41ournal ofNeurology, Neurosurgery, and Psychiatry 1993;56:410-415

Long-term outcome of head injuries: a 23 yearfollow up study of children with head injuries

Harry Klonoff, Campbell Clark, Pamela S Klonoff

Department ofPsychiatry, Universityof British Columbia,Vancouver, BritishColumbia, CanadaH KlonoffC ClarkBarrow NeurologicalInstitute, St. Joseph'sHospital and MedicalCentre, Phoenix,Arizona, USAP S KlonoffCorrespondence to:Professor Klonoff,University of BritishColumbia, FairmontMedical Building, 706-750West Broadway, Vancouver,British Columbia, CanadaV5Z 1H6.

Received 7 February 1992and in revised form6 July 1992.Accepted 20 July 1992

AbstractThe purpose of the 23 year follow upstudy was to determine the relationshipbetween trauma variables including mea-sures of head injury and very long-termsequelae. The study included 159 individ-uals with a mean age 31-40 years, ofwhom approximately 90% were admittedto hospital with a mild head injury dur-ing childhood (mean age 7.96). Extent ofhead injury was determined by uncon-sciousness, neurological status, skullfracture, EEG, post-traumatic seizuresand a composite measure. The compositemeasure of neurological variables wasthe best predictor of long-term outcome.In addition, IQ recorded in the post-acute phase was a reliable predictor oflong-term outcome. Of the sample,32-7% reported physical complaints and17*6% reported current psychological/psychiatric problems unrelated to thehead injury. Subjective sequelae (physi-cal, intellectual and emotional) specifiedas due to the head injury were reportedby 31% of the sample, and the sequelaewere found to be related to the extent ofthe head injury and initial IQ. Therewere no discernible relationships bet-ween attribute variables including pre-morbid status and age with subjectivesequelae. There were, however, signifi-cant relationships between subjectivesequelae and objective, psychosocialmeasures of adaptation including educa-tional lag, unemployment, current psy-chological/psychiatric problems andrelationships with family members.Finally, there appeared to be continuityof complaints elicited during the five yearfollow up of the original project and cur-rent sequelae. The severity of the headinjury was identified as the primary con-tributory factor in the reconstitutionprocess and in the prediction of longterm outcomes.

(7 Neurol Neurosurg Psychiatry 1993; 56:410-415)

Despite extensive research on the diverseeffects of head injuries, particularly mild headinjuries, there is a profound paucity of verylong-term outcome studies in children sus-taining traumatic brain injuries. In 1967 aprospective study of 231 children who sus-tained closed head injuries was initiated. Insubsequent years a number of publications

including the final results of the five year fol-low up study were published.'2 The currentstudy is a 23 year follow up of the childrenincluded in the original project. The purposeof the study was to determine the relationshipbetween trauma variables including measuresof extent of head injury and very long-termsequelae.The acute neurological, cognitive, emo-

tional and physical sequelae of closed headinjury34 including mild head injuries5 are welldocumented for children' and adult6 samples.

Recent interest has focused on the shortand long-term outcome following traumaticbrain injury. Publications range from the sub-acute phase of one month after the headinjury,7 to follow up for periods up to fouryears,8 and late outcome studies from six to15 years post-trauma.9-" No published studyhas evaluated outcome as long as 23 yearsafter injury. In addition, no study has docu-mented the long-term consequences of headinjury from childhood to adulthood in thesame patient sample.

In follow up studies, enduring neuropsy-chological deficits have been reported.6 Othercommonly identified problems at follow upare in the areas of behavioural dysfunction,including anxiety, depression and social with-drawal.8'3' Families identify the behaviouralchanges as the most enduring and trouble-some compared with the physical and cogni-tive sequelae.9

