BRAIN INJURY AND HAZARDOUS ALCOHOL DRINKING IN TRAUMA PATIENTS

OLLISAVOLA

Department of Neurology,University of Oulu

OULU 2004

OLLI SAVOLA

BRAIN INJURY AND HAZARDOUS ALCOHOL DRINKING IN TRAUMA PATIENTS

Academic Dissertation to be presented with the assent ofthe Faculty of Medicine, University of Oulu, for publicdiscussion in the Auditorium 8 of Oulu UniversityHospital, on June 11th, 2004, at 12 noon.

OULUN YLIOPISTO, OULU 2004

Copyright © 2004University of Oulu, 2004

Supervised byProfessor Matti Hillbom

Reviewed byProfessor Kaija SeppäDocent Jukka Turkka

ISBN 951-42-7378-8 (nid.)ISBN 951-42-7379-6 (PDF) http://herkules.oulu.fi/isbn9514273796/

ISSN 0355-3221 http://herkules.oulu.fi/issn03553221/

OULU UNIVERSITY PRESSOULU 2004

Savola, Olli, Brain injury and hazardous alcohol drinking in trauma patients Department of Neurology, University of Oulu, P.O.Box 5000, FIN-90014 University of Oulu,Finland 2004Oulu, Finland

AbstractHead injury is the leading cause of death and disability in trauma patients, and alcohol misuse is oftenassociated with such injuries. Despite modern diagnostic facilities, the extent of traumatic brain injury(TBI) is difficult to assess and supplementary diagnostic tools are warranted. The contribution ofalcohol misuse to traumas also needs to be elucidated, as the role of different patterns of alcoholdrinking in particular has received less attention.

We investigated the clinical utility of a novel serum marker of brain damage, protein S100B, as atool for assessing TBI in patients with trauma. We also investigated the patterns of alcohol drinkingamong trauma patients and the trauma mechanisms in relation to blood alcohol concentration (BAC),with special emphasis on head traumas. Finally, we studied the early identification of hazardousdrinkers among trauma patients.

Serum protein S100B was found to be a feasible supplementary method for assessing TBI, as thelatter was shown to elevate its levels significantly, the highest values being found in patients withsevere injuries. S100B was also found to be elevated in patients with mild head injury, where it wasassociated with an increased risk of developing post-concussion symptoms (PCSs). Extracranialinjuries also increased S100B values in patients with multitrauma. Accordingly, S100B was notspecific to TBI. The more severe the extracranial injury, the higher the S100B value that was found.

Binge drinking was found to be the predominant pattern in trauma patients. Alcohol intoxicationon admission and hazardous drinking patterns were more often present in patients with head injurythan in those with other types of trauma. The risk of sustaining a head trauma significantly increasedwith increasing BAC. The results also demonstrated that BAC on admission is the best marker ofalcohol misuse in trauma patients. The BAC test depicts hazardous alcohol drinking better thanconventional biochemical markers of alcohol misuse such as gamma-glutamyl transpeptidase (GGT),aspartate aminotransferase (AST), carbohydrate-deficient transferrin (CDT), or mean corpuscularvolume (MCV) of erythrocytes.

The findings support the use of S100B as a supplementary method for assessing TBI and the useof BAC as a marker of alcohol misuse in trauma patients.

Keywords: alcohol misuse, brain injury, laboratory markers, patterns of alcohol drinking,post-concussion symptoms, serum protein S100B

To my family and friends

Acknowledgements

The present study was carried out at the Department of Neurology, University of Oulu,during the years 1998–2004.

While I am very grateful to everyone who has contributed to this work, I would like tothank the following persons in particular.

First I wish to express my most sincere gratitude to my supervisor, Professor MattiHillbom, Head of the Department of Neurology, who introduced me to the subject andgave his support and scientific guidance at all stages in the work. His criticism and end-less enthusiasm for medical science have made this research very pleasant. I am also verygrateful to Professor Onni Niemelä for his guidance. His broad knowledge of alcoholismhas made a great impression during these years.

I am very grateful to my co-authors Professor Juhani Pyhtinen, Tuomo K. Leino,M.D., Ph.D. and Simo Siitonen, M.D., Ph.D. for their collaboration and encouragementduring the research.

I owe particular thanks to Professor John Koivukangas, Head of the Department ofNeurosurgery, and Professor Esa Heikkinen for allowing me to use the facilities of thatdepartment, and Docent Juhani Valkama and the whole staff of the Accident and Emer-gency Department of Oulu University Hospital for their excellent co-operation.

I also owe particular thanks to Professor Kaija Seppä and Docent Jurkka Turkka, whoreviewed the manuscript of this thesis and provided constructive criticism and advice. Iwould similarly like to thank Mr Malcolm Hicks for revising the English language of themanuscript.

