also may be a role for therapeutichypothermia in reducing intracranial hy-pertension in severe pediatric TBI. Evi-dence in both the pediatric and adultliterature is evaluated.

III. PROCESS



There was one retrospective study fromthe 1950s by Hendrick (2) indicating thatmoderate hypothermia (32–33°C) was ef-fective in the treatment of children follow-ing severe TBI. This initial investigationwas of 18 children with severe TBI (Glas-gow Coma Scale score � 4) who presentedwith decerebrate posturing. There were tenlong-term survivors with only one severelyimpaired. Hendrick concluded that sys-temic cooling following injury was a “use-ful adjunct” and could improve outcome inchildren after TBI. Since that time, therehas been a lack of subsequent randomizedor other trials to further evaluate this pre-liminary finding.

In 1973, Gruszhiewicz et al. (3) con-ducted a prospective, randomized studyof 20 children �16 yrs of age who suf-fered a severe TBI, presenting with a clin-ical exam of decerebrate rigidity (GlasgowComa Scale � 4). The children were ran-domized to one of two groups: hypother-mia vs. hypothermia combined withdexamethasone (2 mg twice daily). Therewas no normothermic group. Nineteen ofthese 20 patients were hyperthermic atpresentation and suffered various mech-anisms of injury. Outcome was deter-mined by duration of coma and time until“recovery,” although the length of fol-low-up was �7 months in all instances.Although no statistical analysis was per-formed, the authors described similar du-ration of coma and neurologic recoveryfor the two groups. Surprisingly, 19 pa-tients survived.

Since 1973, no further studies haveevaluated the specific efficacy of hypo-thermia following head injury fromwhich results could be gleaned for pedi-atric patients. No other studies comparedtemperature control (e.g., hypothermiawith normothermia or hyperthermia) asit relates to outcome. In all other studies,either only adults were studied, or resultsfor children and adults were so con-founded that no conclusions can bedrawn specifically for pediatric cases.Thus, only two studies met the criteriafor inclusion in this chapter.


There was no section on temperatureregulation in the adult guidelines (4) andthere has not been an evidence report onthis topic in adults. However, a numberof studies have assessed the efficacy oftherapeutic hypothermia following severeTBI in adults.

The induction of hypothermia clini-cally to treat patients with TBI was orig-inally reported �50 yrs ago (2), but use oftherapeutic hypothermia did not becomeestablished because early studies lackedmodern scientific methods and adequateoutcome measures. Renewed interest inmoderate hypothermia after severe TBIdid not occur until the early 1990s, whenpreliminary data from single-center clin-ical trials were published in adults.Shiozaki et al. (5) used therapeutic mod-erate hypothermia to 34°C for �2 days ina group of severe TBI patients who hadintracranial hypertension and were re-fractory to barbiturate therapy. Theyfound an improvement in cerebral perfu-sion pressure, compared with normo-thermic patients, which was sustainedduring and after rewarming. Marion et al.(6) reported that moderate hypothermiain adult patients with severe TBI reducedintracranial pressure and showed a trend


toward improved outcome at both 3 and 6months after injury. Clifton et al. (7)cooled patients to 32–33°C after severeTBI for 48 hrs and similarly reported atrend toward improved outcome. Marionet al. (8) later demonstrated that moder-ate hypothermia for 24 hrs specificallyhastened neurologic recovery in patientswho presented with a Glasgow ComaScale score of 5–7 and that the treatedpatients tended to have improved overalloutcome. Although these single-centerstudies provided evidence of efficacy,Clifton et al. (9) more recently reportedlack of effectiveness in adults in a multi-ple-center clinical trial of moderate hypo-thermia following severe TBI. Despitefailure to replicate the earlier single-center findings in the larger multiple-center trial, there was a suggestion ofimproved outcome in those patients whopresented as hypothermic and were thenkept cool and in the younger age groupswithin the study (�40 yrs of age). Chil-dren (�16 yrs) were not included in theClifton et al. (9) study.

V. SUMMARY

There is presently no published sup-port for temperature control or therapeutic

hypothermia in pediatric TBI. Based onstudies in adults, therapeutic optionsinclude the avoidance of hyperthermiaand the consideration of hypothermiafor refractory intracranial hyperten-sion.


● The effect of temperature control onoutcome following pediatric TBI needsto be studied.

● The role of therapeutic hypothermia,both as a neuroprotective measure andfor refractory intracranial hyperten-sion, deserves investigation in pediatricTBI. Direct comparisons to other ther-apies should be conducted.

● Evaluations of therapeutic hypother-mia should be age stratified. Additionaldocumentation of the effect of hypo-thermia and temperature regulation instudies restricted to infants and chil-dren is needed.

● In addition, studies will be needed tobetter understand the effect of temper-ature regulation on other physiologicvariables (e.g., intracranial pressure,cerebral perfusion pressure, cardiac

output, and immune status) and howthis might affect long-term outcome.

REFERENCES

1. Jones PA, Andrews PJ, Midgley S, et al: Mea-suring the burden of secondary insults in headinjured patients during intensive care. J Neu-rosurg Anesthesiol 1994; 6:4–14

2. Hendrick EB: The use of hypothermia in se-vere head injuries in childhood. AMA ArchSurg 1959; 17–20

3. Gruszhiewicz J, Doron Y, Peyser E: Recoveryfrom severe craniocerebral injury with brainstem lesions in childhood. Surg Neurol 1973;1:197–201


5. Shiozaki T, Hisashi S, Taneda M, et al: Effectof mild hypothermia on uncontrollable intra-cranial hypertension after severe head injury.J Neurosurg 1993; 79:363–368

6. Marion DW, Obrist WD, Carlier PM, et al: Theuse of moderate therapeutic hypothermia forpatients with severe head injuries: A prelimi-nary report. J Neurosurg 1993; 79:354–362

7. Clifton GL, Allen S, Barrodale P, et al: A phaseII study of moderate hypothermia in severebrain injury. J Neurotrauma 1993; 10:263–271

8. Marion DW, Penrod LE, Kelsey SF, et al:Treatment of traumatic brain injury withmoderate hypothermia. New Engl J Med 1997;336:540–546

9. Clifton GL, Miller ER, Choi SC, et al: Lack ofeffect of induction of hypothermia after acutebrain injury. New Engl J Med 2001; 344:556–563



Gruszhiewicz et al. (3),1973

Uncontrolled case series of 20patients treated withhypothermia vs. hypothermiaplus glucocorticoids/steroids(dexamethasone, 2 mg twicedaily).

Outcome was determined byduration of coma and timeto recovery.

No long-term outcome cited.

III No difference in outcomewith the addition ofsteroids to hypothermia.

No comparisons oranalysis with relation tocontrolled ornormothermic patients.

Hendrick (2), 1959 Uncontrolled retrospective caseseries of 18 children with asevere TBI who presentedwith decerebrate posturingand were cooled to 32–33°C.

III Hypothermia is a usefuladjunct with thepotential for improvedoutcome in childrenwith severe TBI.

TBI, traumatic brain injury.

T here is presently

no published sup-

port for tempera-

ture control or therapeutic

hypothermia in pediatric

traumatic brain injury.





Chapter 14. Temperature Control

1. exp craniocerebral trauma/2. head injur$.tw.3. brain injur$.tw.4. 1 or 2 or 35. hypothermia, induced/6. 4 and 57. limit 6 to (newborn infant �birth to 1 month� or infant �1 to 23 months� or preschool child �2 to 5 years� or



Chapter 15. Surgical treatment of pediatric intracranialhypertension

I. RECOMMENDATIONS



C. Options. Decompressive craniec-tomy should be considered in pediatricpatients with severe traumatic brain in-jury (TBI), diffuse cerebral swelling, andintracranial hypertension refractory tointensive medical management.

Decompressive craniectomy shouldbe considered in the treatment of severeTBI and medically refractory intracranialhypertension in infants and young chil-dren with abusive head trauma.