Increasingly, studies have evaluatedchanges in work status after head injury.Results are variable, but most studies report asignificant degree of unemployment orreduced work capacity.9 '1-'7 Several studieshave evaluated factors affecting return towork after head injury. Decreased rates ofemployment have been found with increasedseverity of injury, as measured by length ofcoma,'5 by initial Glasgow Coma Scale score8and length of post-traumatic amnesia."'Greater cognitive dysfunction has also beenrelated to poorer vocational outcome.'920A growing body of literature has evaluated

predictors of outcome after head injury,including severity of head injury and age attime of trauma.4 It has been suggested thatchildren sustaining brain injuries showimproved neurological and cognitive recoverycompared with older age groups.4 -23 Onestudy, however, reported no significant rela-tionship between age and outcome.24

Typically, the recovery of patients sustain-ing head injury has been determined by inter-viewer ratings.25 The best known of these is

410

4

on April 4, 2021 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.56.4.410 on 1 A

pril 1993. Dow

nloaded from

http://jnnp.bmj.com/

Long-term outcome of head injuries: a twenty three yearfollow up study of children with head injunres

Table 1 Demographic charactenrstics of cohort at time ofhead injury (n=159)

CharacteristicsAge (years)Mean 7-96SD 3-28Range 2-7-15-9

Educational placement (%)Pre-school 29-6Primary (lower secondary) 60-3Secondary (upper secondary) 10-1

Occupations of fathers (%)Professional and semi-professional 24-5Clerical and skilled 44-1Semi-skilled and unskilled 27-6Unemployed 3-8

Agent of injury (%)Automobile 44Fall 48Other 8

Length of hospitalizationMean days 12 7

Litigation (%) 21-4Pre-morbid anomalies (%) 27-0Multiple injuries (%) 26-4Musculoskeletal 19-5Abdomen 2-5Ears/nose 1-9Eyes 1*3Skin 0-6Respiratory 0-6

the Glasgow Outcome Scale.26 Recently,research has started to focus on outcomebased on patient interviews,8 and psychoso-cial changes in head injured adults usingquestionnaire data from patients comparedwith their families.27

Methods and resultsThe initial project set out to investigateprospectively a head-injured group of chil-dren from the time of trauma to the fifth yearafter trauma, within the context of antecedentfactors (pre-morbid anomalies, age, sex), cir-cumstances at time of head injury (extent ofinjury) and consequence factors (education,interpersonal transactions, sequelae). Theoriginal sample comprised 231 children- 147boys and 84 girls-with a mean age of 8-32years at the time of the head injury. Childrenincluded in the project were consecutiveadmissions to two university hospitalsbetween August 1967 and November 1968with a diagnosis of head injury. The number

Table 2 Neurological indicators of extent of head injury (%/6)

Loss of Neurological Skull EEGConsciousness status fracture rating Seizure

Not proven or No clinical None Normal/ Absentmomentary evidence of Equivocal

trauma(52-2) (24 5) (59-7) (40 3) (93 7)

Unconsciousness5 minutes to 5 minutes to less Simple Minimal Petit malless than than 30 minutes linear abnormal30 minutes or

concussionor

skull fracture(simple)

(37-1) (60 4) (22 0) (38-4) (4 4)Unconsciousness

more than more than Basal/ Moderate/ Grand30 minutes 30 minutes depressed marked mal

or abnormalconcussionor

skull fracture(basal/depressed)with

other symptoms(for example aphasia)

(107) (15-1) (18-3) (21-3) (1 9)

of re-examinations varied among the groupand 117 children were examined during thefifth year of follow up.

During 1990-91, 175 (76%) of the indi-viduals included in the original project weretraced by a variety of means and 159 (91%)volunteered to participate in the currentstudy. Of the remaining 16, contacts weremade with parents or relatives but the mem-ber of the cohort did not return telephonecalls. A university approved consent form wascompleted by the volunteers who were inter-viewed. Geographic locations of the cohortincluded Canada, USA, Europe, Australiaand the Middle East. The senior author inter-viewed all the volunteers either in person(n = 82) or by telephone (n = 77). In twoinstances information was provided by a par-ent with the volunteer, while in two addition-al instances only the parent providedinformation because of the volunteer's mentalstatus.A standardised interview was conducted

with a predetermined format. Details of theoriginal data base obtained during admissionto hospital and on follow up were unknownto the interviewer and the volunteer. Any dif-ference therefore in subjective impressionsshould be randomly distributed among thosewith or without elicited sequelae.