Decompressive craniectomy may beparticularly appropriate in childrenwith severe TBI and refractory intracra-nial hypertension who have a poten-tially recoverable brain injury. Decom-pressive craniectomy appears to be lesseffective in patients who have experi-enced extensive secondary brain in-sults. Patients who experience a sec-ondary deterioration on the GlasgowComa Scale (GCS) and/or evolving ce-rebral herniation syndrome within thefirst 48 hrs after injury may represent afavorable group. Patients with an unim-proved GCS of 3 may represent an un-favorable group.

II. OVERVIEW

The Traumatic Coma Data Bank hasestablished the poor prognosis (34%mortality rate, 16% good or moderatelydisabled) of pediatric and adult patientswith severe TBI and diffuse cerebral in-jury on computed tomography (CT)scan (compressed cisterns, �5 mmmidline shift, mass lesion �25 mL) (2,3). Because maximum postinjury intra-cranial pressure (ICP) is a leading pre-dictor of outcome in severe TBI, somehave advocated the use of decompres-

sive craniectomy to treat medically re-fractory intracranial hypertension inchildren (4 –7).

The main objective of decompressivecraniectomy is to control ICP and thusmaintain cerebral perfusion pressureand cerebral oxygenation, as well asprevent herniation, in the face of re-fractory cerebral swelling. There are anumber of surgical interventions forthe treatment of refractory intracranialhypertension. This chapter addressesonly the use of decompressive craniec-tomy. Four questions regarding the useof decompressive craniectomy in chil-dren are evaluated:

1. Is decompressive craniectomy suc-cessful in controlling ICP?

2. Does decompressive craniectomyimprove clinical outcomes?

3. What surgical technique is appro-priate?

4. Which patients are appropriate candi-dates for decompressive craniectomy?

III. PROCESS

We searched Medline and Healthstarfrom 1966 to 2001 by using the searchstrategy for this question (see Appendix A)and supplemented the results with litera-ture recommended by peers or identifiedfrom reference lists. Of 21 potentially rele-vant studies, three were used as evidencefor this question (Table 1).


Is Decompressive CraniectomySuccessful in Lowering ICP?

The measured value of ICP may beartifactually altered due to the cranialdefect in patients who have undergonedecompressive craniectomy. However,given that this surgical procedure isgenerally undertaken with the goal ofcontrolling severe refractory intracra-

nial hypertension, its effect on ICP is ofinterest. Taylor and colleagues (8) re-ported a significant reduction in meanICP after decompressive craniectomyfor severe TBI in children (averagemean decrease, 9 mm Hg). Hieu andcolleagues (9) illustrated graphicallythe intraoperative and immediate post-operative ICP at the time of decompres-sive craniectomy in two pediatric pa-tients. Sequential decreases in ICP seenat the point of craniectomy and of du-raplasty were sustained in the immedi-ate postoperative period.

Cho and colleagues (10) reported sig-nificantly decreased ICP after decompres-sive craniectomy (preoperative mean, 59mm Hg; postoperative mean, 12 mm Hg)in 10 children �2 yrs of age with severeTBI and refractory intracranial hyperten-sion from abusive head trauma.

Key Elements from AdultEvidence Relevant to PediatricTBI

Polin et al. (4) documented a statisti-cally significant decrease in ICP from 32to 21 mm Hg after decompressive crani-ectomy in 26 pediatric and adult patients.In their study, ICP after craniectomy wasalso lower than ICP at an equivalentpostinjury interval in a matched controlgroup taken from the Traumatic ComaData Bank. Kunze and colleagues (6) re-ported decreased ICP (mean preoperative,42 mm Hg; mean postoperative, 21 mmHg) and increased cerebral perfusionpressure in 28 children and adults whounderwent decompressive craniectomy totreat severe TBI. Gaab and colleagues (7)illustrated graphically a single “represen-tative” case from their study of 37 pedi-atric and adult severe TBI patients show-ing immediate decrease in ICP andincrease in GCS after decompressivecraniectomy.


Does DecompressiveCraniectomy Improve ClinicalOutcomes?

Three class III studies evaluated out-come after decompressive craniectomyfor the treatment of severe TBI in chil-dren. Taylor and colleagues (8) per-formed a single-center, prospective, ran-domized clinical trial of decompressivecraniectomy in the treatment of pediatricpatients (age 1–18) with severe TBI andrefractory intracranial hypertension (n �27). These patients were randomized toreceive maximal medical therapy andventricular drainage alone or in additionto decompressive bitemporal craniec-tomy. Patients in this study who under-went craniectomy had a trend toward bet-ter clinical outcome at 6 months afterinjury (modified Glasgow Outcome Scale[GOS]). Although this study is a prospec-tive, randomized clinical trial, concernsabout size and generalizability of thesample limited the evidence to class IIIwith respect to outcome.

Polin et al. (4) reported the outcomesof 35 patients with mean age 18 yrs whounderwent bifrontal decompressivecraniectomy for severe TBI and medicallyrefractory intracranial hypertension. Fa-vorable outcome (GOS at discharge fromrehabilitation) was more frequent in pe-diatric (44%) than adult (29%) patients.There was no concurrent control groupin this study. However, the authorsmatched control patients from the Trau-matic Coma Data Bank to each studypatient. They reported a significantlyhigher rate of favorable outcome in pedi-atric patients undergoing decompressivecraniectomy vs. controls, based on a uni-variate analysis. This beneficial effect was

also evident in a multivariate analysis re-stricted to pediatric patients operated onwithin 48 hrs of injury and without sus-tained ICP elevation beyond 40 mm Hg.Hieu and colleagues (9) reported goodneurologic recovery in two 8-yr-old pa-tients who underwent decompressivecraniectomy within 12 hrs of TBI becauseof severe intracranial hypertension andevolving transtentorial herniation syn-drome. Their operative procedure also in-cluded the resection of severely contusedbrain.

Cho and colleagues (10) reported out-comes in 23 children �2 yrs of agetreated with medical therapy or medicaltherapy plus decompressive craniectomyfor severe TBI due to abusive headtrauma. In this prospective, single-center, case control study, ten patientswith severe intracranial hypertension(�30 mm Hg) received medical ICP man-agement plus decompressive craniec-tomy, whereas seven patients with severeintracranial hypertension and six patientswith less severe intracranial hypertension(�30 mm Hg) were treated with medicaltherapy alone. Patients with severe intra-cranial hypertension managed with sur-gery, and patients with less severe intra-cranial hypertension, had Child OutcomeScores significantly higher than patientswith severe intracranial hypertensionmanaged medically. Among children withsevere intracranial hypertension, survivalwas also significantly improved in thosechildren undergoing decompressive sur-gery. Decompressive craniectomy in thisstudy was generally performed within 24hrs of injury. A mean of 32 mL of sub-dural hematoma was also evacuated atthe time of decompressive surgery.

Key Elements from AdultScientific Literature Relevant toPediatric TBI

Venes and Collins (11) reported a se-ries of 13 severe TBI patients who under-went decompressive craniectomy, includ-ing six children. Although the authorssuggested that survival in this retrospec-tive, uncontrolled study was increasedrelative to historical experience, only twopatients, including one child, made a sig-nificant neurologic recovery. Kunze et al.(6) performed decompressive craniec-tomy on 28 severe TBI patients with re-fractory intracranial hypertension andmean age of 22 (range, 8–44). Althoughstatistical analysis was not performed inthis uncontrolled study, the descriptivedata suggest markedly better GOS in pa-tients �30 yrs of age. Guerra et al. (5)prospectively studied the results of de-compressive craniectomy in 57 pediatricand adult patients with severe TBI andmedically refractory intracranial hyper-tension. These authors excluded patientswith CT demonstration of severe brain-stem injury, absent brainstem auditoryevoked responses, absence of oscillatorycerebral blood flow on transcranial Dopp-ler ultrasound, initial GCS of 3 withoutimprovement, or bilateral fixed and di-lated pupils. Using these exclusion crite-ria, the authors estimated that only 3% ofseverely TBI patients presenting to theirinstitution over a 20-yr period were en-tered into the decompression protocol.Fifty-eight percent of this highly re-stricted patient population, however,achieved a GOS of 4 or 5. Logistic regres-sion analysis in this study failed to sup-port young age as a predictor of improvedoutcome.



Polin et al. (4),1997

In a single-center, case-controlled study, 35 severely head-injured patients underwent decompressive craniectomy withpre- and postoperative ICP monitoring and medicalmanagement.

III A significantly increased rate of favorable outcome was seenin surgical patients compared with matched controls.Young age, early operation, and avoidance of ICP�40 mm Hg may improve outcome.

Cho et al. (10),1995

In a single-center, case-controlled study, 23 severely head-injured children with shaking-impact syndrome underwenteither decompressive craniotomy or medical management.

III Of patients with severe (�30 mm Hg) intracranialhypertension, those undergoing surgery had improvedsurvival and neurological outcomes compared with thoseundergoing medical therapy alone.

Taylor et al. (8),2001

In a single-center PRCT, 27 severely head-injured childrenwith intracranial hypertension refractory to medicalmanagement and ventricular drainage were randomized tobitemporal decompressive craniectomy vs. no surgery.

III Decompressive craniectomy significantly lowered mean ICPin the 48 hrs after randomization and resulted in amarginally nonsignificant trend toward improved clinicaloutcome at 6 mos.

ICP, intracranial pressure; PRCT, prospective, randomized controlled trial.


What Surgical Technique IsAppropriate?

Studies from the CT imaging era havegenerally recommended unilateral fron-tal-temporal-parietal decompressivecraniectomy for unilateral cerebral swell-ing or bilateral frontal craniectomy forbilateral cerebral swelling in both chil-dren and adults (5–7). The historical lit-erature is cited as a caution against smallcraniectomies, due to the potential forinadequate relief of intracranial hyper-tension and for cerebral incarcerationand infarction (5, 12). However, one pro-spective study of decompressive craniec-tomy in pediatric patients demonstrated atrend toward improved outcome after4-cm bitemporal craniectomies (8). Mostauthors describe a combined craniectomyand expansion duraplasty (4–7, 9, 10).Bilateral procedures used by various au-thors include separate bilateral craniec-tomies with a strip of intact bone over thesagittal sinus (5, 7) vs. bifrontal craniec-tomy with section of the anterior falxcerebri at the skull base (4, 11, 13). Nostudies have evaluated the differential ef-ficacy of these various techniques.

Which Patients Are AppropriateCandidates for DecompressiveCraniectomy?

Three studies of outcome in pediatricpatients suggest specific criteria for theperformance of decompressive craniec-tomy. After conducting logistic regres-sion analysis of 35 severe TBI patientstreated in their institution, Polin and col-leagues (4) recommended decompressivecraniectomy for pediatric patients withcerebral swelling and medically refrac-tory intracranial hypertension who arewithin 48 hrs of injury and who have notexperienced a sustained ICP elevation�40 mm Hg. They also recommendedagainst decompressive craniectomy forpatients with initial and sustained GCS of3. Taylor et al. (8) recommended decom-pressive craniectomy for pediatric pa-tients with refractory intracranial hyper-tension (ICP 20 –24 mm Hg for �30mins, 25–29 for �10 mins, �30 for �1min) or cerebral herniation syndrome onthe first day after injury, despite ventric-ular drainage.

Cho and colleagues (1995) suggestedthe use of decompressive craniectomywithin 24 hrs of injury in children �2 yrsof age with severe TBI and medically re-

fractory intracranial hypertension (�30mm Hg) from nonaccidental trauma.

Key Elements from AdultEvidence Relevant to PediatricTBI

Guerra and colleagues (5) recom-mended decompressive craniectomy forpediatric and adult patients with severeTBI, cerebral swelling on CT imaging,refractory intracranial hypertension, andwitnessed deterioration in clinical vari-ables (GCS, neurologic examination),electrophysiological variables (electroen-cephalogram, somatosensory evoked po-tentials), and/or transcranial Doppler ul-trasound variables (increased pulsatility,decrease in diastolic flow). These authorsexcluded patients with CT imaging dem-onstration of severe brainstem injury, ab-sent brainstem auditory evoked re-sponses, absence of oscillatory cerebralblood flow on transcranial Doppler ultra-sound, initial GCS of 3 without improve-ment, and bilateral fixed and dilated pu-pils.

V. SUMMARY

Decompressive craniectomy for severeTBI and medically refractory intracranialhypertension in children lowers ICP andmay improve outcome. Decompressivecraniectomy also may be appropriate inyoung children with severe TBI and re-fractory intracranial hypertension fromabusive head trauma. Insufficient evi-dence is available to evaluate the efficacyof various described surgical techniquesfor decompressive craniectomy. Decom-pressive craniectomy for children withsevere TBI and refractory intracranial hy-pertension may be most appropriate inpatients meeting some or all of the fol-lowing criteria:

1. Diffuse cerebral swelling on cranialCT imaging

2. Within 48 hrs of injury

3. No episodes of sustained ICP �40mm Hg before surgery

4. GCS �3 at some point subsequentto injury

5. Secondary clinical deterioration

6. Evolving cerebral herniation syn-drome


● Randomized controlled trials of thesafety and efficacy of decompressive

craniectomy in severe pediatric TBIshould be undertaken. The onlystudy of this type may have failed todemonstrate a statistically signifi-cant benefit of decompression onlong-term outcome due to smallsample size.

● It may be useful for future studies tocompare decompressive craniectomyto other “second-tier” interventionsfor severe, refractory intracranial hy-pertension, such as barbiturate ther-apy, hypothermia, or lumbar CSFdrainage.

● The safety and efficacy of decompres-sive craniectomy for severe TBI ininfants and for severe TBI due toabusive head trauma should be fur-ther studied.

● Studies of decompressive craniec-tomy for severe TBI should includecareful monitoring of ICP, cerebralperfusion pressure, cerebral bloodflow, and other important physio-logic variables to correlate alter-ations in the latter variables withsuccessful clinical outcome. Suchdata may clarify the pathophysiolog-ical variables involved and providebetter information about the indica-tions for and appropriate timing ofdecompressive surgery.

● Studies are needed to evaluate theoptimal surgical approach to decom-pressive craniectomy.

REFERENCES

1. Deleted in proof2. Marshall LF, Gautille T, Klauber MR, et al:

D ecompressive

craniectomy for

severe traumatic

brain injury and medically

refractory intracranial hy-

pertension in children lowers

intracranial pressure and

may improve outcome.


The outcome of severe closed head injury.J Neurosurg 1991; 75:S28–S36

3. Levin HS, Aldrich EF, Saydjari C, et al: Se-vere head injury in children: Experience ofthe Traumatic Coma Data Bank. Neurosur-gery 1992; 31:435–443

4. Polin RS, Shaffrey ME, Bogaev CA, et al:Decompressive bifrontal craniectomy in thetreatment of severe refractory posttraumaticcerebral edema. Neurosurgery 1997; 41:84–94

5. Guerra WKW, Gaab MR, Dietz H, et al: Sur-gical decompression for traumatic brainswelling: Indications and results. J Neuro-surg 1999; 90:187–196

6. Kunze E, Meixensberger J, Janka M, et al:

Decompressive craniectomy in patients withuncontrollable intracranial hypertension.Acta Neurochir 1998; 71:16–18

7. Gaab MR, Rittierodt M, Lorenz M, et al: Trau-matic brain swelling and operative decom-pression: A prospective investigation. ActaNeurochir Suppl 1990; 51:326–328

8. Taylor A, Warwick B, Rosenfeld J, et al: Arandomized trial of very early decompressivecraniectomy in children with traumaticbrain injury and sustained intracranial hy-pertension. Childs Nerv Syst 2001; 17:154–162

9. Hieu PD, Sizun J, Person H, et al: The placeof decompressive surgery in the treatment ofuncontrollable post-traumatic intracranial

hypertension in children. Childs Nerv Syst1996; 12:270–275

10. Cho DY, Wang YC, Chi CS: Decompressivecraniotomy for acute shaken/impact syn-drome. Pediatr Neurosurg 1995; 23:192–198

11. Venes JL, Collins WF: Bifrontal decompres-sive craniectomy in the management of headtrauma. J Neurosurg 1975; 42:429–433

12. Kerr FWL: Radical decompression and duralgrafting in severe cerebral edema. Proc MeetMayo Clin Staff 1968; 43:852–864

13. Kjellberg RN, Prieto A: Bifrontal decompres-sive craniectomy for massive cerebral edema.J Neurosurg 1971; 34:488–493



Chapter 15. Surgical Treatment

1. exp craniocerebral trauma/2. head injur$.tw.3. brain injur$.tw.4. 1 or 2 or 35. exp intracranial hypertension/ or “intracranial hypertension”.mp.6. 4 and 57. limit 6 to (newborn infant �birth to 1 month� or infant �1 to 23 months� or preschool child �2 to 5 years� or

child �6 to 12 years� or adolescence �13 to 18 years�)8. limit 7 to English language9. su.fs.