Table 1 describes the demographic charac-teristics of the cohort of 159 adults at thetime of trauma.

Although the Glasgow Coma Scale is nowwidely used to determine the severity of headinjury, this project was conducted before itsdevelopment. Therefore, four uni-dimension-al neurological indexes (length of uncon-sciousness, skull fractures, EEG ratings2 andpost-traumatic seizures) and one global mea-sure (neurological status) were used as indi-cators of the extent of head injury. The lastmeasure was derived from the medical opin-ion of one examiner. These variables andrespective percentages are itemised in table 2.

In addition to the ratings of each variablefrom 1 to 3, a composite score of 5-15 wasalso derived by summing the 5 variables. Thecomposite score ranged from 5-13 with amedian of 7, with the following distribution:5-8 (63-5%); 9-10 (27-6%); 11-13 (8-9%).The initial neuropsychological examination

revealed a mean (SD) IQ of 103-0 (15-0),with a range from 46 to 136. The fifth yearfollow up mean (SD) IQ was 111-2 (11-7),with a range from 77 to 137.

Table 3 describes the current demographicand personal-social characteristics of thecohort.

Health historySubsequent (recurrent) head injuries werereported by 15-1% (8-8% with loss of con-sciousness) of the sample and the number ofsuch head injuries were as follows: 1-10-7%;2-2-5%; 3-1-3%; 4-0-6%.

Table 4 summarises the 66 interveningphysical complaints (not mutually exclusive)specified by the sample as unrelated to thehead injury. The physical complaints were

41 on A

pril 4, 2021 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.56.4.410 on 1 April 1993. D

ownloaded from

http://jnnp.bmj.com/

2Klonoff,Clark, Konoff

Table 3 Current demographic and personal-socialcharacteristics (n=159)

CharacteristicsAge (years)MeanSDRange

Male:Female (n)Education (%)Grades 7-9 (lower secondary)Grades 10-12 (upper secondary)Post-secondary (tertiaryvocational professional)

Bachelors degreesPost-graduate degrees

Grade failure (retention) (%)Marital status (%)Married/common-lawSingleDivorced/separated

Occupation (%)Professional and semi-professionalClerical and skilledSemi-skilled and unskilledHomemakerStudent

Employment status (%)Full timePart timeHomemakerStudentDisability unrelated to traumaUnemployed

Interpersonal relationships (%)Strained relationship with spouseStrained relationship with family members

Leisure activities restricted (%)Alcohol problem (%)PastCurrent

Illicit drug use (%)Past socialPast heavyCurrent socialCurrent heavy

Contact with legal authorities (%)One criminal chargeMultiple criminal charges

31 403-23

25-40105:54

7-650 3

28-910-13-1

25-8

59-228-911-9

27-138-427-64.42-5

80-17-34.42-51.93 8

11-917-930-2

10-14-4

2 56 3

15 12-5

17-02-64-4

itemised in a systems schema.28 However,some of these complaints reported by 32-7%of the sample may be related to the multipleinjuries sustained at the time of the headinjury or in a number of instances, subse-quent head injuries. Seizures were reportedby one individual who had a severe headinjury in 1989 and dizzy spells were reportedby another individual who had relativelysevere head injuries in 1976 and 1990.

Resolved psychological/psychiatric prob-lems specified as unrelated to the initial headinjury were reported by 31-4% of the sample.The 60 reported problems (not mutually

Table 4 Physical complaints unrelated to head injury(n=159)

System Complaints Frequency

Neurological Seizures 1Dizzy spells 1

Musculo-skeletal Leg/back pain 20Arthritis 8

Head and Neck Headaches 5Neck pain 3Thyroid 2

Ears/Nose Reduced hearing 4Sinusitis 2

Abdomen Colitis 1Crohn's disease 1Hepatitis 1Ulcers 1Liver 1

Skin Psoriasis/eczema 3Chest and Lungs Asthma 8Lymphatic Lymphoma 1Heart/Blood vessels Rheumatic fever 2Metabolism Diabetes 1

exclusive) were categorised as: emotional dis-orders (40'0%); problems with spouse orchild (36&6%); problems with parents(15'0%); substance abuse (5 0%); and rela-tionships in general (3-4%).