10. drain$.mp.11. exp cerebrospinal fluid shunts/ or “cerebrospinal fluid shunts”.mp.12. Neurosurgery/ or “neurosurgery”.mp.13. shunt$.mp.14. 9 or 10 or 11 or 12 or 1315. 8 and 14


Chapter 16. The use of corticosteroids in the treatment of severepediatric traumatic brain injury

I. RECOMMENDATIONS

1. Standards. There are insufficientdata to support a treatment standard forthis topic.

2. Guidelines. There are insufficientdata to support a treatment guideline forthis topic. The use of steroids signifi-cantly reduces endogenous cortisol pro-duction. The use of steroids may have anassociated increased risk of complicationsof infection in children.

C. Options. The use of steroids is notrecommended for improving outcome orreducing intracranial pressure (ICP) inpediatric patients with severe traumaticbrain injury (TBI). Despite two class IIstudies failing to show efficacy, the smallsample sizes preclude support for a treat-ment guideline for this topic.

D. Indications from Adult Guidelines.The majority of available evidence indi-cates that steroids do not improve out-come or lower ICP in severely head-injured adult patients (1). The routineuse of steroids is not recommended forthese purposes.

II. OVERVIEW

Corticosteroids have been commonlyused in children, for a wide range of neu-rologic diseases, to reduce edema (due totumors, infection, inflammation) and tolessen its neurologic effects. The poten-tial of steroid use in adults following TBIwas first indicated in literature reportingthe benefits of edema reduction and clin-ical improvement in brain tumor pa-tients. As summarized in the “Guidelinesfor the Management of [Adult] SevereTraumatic Brain Injury” (1), there is ev-idence that steroids are useful in reduc-ing cerebral edema, attenuating free rad-ical production, and affording otherbeneficial effects in experimental modelsof TBI, but clinical evidence did not sup-port its use. In the adult literature re-viewed, corticosteroids did not improvefunctional outcome or prove useful in

reducing ICP in patients with severe TBI.The studies cited in the adult guidelinesdid not specifically report on the use ofcorticosteroids in pediatric patients fol-lowing severe TBI, but there was a sug-gestion of potential efficacy in youngerpatients. A specifically pediatric reviewwas necessary to determine whetherthere is adequate evidence to support rec-ommendations for children.

III. PROCESS

We searched Medline and Healthstarfrom 1966 to 2001 by using the searchstrategy for this question (see AppendixA) and supplemented the results with lit-erature recommended by peers or identi-fied from reference lists. Of 45 potentiallyrelevant studies, eight were used as evi-dence for this question (Table 1).


In clinical practice, steroids have beenused in an attempt to reduce posttrau-matic swelling and improve outcome inboth adults and children. The role of ste-roids remains uncertain in the treatmentof TBI, particularly as it relates to thepediatric population.

Cooper et al. (2) performed a prospec-tive, randomized, double-blind clinicaltrial with adults and children using dexa-methasone. There were 76 total patients,ten of whom were �10 yrs of age, and 32who were less �20 yrs of age. Only se-verely injured patients were included,and each of the patients was randomizedto one of three groups—placebo, low-dose steroids, and high-dose steroids. Theadults were given standard doses,whereas the children were given weight-related doses. Assignment to the groupswas randomized on entry into treatment,but there was no stratification by age.Glasgow Outcome Scale (GOS) was as-sessed at 6 months. The analysis per-formed was on the relation of treatmentto outcome. The authors reported that in

older patients there was no difference inoutcome with the use of steroids. Be-cause of the small number of childrenincluded, no conclusions could be drawnregarding steroid use in pediatric TBI.

Fanconi et al. (3) performed a ran-domized, prospective clinical trial on 25pediatric patients using dexamethasoneat 1 mg·kg�1·day�1 for 3 days (n � 13)and 12 controls treated with an alternatestandard regimen. Outcome was deter-mined by 6-month GOS and response byendogenous free cortisol levels. In thisstudy, there was no difference in effect onICP or cerebral perfusion pressure and nodifference in 6-month outcome. Therewas a statistically significant suppressionof cortisol levels up to 6 days posttreat-ment. As well, there was a significantlyincreased bacterial infection rate in thosepatients treated with steroids.

Gobiet (4) reported on a case controlseries of 205 children: 139 who did notreceive steroids, and 66 who did. Theintervention was high-dose dexametha-sone, although no specific dose was re-ported. The nonsteroid treatment groupwas a consecutive sample of 139 patientsfrom the years 1972–1974 who weretreated with neither ICP monitoring noraggressive intensive care unit therapy.The later steroid treatment group (re-cruited 1974–1975) received ICP moni-toring, intensive care unit management,and steroid therapy. There was no lengthof follow-up reported, but the measuresof outcome included mortality rate,length of intubation, ICP, seizures, andscholastic performance. The authors re-ported that there was decreased mortalityrate with the use of steroids but no dif-ference with regard to length of intuba-tion, time of unconsciousness, incidenceof seizures, or neurologic outcome. Acomparison between the groups in thisstudy is compromised because significantdifferences in timing and approach mayconfound many subtle variables of treat-ment. However, given the lack of differ-




Cooper et al.(2), 1979

Prospective, double-blind study of 76 patients with severehead injury (42 of whom were �20 yrs of age, ten ofwhom were �10 yrs of age). The patients were stratifiedfor severity and treated with placebo or weight-relateddoses in the children. Six-month GOS was thendetermined.

III No significant difference in 6-month outcome in the olderchildren and adults. Potentially improved outcome inchildren �10 yrs of age, although numbers were toosmall to determine true differences.

Fanconi et al.(3), 1988

Prospective, randomized clinical trial of 25 patientstreated with placebo or 1 mg�kg�1�day�1 � 3 days.Endogenous free cortisol and 6-month GOS weredetermined.

II Steroid treatment resulted in no difference in ICP, CPP,or outcome. Steroid treatment significantly suppressedendogenous free cortisol and increased infection rate.

Gobiet (4), 1977 Retrospective review of 205 children who received “high-dose” dexamethasone vs. no steroids. The earlynontreated group was from 1972 to 1974; the treatedgroup was the later, more aggressively treatedpopulation with ICP monitors in treatment ofintracranial hypertension.

III No difference in acute variables or outcome, althoughthere was a suggestion of decreased mortality rate inthe treated group.

Group INo ICP Monitoring or Steroids

Group IIICP Monitoring and Steroids

Number 139 66Died 58 10

% mortality 41.7% 15.8% (p �.001)


Gobiet (5), 1977 Retrospective review of 100 patients using no steroids,normal dose steroids, and “high-dose” steroids indiffering doses. The different treatment regimens werenot specifically delineated, nor was the relationshipbetween the use of other therapies. Only acutemeasures were performed.

III There was no difference in outcome. Steroids reducedbrain edema, but there were no data confirming this.

Hoppe et al.(6), 1981

Case series of 22 patients maximally treated with aconglomeration of regimens.

III Younger patients had a better outcome.

Age, yrs GOS 1–3 GOS 4–5 % Good Outcome

�20 3 19 86�20 9 14 61

Total 12 33


James et al.(7), 1979

Case control study of nine patients. Group 1 received noor low-dose steroids, and group 2 received high-dosesteroids. Neurologic exam and 6-month GOS weredetermined.

III No improved long-term outcome based on GOS due tolow numbers, although reported improved GCS, meanICP and ICP wave fluctuations, acute neurologic exam,and ICU and hospital course in the acute period.

Group GOS 1–3 GOS 4–5 % Good Outcome

Group 1, no or low-dose steroids (n � 4) 3 1 25Group 2, high-dose steroids (n � 5) 0 5 100

Total 3 6


ence between the groups on other vari-ables measured, one might fairlyconclude that the addition of an aggres-sive treatment protocol with ICP moni-toring may reduce mortality rate.

Gobiet et al. (5) reported on the studyof 100 patients, including 40 children,but there was no separate report of re-sults in the pediatric group. The authorsreviewed their experience with ICP mon-itoring, hyperosmolar therapy, and treat-ment with or without high-dose steroidsin two consecutive patient groups. Theearlier patients in the series (1973–1974)all had ICP monitoring and a “standard”therapeutic regimen. The later patients(1975) received steroids in addition to theprevious therapies. No conclusion for theuse of dexamethasone in pediatric TBIcan be drawn from this study because thedata for children are confounded withadult measures.

Hoppe et al. (6) reported a case seriesof 22 patients �19 yrs of age who re-ceived intensive therapy in a multiple-treatment regimen including steroids,barbiturates, and hyperventilation. Thesteroid treatment was dexamethasone120 mg, given at admission, 6 and 72 hrsafter injury, combined with 4 mg every 6hrs. They reported outcomes measuredby 3- and 6-month GOS. Their only con-clusion with regard to children was thatyounger patients had better outcomes,but the authors did not report a specificrelation of outcome to the use of dexa-methasone.

James et al. (7) reported a retrospec-tive case series of nine pediatric patientswith severe TBI. Group 1 received no orlow-dose steroids (dexamethasone � 0.25mg·kg�1·day�1), and group 2 receivedhigh-dose steroids (1 mg/kg every 6 hrsfor two doses and then 1 mg·kg�1·day�1).

The children otherwise received standardtreatment for severe TBI. Outcome wasdetermined by their course and length ofstay in the intensive care unit and in thehospital and 6-month GOS. The authorsconcluded that steroids improved Glas-gow Coma Scale (GCS) and neurologicexam by 7-days postinjury, shortened in-tensive care unit and hospital stay, anddecreased mean ICP and ICP wave fluc-tuations. There was no increase in gas-trointestinal hemorrhage or in pulmo-nary infection. There was no significantdifference in GOS outcome at 6 months,although group 2 tended to have betteroutcomes.

Kloti et al. (8) performed a prospec-tive, randomized clinical trial in 24 se-verely head-injured children. Group 1 re-ceived steroids at 1 mg·kg�1·day�1, andgroup 2 received no steroids. The chil-dren were otherwise treated with stan-

Table 1. Continued


Kloti et al. (8),1987

Prospective, randomized clinical trial of 24 patients. Group1 received dexamethasone (1 mg�kg�1�day�1), and group2 received no steroids. Urinary-free cortisol in the acuteperiod and 6-month GOS were used for outcome.

IIA There was near complete suppression of endogenouscortisol, and no difference in long-term outcome.

Group GOS 1–4 GOS 5% GoodOutcome

Group 1, steroids 1 (n � 12) 3 9 75Group 2, no steroids 2 (n � 12) 4 8 67Total 7 17


Kretschmer (9),1983

Retrospective review of 107 patients, 51 of whom receiveda loading dose of dexamethasone, 20–25 mg, and thenreceived a dosing based on whether body weight was �or �35 kg (not based on a mg/kg schedule), inconjunction with standard therapy. This series includedpenetrating injuries, mild to moderate head injuries, aswell as differences in severity between treated andnontreated groups.

III There was lowered mortality rate in the steroid group inpatients with intracranial hematomas and severeinjuries (GCS 5–7), although the small numbersprecluded conclusion of efficacy.

IntracranialHematomas

% ofPatients GCS 5–7

% ofPatients

Group 1 (steroids)Favorable 15 88.2 14 63.6Unfavorable 2 11.8 8 36.4Mortality 2 11.8 3 13.6

Group 2 (no steroids)Favorable 11 57.8 9 60Unfavorable 8 42.1 6 40Mortality 7 36.8 5 33

GOS, Glasgow Outcome Scale; ICP, intracranial pressure; CPP, cerebral perfusion pressure; GCS, Glasgow Coma Scale; ICU, intensive care unit.


dard therapy. In this study, outcomemeasures included urinary-free cortisolto assess cortisol excretion as well asfunctional outcome using the 6-monthGOS. There was no difference betweenthe two groups with regard to mean ICPstay, duration of intubation, or 6-monthGOS. In group 1, there was completeendogenous cortisol suppression. Ingroup 2 patients, free cortisol was in-creased 20-fold from normal basal condi-tions, reaching maximum levels at 1–3days after injury. In addition, 50% of pa-tients in group 1 developed pneumoniacompared with only 15% in group 2.

Kretschmer (9) reported a retrospec-tive review of 107 head-injured pediatricpatients with a pathologic finding oncomputed tomography scan, 51 of whomreceived steroids. Dexamethasone wasgiven as a bolus (20–25 mg) and then asan infusion based on body weight (aboveor below 35 kg). The patients otherwisereceived standard treatment for severeTBI. The two groups differed signifi-cantly: a) the study included 29 patientswith penetrating injuries, 24 of whomwere in the no-steroid group; b) for con-tusive injury, 29 of 42 patients weretreated with steroids; c) mild to moderatehead injuries (GCS 8–15) were includedbut not equally distributed in assign-ment, resulting in different average se-verity in the treated and nontreatedgroups (mean GCS of 7.4 vs. 9, respec-tively). Outcome was measured by theGOS, but length of follow up was not

reported. Although the overall mortalityrate between groups did not differ (24%vs. 23%, treated vs. untreated, respective-ly), the authors concluded that steroidtreatment reduced mortality rate in pa-tients with intracranial hematomas(36.8% vs. 11.8%) and in those with ini-tial GCS 5–7 (33% vs. 14%). Dexameth-asone, however, was not useful in themost severely injured patients (GCS3–4), the mild and moderately injured, orthose with penetrating injuries. Becausethe groups were not evenly balanced onmechanism of injury or severity, conclu-sions on the efficacy of using steroids inpediatric TBI cannot be drawn from thisstudy.

V. SUMMARY

The majority of available evidence in-dicates that steroids did not improvefunctional outcome in pediatric patientswith severe TBI. A few studies reportedbeneficial effect on outcome, but they allhad design problems, so recommenda-tions for steroid use cannot follow fromtheir results. In addition, there were asmany studies that were inconclusive orlacking any evidence of efficacy. A fewstudies did not show evidence of compli-cations from steroid use, but two othersreported significantly increased rates ofinfection (bacterial infections and pneu-monia) and suppression of endogenouscortisol, which further lessens any enthu-siasm for the use of this treatment. Withthe lack of sufficient evidence for benefi-cial effect and the potential for increasedcomplications and suppression of adrenalproduction of cortisol, the routine use ofsteroids is not recommended for childrenfollowing severe TBI.


Efficacy

Despite the lack of sufficient clinicalevidence of efficacy, there may be sub-groups of children with severe TBI whomight benefit from the use of high-dosesteroids in treatment. Examples of candi-date conditions are certain types of pa-thology (like diffuse swelling or intracra-nial hematomas), different levels ofseverity (moderately severely injured,GCS 5–7), and age at injury (school-agechildren). Further experimental and clin-

ical studies that use high-dose steroidswith stratification of these variablesamong the comparison groups will benecessary before recommendations fortreatment can be made.

Complications

Future trials also will need to addressthe issue of complications, specifically in-fection and gastrointestinal hemorrhage,and whether preventive interventions(e.g., antibiotics and/or H2 blockers) areeffective in reducing or eliminating oc-currences.

Endogenous Cortisol

There is evidence that endogenouscortisol production is suppressed withthe administration of corticosteroids fol-lowing severe TBI. The significance of thesuppression of endogenous cortisol pro-duction and its effect on clinical courseand outcome, as well as the effect of anincreased catabolic response, needs to beanswered.