Current psychological/psychiatric problemsspecified as unrelated to the initial headinjury were reported by 17-6% of the sample.The 33 reported problems (not mutuallyexclusive) were categorised as: emotional dis-orders (42A4%) including one chronic schizo-phrenic; problems with spouse or child(21-2%); substance abuse (12-1%); problemswith parents (9 1%); relationships in general(9 1%); and sleep problems (6 1%).A comparison of those in the sample listing

no psychological/psychiatric problems orspecifying problems both in the past and thepresent, revealed the following: 62-3% listedno problems at any time; 11-3% listed prob-lems both in the past and present; 20- 1% list-ed a problem in the past but not currently;and, 6&3% did not list a problem in the pastbut did currently.Mood was determined during the interview

by asking individuals to rate their mood on ascale from one to 10 where one would be verydepressed and 10 very happy. The distribu-tion was distinctly skewed towards the upperend of the scale with a median rating of 7.

Stressors were reported by 53-5% of thesample and the number ranged from 1-4.The 119 identified stressors were distributedamong the following: work (30 3%);spouse/child (16-8%); finances (15-1%); psy-chological (12-6%); relationships (8 4%);physical (7 6%); parents (5 9%); livingarrangements (2 5%); and substance abuse(0-8%).

Outcome measuresThe subjective measure of outcome wasderived from the complaints by the respon-dents elicited during the interview whenasked whether they had noted post-accidentsequelae. The reported subjective sequelaewere then categorised as physical, intellectualand emotional. The physical complaints wereitemised in a systems schema.28 The intellec-tual complaints included difficulties withlearning, memory, intellectual functioningand slowed thinking. The emotional com-plaints included anxiety, depressive andbehavioural disorders and problems with self-esteem and feelings of rejection.The details of the sequelae are itemised in

table 5. Fifty individuals (31%) reported 96sequelae: 36 were physical, 30 intellectualand 30 emotional. The subjective sequelaewere distributed as follows: 1 -(17%), 2-(8%), 3 -(2%), 4 -(2%), 5 -(1%), and 7-(1 %). The overlap of sequelae among the 50individuals was: exclusively intellectual and/oremotional (27); intellectual and/or emotionalas well as physical ( 11); and exclusivelyphysical (12).Of the 7 physical systems, the highest

number of subjective sequelae were recordedunder neurological, with two individualsreporting seizures: one is still on anti-convul-

412

14

on April 4, 2021 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.56.4.410 on 1 A

pril 1993. Dow

nloaded from

http://jnnp.bmj.com/

Long-term outcome ofhead injuries: a twenty three yearfollow up study of children with head injuries

Table S Types of subjective sequelae (n=SO)

System Complaints Frequency

PhysicalNeurological Coordination or speed 4

Seizures 2Handedness changed 2Dizziness 1Speech 1

Musculo-skeletal Arthritis/osteoarthritis 3Back pain 2Leg/hip pain 2Leg shorter/deformity 2

Head and Neck Headaches 3Plate area sensitive 3Neck pain 1

Eyes Diplopia 2Retinal damage 1Fields/depth perception 2

Ears/Nose Tinnitus/sinusitis 2Abdomen Kidney/spleen 2Skin Sweating 1Intellectual

Learning problems/disabilities 12Memory/attention/concentration 10Intelligence/brain affected 4Thinking/problem solving slowed 4

EmotionalAnxiety disorders Anxieties, phobias, nightmares 8Depressive disorders Depression 6

Depression with suicidal thoughts 3Bereavement, breakup of family 3

Self esteem problems Insecure/self-conscious/introverted 5Parent-child problems Rejection by parent(s) 2Behaviour disorders Aggression (as child) 2