REFERENCES


2. Cooper PR, Moody S, Clark WK, et al: Dexa-methasone and severe head injury. J Neuro-surg 1979; 51:307–316

3. Fanconi S, Kloti J, Meuli M, et al: Dexameth-asone therapy and endogenous cortisol pro-duction in severe pediatric head injury. Inten-sive Care Med 1988; 14:163–166

4. Gobiet W: Advances in management of severehead injuries in childhood. Neurochirurgica1977; 39:201–210

5. Gobiet W: Monitoring of intracranial pressurein patients with severe head injury. Neuro-chirurgica 1977; 20:35–47

6. Hoppe E, Christensen L, Christensen KN: Theclinical outcome of patients with severe headinjuries, treated with high-dose dexametha-sone, hyperventilation and barbiturates. Neu-rochirurgica 1981; 24:17–20

7. James HE, Maudauss WC, Tibbs A, et al: Theeffect of high dose dexamethasone in childrenwith severe closed head injury. Acta Neuro-chir 1979; 45:225–236

8. Kloti J, Fanconi S, Zachmann M, et al: Dexa-methasone therapy and cortisol excretion insevere pediatric head injury. Childs Nerv Syst1987; 3:103–105

9. Kretschmer H: Prognosis of severe head inju-ries in childhood and adolescence. Neuropedi-atrics 1983; 14:176–181

W ith the lack of

sufficient evi-

dence for bene-

ficial effect and the potential

for increased complications

and suppression of adrenal

production of cortisol, the

routine use of steroids is not

recommended for children

following severe traumatic

brain injury.




Chapter 16. Steroids

1. exp craniocerebral trauma/2. head injur$.tw.3. brain injur$.tw.4. 1 or 2 or 35. exp steroids/ or “steroids”.mp.6. glucocorticoics, synthetic/ or “synthetic glucocorticoids”.mp.7. 5 or 68. 4 and 79. limit 8 to (newborn infant �birth to 1 month� or infant �1 to 23 months� or preschool child �2 to 5 years� or



Chapter 17. Critical pathway for the treatment of establishedintracranial hypertension in pediatric traumatic brain injury

Acritical pathway, developed byconsensus, is presented in Fig-ures 1 and 2. We developed atreatment algorithm for estab-

lished intracranial hypertension, whereinthe order of steps is determined by therisk/benefit ratio of individual treatmentmaneuvers. The considerations involvedare outlined in the chapter specific toeach step.

As discussed in the section on intra-cranial pressure (ICP) treatment thresh-old, the absolute value defining unaccept-able intracranial hypertension is unclear.Although a general threshold of 20 mmHg has been presented, there will be sit-uations where such pressures are toohigh as well as instances where higherICP values are acceptable. These consid-erations are relevant to the decision topursue any step in the escalated treat-ment of ICP.

This critical pathway is a committeeconsensus and, therefore, must be viewedas class III (“expert opinion”) evidence. Assuch, it should be interpreted as a frame-work that may be useful in guiding anapproach to treating intracranial hyper-tension. It can and should be modified inan individual case by any circumstancesunique to the patient as well as by theresponse of the ICP to individual treat-ment steps.

This algorithm applies to severe trau-matic brain injury. The decision to mon-itor ICP and apply this algorithm to chil-dren or infants with lesser degrees ofbrain injury is left to the physician.

Critical Pathway

The most fundamental maneuver inmanaging the pediatric patient with se-vere traumatic brain injury, outside ofthe surgical evacuation of intracranial

mass lesions, is the insertion of an ICPmonitor. Once this has been accom-plished, treatment can be directed at con-trolling ICP and cerebral perfusion pres-sure (CPP). A number of generalmaneuvers may be applied to this groupof patients early during their treatment,including control of fever, avoidance ofjugular venous outflow obstruction, andmaintenance of adequate arterial oxygen-ation. The initial PaCO2 should be main-tained at the low end of eucapnia (35 mmHg). In addition, regardless of the presenceor absence of intracranial hypertension,CPP should be maintained. The exact valueto be targeted should be predicated on ageand may be modified by advanced cerebralphysiologic monitoring.

When intracranial hypertension oc-curs, the adequacy of sedation and anal-gesia should be checked and augmentedas needed. In euvolemic patients, thehead of the bed may be elevated to ap-proximately 30° and the response of ICPand CPP monitored for efficacy. The ad-dition of neuromuscular blockade to thetreatment regimen also may be consid-ered at this point.

If ICP remains elevated despite ade-quate sedation, analgesia, and elevationof the head of the bed (with or withoutneuromuscular blockade), the initial ICP-specific therapeutic intervention shouldbe CSF drainage when ventricular accessis available. Should an ICP monitor lack-ing the capability of CSF drainage havebeen placed, consideration should begiven to obtaining ventricular access.

If CSF drainage is ineffective in con-trolling ICP or is not available, hyperos-molar therapy should be considered.There is insufficient evidence to supportprioritizing the use of mannitol vs. hy-pertonic saline as a first choice. The pa-tient’s volume status should be closely

observed during mannitol administra-tion, and the upper limits of 320 mOsm/Lfor mannitol and 360 mOsm/L for hyper-tonic saline should be observed. If hyper-osmolar therapy proves ineffective, thelevel of ventilation may be increased toobtain PaCO2 levels of 30–35 mm Hg.Measurement of cerebral blood flow, jug-ular venous saturation, or tissue oxygentension should be considered when hy-perventilation is increased.

At all times during the treatment ofintracranial hypertension, the possibilitythat a surgical mass or an unexpectedintracranial lesion may have developedshould be considered. Therefore, underconditions of intractability or loss of ICPcontrol or when second-tier therapy isbeing contemplated, clinicians shouldconsider repeating a computed tomogra-phy scan.

For intracranial hypertension refrac-tory to the previously described tech-niques, second-tier therapies should beconsidered when the physician believesthat the patient may benefit if ICP controlcan be accomplished. Second-tier therapyincludes ICP-lowering treatment modali-ties that have been demonstrated to im-prove outcome at some level of evidencebut that have not been subjected to trialscomparing them to alternate second-tiertreatments to establish relative risk/benefit ratios. Details of the literatureaddressing the various second-tier thera-pies presented here are found in otherchapters. Some considerations relevantto differentiating between these differingsecond-tier maneuvers are contained inthe algorithm. The precise indications forselecting and applying second-tier thera-pies in an individual patient are left to thediscretion of the managing physician.


Figure 1. First tier. GCS, Glasgow Coma Scale; ICP, intracranial pressure; CPP, cerebral perfusion pressure; HOB, head of bed; CSF, cerebrospinal fluid;CT, computed tomography; PRN, as needed.


Figure 2. Second tier. ICP, intracranial pressure; CT, computed tomography; EEG, electroencephalogram; CBF, cerebral blood flow; SjO2, jugular venousoxygen saturation; AJDO2, arterial-jugular venous difference in oxygen content.


Chapter 18. Nutritional support

I. RECOMMENDATION



C. Options. Replace 130–160% of rest-ing metabolism expenditure after trau-matic brain injury (TBI) in pediatric pa-tients. Weight-specific resting metabolicexpenditure guidelines can be found inTalbot’s tables (1).

Based on the adult guidelines, nutri-tional support should begin by 72 hrswith full replacement by 7 days.

D. Indications from Adult Guidelines.At a guideline level, replace 140% of rest-ing metabolism expenditure in nonpara-lyzed patients and 100% of resting me-tabolism expenditure in paralyzedpatients, by using enteral or parenteralformulas containing �15% of calories asprotein by day 7 after injury (2).

At an option level, jejunal feeding bygastrojejunostomy is preferred due to easeof use and avoidance of gastric intolerance.

II. OVERVIEW

The nutritional status of pediatric TBIpatients may be critical to the recovery pro-cess. The question remains unansweredwhether nutritional formulations; glucosemetabolism; the amount, type, method, ortiming of feeding; or any other specific nu-tritional intervention influences outcomeof pediatric TBI patients. There are onlytwo studies (one class II and one class IIIevidence) that adequately addressed metab-olism in children with a brain injury, andno studies that looked at differences inmorbidity or mortality rates (3, 4).

III. PROCESS

We searched Medline and Healthstarfrom 1966 to 2001 by using the searchstrategy for this question (see AppendixA) and supplemented the results with lit-erature recommended by peers or identi-

fied from reference lists. Of 35 potentiallyrelevant studies, two were used as evi-dence for this question (Table 1).


Two studies (one class II and one classIII) addressed hypermetabolism and nu-tritional support in pediatric TBI (3, 4).Phillips et al. (3) examined energy expen-diture, nitrogen excretion, and serumprotein levels in pediatric TBI patientswith a Glasgow Coma Scale of 3–8. Thestudy included eight adolescents (age11–17 yrs) and four children (age 2–5yrs). Eleven sustained a blunt head in-jury, and one had a penetrating head in-jury by gunshot. All patients were ini-tially intubated and mechanicallyventilated for 5–14 days. No patients re-ceived corticosteroids. Four patients re-ceived neuromuscular blockade. Sevenpatients were started on total parentalnutrition on days 2–6 after injury, andfive were started on enteral nutrition ondays 3–12 after injury. The study demon-strated a significant elevation in themean energy expenditure of approxi-mately 130% above predicted levels forthe group as a whole (Harris-Benedictequation). Seventy percent achieved ni-trogen balance by 4–14 days. The meanurinary nitrogen excretion was 307mg·kg�1·day�1 for the adolescents (meannitrogen balance of �13.6) and 160mg·kg�1·day�1 for the children (mean ni-trogen balance of �4.1). Mean serum al-bumin levels decreased slightly (from 2.9mg/dL down to 2.4 mg/dL) whereas totalprotein levels increased (from 5.4 mg/dLto 6.0 mg/dL). Weight loss ranged from 2to 26 lb during the 2-wk period showing,despite aggressive nutritional support, asignificant decrease in body weight of 9%for adolescents and 4% for children. Theeffect of neuromuscular blockade on me-tabolism was not addressed.

Moore et al. (4) measured the meta-bolic profiles of 20 severe TBI patients(Glasgow Coma Scale �7). Although theystudied both adult and pediatric patients(13 and seven, respectively), the data spe-

cific to the pediatric patients were suffi-ciently reported to allow inclusion asclass III evidence. All patients were me-chanically ventilated. Two of the patientsreceived corticosteroids. Neuromuscularblockade was not addressed. In all pa-tients, feeding was initiated within 48 hrsof admission to the trauma units. It wasdifficult to ascertain which patients re-ceived total parental nutrition vs. enteralnutrition. In the pediatric group (age3–16 yrs), the study demonstrated an av-erage of 180% of predicted for oxygenconsumption and 173% predicted of rest-ing energy expenditure. The average re-spiratory quotient was 0.68 for both thewhole group and the pediatric subgroup.

Although neither of these studies ad-dressed the effect of nutritional support onoutcome, the data demonstrate a sizableincrease in energy expenditure. The in-crease in expenditure was highly variableamong patients. These studies suggest anincreased need for nutritional support inpediatric TBI patients. Insufficient data pre-vent thorough comparison between themetabolism in adults and children, but thepediatric findings are similar to those wellestablished in the adult literature.


Hypermetabolism after TBI has beenwell documented in the adult literature(2–20). At least 12 class I studies havebeen completed. Nine class I studies ex-amined the effect of amount of feeding,type of feeding, route of feeding, and cor-ticosteroids on nitrogen balance and se-rum biochemistries. One class I studyexamined the effect of insulin growth fac-tor I on the catabolic state and on out-come (21). Two class I studies examinedthe relation between extent of nutritionalreplacement and outcome (22, 23).

Data from investigators measuringmetabolic expenditure in rested comatosepatients with isolated TBI showed a meanincrease of approximately 140% of the


expected metabolic expenditure, withvariations from 120 to 250% of that ex-pected. In TBI patients, neuromuscularblockade or barbiturate coma decreasedmetabolic expenditure from a mean of160% of the expected to 100 –120%.These findings suggest that a major partof the increased metabolic expenditure isrelated to muscle tone. Even with neuro-muscular blockade, energy expenditureremained elevated by 20–30% in somepatients (7). In the first 2 wks after injury,energy expenditure seems to increase re-gardless of neurologic course.

Class I data suggest that when TBIpatients are not fed within the first week,mortality rate is increased. Data in criti-cally ill patients without TBI show that a30% weight loss was associated with anincreased mortality rate. After severe TBI,both energy requirements and nitrogenexcretion markedly increase. Fasting pa-tients with severe TBI continue to lose14 –25 g N/day (11). Data show thatstarved TBI patients lose sufficient nitro-gen to reduce weight by 15% per week.Given adequate nutritional support, thecontribution of protein to consumed cal-ories after TBI increases to levels as highas 30% (24). Class II data show that 100–140% replacement of resting metabolismexpenditure with 15–20% nitrogen calo-ries reduces nitrogen loss. The datastrongly support full nutritional replace-ment by the end of the first week afterinjury.

The optimal method of feeding has notbeen established. Similarly, it has notbeen shown that earlier feeding (full feed-ing before 7 days) improves outcome.Studies showed that, with nearly equiva-lent quantities of feeding, the mode ofadministration (total parental nutritionor enteral nutrition) had no effect onneurologic outcome, despite other poten-tial advantages of enteral nutrition (de-creased risk of hyperglycemia, lowercosts, lower risk of infection). For enteralnutrition, jejunal feeding was better tol-erated than gastric feeding (25, 26).

Based on the level of nitrogen wast-ing documented in TBI patients and thenitrogen sparing effect of feeding, it is aguideline that full nutritional replace-ment be instituted in the adult TBI pa-tient by day 7. It is suggested that en-teral nutrition begin no later than 72hrs.

Finally, studies of glucose homeostasissuggest that serum glucose control may becritical to limiting secondary neurologicdamage. In animal models, hyperglycemiahas been shown to worsen ischemic braininjury (27–29). Hyperglycemia was associ-ated with a worse outcome in critically illpatients (30, 31), and tight control of glu-cose levels (�110) in nondiabetic patientswas associated with improved outcome(32). In two studies of severe head injury,hyperglycemia also was associated withworse outcome (33, 34). These studies sug-

gest that serum glucose levels should betightly controlled in TBI patients.

V. SUMMARY

Due to the limited data that exist onthe nutritional requirements of pediatricTBI patients, recommendations can onlybe made at the option level. Of the twostudies addressing pediatric patients,both showed a significant increase in themetabolic rate associated with TBI. Thesefindings are similar to those reported inthe adult literature. Without further data,the adult guidelines, adjusted for weight,should be considered when providing nu-tritional support to pediatric patientswith TBI.


● Little research has been conducted onnutritional support and its influenceon outcome of pediatric TBI. These pa-tients are likely to be highly dependenton precise nutritional interventionstailored to age, weight, body surfacearea, and other factors.

● Recent studies of glucose homeostasissuggest that glucose control may becritical to secondary damage and neu-rologic outcome. This topic needs in-vestigation in pediatric patients.

● The issue of hypermetabolism in pedi-atric TBI needs further confirmation



Phillips et al. (3),1987

Energy expenditure, nitrogen excretion, and serum proteinlevels in pediatric TBI patients with a GCS of 3–8. Total n� 12 (eight adolescents age 11–17 yrs and four childrenage 2–5 yrs). Eleven blunt head injuries and one GSW. Allpatients were initially intubated and mechanicallyventilated in the emergency department and continued for5–14 days. No patients received steroids. Four patientswere paralyzed with Pavulon. Seven patients were startedon TPN on days 2–6 after injury, and five were started onEN on days 3–12 after injury.

II The mean energy expenditure was approximately 130% abovepredicted for the whole group. Seventy percent achievednitrogen balance by 4–14 days. The mean urinary nitrogenexcretion was 307 mgdb1�kg�1�day�1 for the adolescents(mean nitrogen balance of �13.6) and 160 mg�kg�1�day�1

for the children (mean nitrogen balance of �4.1). Meanalbumin levels decreased slightly (2.9 mg/dL down to 2.4mg/dL), whereas total protein levels increased (from 5.4 mg/dL to 6.0 mg/dL). Weight loss ranged from 2 to 26 lbduring the 2-wk period showing, despite aggressivenutritional support, a significant decrease in body weight of9% for adolescents and 4%.