Pyromania (as child) I

sant medication and the seizures are currentlycontrolled, whereas the other is not on med-ication and does not have a recent history ofseizures. The second most frequent area ofsequelae was derived from the musculoskele-tal system, followed by sequelae regardinghead and neck.Among the intellectual sequelae, the most

frequent complaint was difficulty in leaming,followed by problems with memory or con-centration, and in turn followed by state-ments that intelligence had been adverselyaffected, and finally a slowing of thoughtprocesses. Among the emotional sequelae,depression was the most frequent complaint.Three of those reporting depression indicateda history of accompanying suicidal ideationand two were currently experiencing suicidalthoughts. Anxiety disorders followed in termsof frequency, followed by problems with selfesteem, then rejection by parents, and finallybehaviour disorders (pyromania and aggres-sion) during childhood.

There was a significant relationship(Kendall's Tau 0 19, p < 0.01) betweenphysical sequelae reported as directly relatedto the head injury and physical complaintsidentified as unrelated to the head injury.Specifically, 12-6% of the sample reportedboth physical sequelae and physical com-plaints unrelated to the trauma, 20- 1%reported physical sequelae but no physicalcomplaints unrelated to the trauma, 18-9%reported no physical sequelae but physicalcomplaints unrelated to the trauma, and48-4% reported neither. These findings areunderstandable in view of the multipleinjuries sustained by 26-4% of the sample.

Relationships between extent of head injury andoutcomeThe first analysis evaluated the relationships

between the trauma variables and subjectivesequelae (physical, intellectual and emotion-al). Table 6 summarises the correlationmatrix and as may be noted the compositemeasure was the most discriminating regard-ing long-term subjective sequelae. This inturn was followed by the neurological variableand seizures and finally fractures as predictorsof sequelae. Unconsciousness was of limitedpredictive value and EEG was of no predic-tive value.The second analysis evaluated the relation-

ship between initial IQ and subjective seque-lae, and the findings are also summarised intable 6. A comparison of those individualswho either reported or did not report seque-lae in each of the three areas revealed signifi-cantly lower IQs for those with sequelae inboth the intellectual and emotional areas.The third analysis evaluated the relation-

ship between attribute variables which pur-portedly relate to the outcome of head injuryand the findings are summarised in table 7.While only 3 of the 21 correlations were sig-nificant, this is greater than chance expectan-cy. Physical sequelae were significantlyrelated to the agent of injury and subsequentlitigation, and these are understandable. Therelationships between intellectual sequelaeand subsequent head injuries is also quitepredictable.The purpose of the fourth analysis was to

determine whether the presence of subjectivesequelae had any measurable effects uponobjective and psychosocial measures of adap-tation. The individuals reporting sequelae(n = 50) were therefore compared on desig-nated demographic, health and psychosocialvariables with individuals reporting no seque-lae (n = 109). Significant differences werefound in grade failures/retention (40% for thesequelae group vs 19% for the non-sequelaegroup, x2 = 6-65, p < -0 1), work status (12%vs 2-8% of unemployed respectively,X2 = 3-89, p < 0-05), current psychological/psychiatric problems (32% vs 11 9% respec-tively, x2 = 7-96, p < 0.01) and strained rela-tionships with family members (24% vs15-6% respectively, x2 = 3-95, p < 005). Nodifferences were found for past but resolvedpsychological/psychiatric problems, physicalcomplaints, substance abuse or contact withlegal authorities.The fifth analysis dealt with continuity of

complaints during the twenty three year inter-val between the original project and the cur-rent study. In the published five year followup project,2 parents of the 117 childrenreported on average 0-89 complaints from thefollowing: personality and mood, headachesand dizziness, memory and learning, sensory-motor, and fatigue and sleep. In the currentstudy, the 159 adults reported on average060 sequelae, and these were categorised asphysical, intellectual and emotional. Of the159 individuals included in the current study,93 also had complaints in the initial project.The informants, however, were different asduring the initial project information aboutcomplaints was provided by a parent, general-