Moore et al. (4),1989

The metabolic profiles of severe closed TBI patients (GCS�7) were measured. Mixed adult and pediatric, total n �20 (13 adults, seven children). All patients weremechanically ventilated. Two received steroids. Paralysiswas not addressed. In all patients, feeding was initiatedwithin 48 hrs of admission to the trauma units. Nodistinction between TPN vs. EN.

III In the pediatric group (age 3–16 yrs), the study demonstratedan average of 180% of predicted for oxygen consumptionand 173% predicted of resting energy expenditure. Theaverage respiratory quotient was 0.68 for both the wholegroup and the pediatric subgroup.

TBI, traumatic brain injury; GCS, Glasgow Coma Scale; GSW, gunshot wound; TPN, total parenteral nutrition; EN, enteral nutrition.


and clarification. The effects of timing,type, quality, and methodology of nu-tritional support on the outcome ofpediatric patients after TBI need to beinvestigated.

● Special issues such as vitamin andmineral replacement and other nutri-tional interventions need to be ex-plored. Authors need to stratify differ-ent age groups (infants, toddlers,children, adolescents) in these studiesto further delineate developmental dif-ferences in nutritional requirements.Without further data, the adult guide-lines, adjusted for weight, should beconsidered when providing nutritionalsupport to pediatric patients with trau-matic brain injury.

REFERENCES

1. Talbot F: Basal metabolism standards forchildren. J Dis Child 1938; 55:455–459


3. Phillips R, Ott L, Young B, et al: Nutritionalsupport and measured energy expenditure ofthe child and adolescent with head injury.J Neurosurg 1987; 67:846–851

4. Moore R, Najarian MP, Konvolinka CW: Mea-sured energy expenditure in severe headtrauma. J Trauma 1989; 29:1633–1636

5. Bruder N, Dumont JC, Francois G: Evolutionof energy expenditure and nitrogen excretionin severe head-injured patients. Crit CareMed 1991; 19:43–48

6. Clifton GL, Robertson CS, Grossman RG, et

al: The metabolic response to severe headinjury. J Neurosurg 1984; 60:687–696

7. Clifton GL, Robertson CS, Choi SC: Assess-ment of nutritional requirements of head-injured patients. J Neurosurg 1986; 64:895–901

8. Deutschman CS, Konstantinides FN, Raup S,et al: Physiological and metabolic responseto isolated closed-head injury. Part 1: Basalmetabolic state: Correlations of metabolicand physiological parameters with fastingand stressed controls. J Neurosurg 1986; 64:89–98

9. Dickerson RN, Guenter PA, Gennarelli TA, etal: Increased contribution of protein oxida-tion to energy expenditure in head-injuredpatients. J Am Coll Nutr 1990; 9:86–88

10. Fell D, Benner B, Billings A, et al: Metabolicprofiles in patients with acute neurosurgicalinjuries. Crit Care Med 1984; 12:649–652

11. Gadisseux P, Ward JD, Young HF, et al: Nu-trition and the neurosurgical patient. J Neu-rosurg 1984; 60:219–232

12. Long CL, Schaffel N, Geiger JW, et al: Meta-bolic response to injury and illness: Estima-tion of energy and protein needs from indi-rect calorimetry and nitrogen balance. JPENJ Parenter Enteral Nutr 1979; 3:452–456

13. Ott LB, Young B, McClain C: The metabolicresponse to brain injury. JPEN J ParenterEnteral Nutr 1987; 11:488–493

14. Piek J, Lumenta CB, Bock WJ: Protein andamino acid metabolism after severe cerebraltrauma. Intensive Care Med 1985; 11:192–198

15. Piek JK, Brachwitz K, Bock WJ: The energyconsumption of patients with craniocerebraltrauma and spontaneous intracranial hemor-rhage in the early postoperative post-traumatic phase. Anasth Intensivther Not-fallmed 1988; 23:325–329

16. Robertson CS, Clifton GL, Grossman RG: Ox-ygen utilization and cardiovascular functionin head-injured patients. Neurosurgery 1984;15:307–314

17. Sunderland PM, Heilbrun MP: Estimatingenergy expenditure in traumatic brain in-jury: Comparison of indirect calorimetrywith predictive formulas. Neurosurgery1992; 31:246–252

18. Young B, Ott L, Norton J, et al: Metabolic andnutritional sequelae in the non-steroidtreated head injury patient. Neurosurgery1985; 17:784–791

19. Young B, Ott L, Twyman D, et al: The effectof nutritional support on outcome from se-vere head injury. J Neurosurg 1987; 67:668–676

20. Deitrick JE, Chestnut GW, Shorr E: Effects ofimmobilization upon various metabolic and

physiological functions of normal men. Am JMed 1948; 4:3–36

21. Hatton J, Rapp RP, Kudsk KA, et al: Intrave-nous insulin-like growth factor-I (IGF-I) inmoderate-to-severe head injury: A phase IIsafety and efficacy trial. J Neurosurg 1997;86:779–786

22. Ott L, Young B, Phillips R, et al: Alteredgastric emptying in the head-injured patient:Relationship to feeding intolerance. J Neuro-surg 1991; 4:738–742

23. Robertson CS, Goodman JC, Narayamn RK,et al: The effect of glucose administration oncarbohydrate metabolism after head injury.J Neurosurg 1991; 74:43–50

24. Duke JH Jr, Jorgensen SB, Broell JR, et al:Contribution of protein to caloric expendi-ture following injury. Surgery 1970; 68:168–174

25. Clifton GL, Robertson CS, Contant CF: En-teral hyperalimentation in head injury.J Neurosurg 1985; 62:186–193

26. Kirby DF, Clifton GL, Turner H, et al: Earlyenteral nutrition after brain injury by percu-taneous endoscopic gastrojejunostomy.JPEN J Parenter Enteral Nutr 1991; 15:298–302

27. Cherian L: Hyperglycemia increases brain in-jury caused by secondary ischemia after cor-tical impact injury in rats. Crit Care Med1997; 25:1378–1383

28. Cherian L: Effect of glucose administrationon contusion volume after moderate corticalimpact injury in rats. J Neurotrauma 1998;15:1059–1066

29. Hovda DA: The increase in local cerebralglucose utilization following fluid percussionbrain injury is prevented with kynurenic acidand is associated with an increase in calcium.Acta Neurochir Suppl 1990; 51:331–333

30. McGonigal MD, Lucas CE, Ledgerwood AM:Feeding jejunostomy in patients who arecritically ill. Surg Gynecol Obstet 1989; 168:275–277

31. Bruno A: Acute blood glucose level and out-come from ischemic stroke. Trial of ORG10172 in Acute Stroke Treatment (TOAST)Investigators. Neurology 1999; 52:280–284

32. van den Berghe G, Wouters P, Weekers F, etal: Intensive insulin therapy in the surgicalintensive care unit. N Engl J Med 200; 345:1359–1367

33. Lam AM, Winn HR, Cullen BF, et al: Hyper-glycemia and neurological outcome in pa-tients with head injury. J Neurosurg 1991;75:545–551

34. Rovlias A, Kotsou S: The influence of hyper-glycemia on neurological outcome in pa-tients with severe head injury. Neurosurgery2000; 46:335–342


W ithout further

data, the adult

guidelines,

adjusted for weight, should

be considered when provid-

ing nutritional support to

pediatric patients with

traumatic brain injury.




Chapter 18. Nutrition

1. exp craniocerebral trauma/2. head injur$.tw.3. brain injur$.tw.4. 1 or 2 or 35. exp parenteral nutrition/ or “parenteral nutrition”.mp.6. nutritional support/ or “nutritional support”.mp.7. 5 or 68. 4 and 79. limit 8 to (newborn infant �birth to 1 month� or infant �1 to 23 months� or preschool child �2 to 5 years� or



A SUPPLEMENT TO · 2019. 6. 12. · P. David Adelson, MD, FACS, FAAP University of...

Documents

Transcript of A SUPPLEMENT TO · 2019. 6. 12. · P. David Adelson, MD, FACS, FAAP University of...