413 on A

pril 4, 2021 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.56.4.410 on 1 April 1993. D

ownloaded from

http://jnnp.bmj.com/

Klonoff, Clark, Klonoff

Table 6 Relationships between subjective sequelae and trauma variables as well as initialIQ (Kendall's Tau and t test)

Subjective sequelae

Trauma variables Physical Intellectual Emotional

Unconscious 0-12 0 04 0-21**Fracture 0 19** 0 20** 0-20**EEG - 0-02 0 07 - 0 06Neurological 0-13* 0-21** 0 35***Seizures 0-20** 0 25*** 0 18*Composite 0 17*** 0 23*** 0 28***Initial IQ - 0 05 -0-21** - 0 17*Initial IQ non-sequelae group 103 5 104 7 104 2Initial IQ sequelae group 99 9 93 3 98 6t value 1 07 3 57** 2.37*

*p

Long-tern outcome of head injuries: a twenty three yearfollow up study of children with head injuries

so that by the fifth follow up a majority of thechildren had improved substantially. Withrespect to the reconstitution process, thequestion posed initially was whether this wasa function of the severity of head injury, indi-vidual differences in disposition, subsequentenvironment or an interaction of these vari-ables. Based on the long-term findings of thecurrent study, individual differences in dispo-sition have not been identified as particularlyrelevant. Intervening life events while inordi-nately complicated are undoubtedly interac-tive, but with the available information it isnot possible to go beyond this statement.However, the severity of the head injury hasbeen identified as the primary contributoryfactor in the reconstitution process. The cur-rent study extends this relationship in the pre-diction of even long term outcomes.Furthermore, the presence of long term sub-jective sequelae is consistent with objectiveindicators of social adaptation.

This research was supported by the Insurance Corporation ofBritish Columbia, British Columbia, Canada. We wish toexpress our appreciation to Mrs Vinetta Lunn for her creativeapproach in tracing individuals and invaluable assistance, toMr Don Gilbert for his consistent support, and to Mrs MaryKlonoff for assisting with the manuscript.

1 Klonoff H, Paris R. Immediate short-term and residualeffects of acute head injuries in children. In: Reitan R,Davison LA, eds. Clitical neuropsychology: current statusand applicatiotns. Washington: Winston, 1974:179-210.

2 Klonoff H, Low MD, Clark C. Head injuries in children:a prospective five year follow-up. Jf Neurol NeurosurgPsychiatry 1977;40:1211-9.

3 Levin HS, Grafman J, Eisenberg HM, eds.Neurobehavioral recovery fromi head injury. New York:Oxford University Press, 1987.

4 Levin HS, Benton AL, Grossman RG. Neurobehavioralconsequences of closed head injury. New York: OxfordUniversity Press, 1982.

5 Levin HS, Eisenberg HM, Benton AL, eds. Mild headinjury. New York: Oxford University Press, 1989.

6 Jennett B, Teasdale G. Managemnent of head inijuries.Philadelphia: FA Davis, 1981.

7 McLean A, Dikmen S, Temkin N, et al. Psychosocialfunctioning at one month after head injury. Neurosurgery1984;14:393-9.

8 Klonoff PS, Snow WG, Costa LD. Quality of life inpatients two to four years after closed head injury.Neurosurgery 1986;19:735-43.

9 Thomsen IV. Late outcome of very severe blunt headtrauma: a 10-15 year second follow up. J NeuirolNeurosuirg Psychiatry 1984;47:260-8.

10 Oddy M, Coughlan T, Tyerman A, et al. Social adjust-ment after closed head injury: A further follow-upseven years after injury. Jf Neurol Neuirosurg Psychiatry1985;48:564-8.

11 Costeff H, Groswasser Z, Goldstein R. Long-term follow-up review of 31 children with severe closed head trau-ma. J7 Neurosurg 1990;73:684-7.

12 Tate RL, Fenelon B, Manning ML, et al. Patterns of neu-ropsychological impairment after severe blunt headinjury. Nerv Menit Dis 1991; 179:117-26.

13 Lishman WA. Brain damage in relation to psychiatric dis-ability after head injury. Br J Psychiatry 1968;114:373-410.

14 Levin HS, Grossman RG. Behavioral sequelae of closedhead injury: a quantitative study. Arch Neuirol 1978;35:720-6.

15 Rimel RW, Giordani B, Barth JT, et al. Moderate headinjury: completing the clinical spectrum of brain trau-ma. Neurosurgery 1982;11:344-51.

16 McKinlay WW, Brooks DN, Bond MR. Post-concussion-al symptoms, financial compensation and outcome ofsevere blunt head injury. J' Neuirol Neurosurg Psvchiatry1983;46:1084-91.

17 Kreutzer JS, Wehman PH, Harris JA, et al. Substanceabuse and crime patterns among persons with traumaticbrain injury referred for supported employment. BrainInj 1991;5:177-87.

18 Lewin W, Marshall TF, Roberts AH. Long-term outcomeafter severe head injury. BMJ7 1979;2:1533-8.

19 Rimel RW, Jane JA. Characteristics of the head-injuredpatient. In: Rosenthal M, Griffith ER, Bond MR,Miller JD, eds. Rehabilitation of the head-injured aduilt.Phila: FA Davis, 1983:9-21.

20 Levin HS, Gary HE, Eisenberg HM, et al. Neuro-behavioral outcome one year after severe head injury.J7 Neurosurg 1990;73:699-709.

21 Eiben C, Anderson T, Lockman L, et al. Functional out-come of closed head injury in children and youngadults. Arch Phys Med Rehabil 1984;65:168-70.

22 Mahoney WJ, D'Souza BJ, Haller JA, et al. Long-termoutcome of children with severe head trauma and pro-longed coma. Pediatrics 1983;71 :756-62.

23 Alberico AM, Ward JD, Choi SC, et al. Outcome aftersevere head injury: relationship to mass lesions, diffuseinjury and ICP course in pediatric and adult patients.J Neurosuirg 1987;67:648-56.

24 Berger MS, Pitts LH, Lovely M, et al. Outcome fromsevere head injury in children and adolescents. _7Neurosurg 1985;62:194-9.

25 Levati A, Farina ML, Vecchi G, et al. Prognosis of severehead injuries. Neurosurg 1982;57:779-83.

26 Jennett B, Bond M. Assessment of outcome after severebrain damage. Lancet 1975;1:480-4.

27 Hendryx PM. Psychosocial changes perceived by closed-head-injured adults and their families. Arch Phvs MedRehabil 1989;70:526-30.

28 Seidel HM, ed. Mosby's guide to physical examninatinon, 2nded. St Louis: Mosby Year Book, 1991.

29 Edna TH, Cappelen J. Late postconcussional symptomsin traumatic head injury: An analysis of frequency andrisk factors. Acta Neurochir 1987;86: 12-7.

30 Klonoff H. Head injuries in children: predisposing fac-tors, accident conditions, accident proneness and seque-lae. AmJ_y Public Health 1971;61:2405-17.

31 Chadwick 0, Rutter M, Shaffer D, et al. A prospectivestudy of children with head injuries. IV: Specific cogni-tive defects. 37 Clin Neuropsychol 1981;3: 101-20.

32 Kraus JF, Fife D, Cox P, et al. Incidence, severity andexternal causes of pediatric brain injury. Am. _7 Dis Child1986;140:687-93.

33 Annegers JF. The epidemiology of head trauma in chil-dren. In: Shapiro K, ed. Pediatric head trauma. NewYork: Futura Publishing, 1983:1-10.

34 Fabian AA, Bender L. Head injury in children: predispos-ing factors. Amil Orthopsychiatn- 1947;17:68-79.

35 Salcido R, Costich JF, Conder R, et al. Recurrent severetraumatic brain injury: series of six cases. Amner _7 PhvsMed Rehabil 1991;70:215-9.

415 on A

pril 4, 2021 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.56.4.410 on 1 April 1993. D

ownloaded from

http://jnnp.bmj.com/