My warm thanks are also due to Professor Kari Majamaa for his advice in scientificdrawing. I also wish to thank all my colleagues and the whole staff of the Department ofNeurology for their sympathetic support during my work. I am especially indebted toMaarit Näppä, Merja Halonen, Mirja Kouvala, Marika Hämeenaho and Ilona Huovinenfor their practical assistance. I am also very grateful to Risto Bloigu for his statisticaladvice.

My special thanks go to my colleagues in Professor Hillbom’s research group, VesaKarttunen, Veli Tuomivaara, Satu Winqvist and Pertti Saloheimo. I will never forget thegreat conversations, more or less scientific, that we have had.

I would also like to express my sincere thanks to the patients and their relatives whoparticipated in this study.

My warm thanks go to my family and friends, especially to my parents and my sisterKaisa, for their support and care during these years.

Finally, I wish to express my dearest thanks to my wife Kaisa-Marja, for her great sup-port and love during my years of study, and to our children Isla and Nikla, who were bornduring this project.

The work was supported financially by the Oulu Medical Research Foundation.

Oulu, April 2004 Olli Savola

Abbreviations

ALT alanine aminotransferaseAST aspartate aminotransferaseAUDIT Alcohol Use Disorders Identification TestBAC blood alcohol concentrationCDT carbohydrate-deficient transferrinCT computed tomographyDAI diffuse axonal injuryEDH epidural haematomaEEG electroencephalographyGCS Glasgow Coma ScaleGGT gamma-glytamyl transpeptidaseGOS Glasgow Outcome ScaleLOC loss of consciousnessMAST Michigan Alcohol Screening TestMCV mean corpuscular volume (of erythrocytes)MHI mild head injuryMRI magnetic resonance imagingNMDA N-methyl-D-aspartateNSE Neuron-specific enolasePCSs post-concussion symptomsPTA post-traumatic amnesiaRGA retrograde amnesiaSAAST Self-Administered Alcoholism Screening TestSDH subdural haematomaSMAST Short Michigan Alcohol Screening TestSPECT single photon emission computed tomographyTBI traumatic brain injury

List of original publications

This thesis is based on the following articles, which are referred to in the text by theirRoman numerals:

I Savola O & Hillbom M (2003) Early predictors of post-concussion symptoms inpatients with mild head injury. Eur J Neurol 10: 175–181.

II Savola O, Pyhtinen J, Leino TK, Siitonen S, Niemelä O & Hillbom M (2003) Effectsof head and extracranial injuries on serum protein S100B levels in trauma patients. JTrauma, in press.

III Savola O, Niemelä O & Hillbom M (2004) Binge drinking as a major risk factor forhead trauma. Submitted for publication.

IV Savola O, Niemelä O & Hillbom M. (2004) Blood alcohol is the best indicator ofhazardous alcohol drinking in young adults and working-age patients with trauma.Alcohol and Alcoholism, in press.

Reprints are included with the permission of the publishers: Blackwell Science Ltd (I),Lippincott Williams & Wilkins (II) and Oxford University Press (IV).

Contents

Abstract Acknowledgements Abbreviations List of original publications Contents1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172 Review of the literature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

2.1 Epidemiology of head injury . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192.2 Classification of head injuries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

2.2.1 Mild head injury (MHI) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212.2.1.1 MHI without brain injury . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212.2.1.2 MHI with mild brain injury . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

2.2.2 Head injury with moderate brain injury . . . . . . . . . . . . . . . . . . . . . . . . . 222.2.3 Head injury with severe brain injury . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

2.3 Clinical features of brain injury . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232.3.1 Diffuse axonal injury (DAI) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232.3.2 Cerebral contusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242.3.3 Intracerebral haematoma (ICH) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242.3.4 Epidural haematoma (EDH) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252.3.5 Subdural haematoma (SDH) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252.3.6 Subarachnoid haemorrhage (SAH) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 262.3.7 Secondary brain injury . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272.3.8 Skull and facial bone fracture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

2.4 Diagnosing brain injury in patients with head injury . . . . . . . . . . . . . . . . . . . . 282.4.1 Clinical examination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 282.4.2 Skull radiography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 292.4.3 Computed tomography (CT) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 292.4.4 Magnetic resonance imaging (MRI) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 312.4.5 Single photon emission computed tomography (SPECT) . . . . . . . . . . . . 322.4.6 Electroencephalography (EEG) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 322.4.7 Biochemical markers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

2.4.7.1 Neuron specific enolase (NSE) . . . . . . . . . . . . . . . . . . . . . . . . . 332.4.7.2 Protein S100B . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

2.4.8 Neuropsychological assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 352.5 Treatment of head injury . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 362.6 Outcome after head injury . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

2.6.1 Mild head injury . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 382.6.1.1 Post-concussion symptoms (PCSs) . . . . . . . . . . . . . . . . . . . . . . 38

2.6.2 Moderate-to-severe head injury . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 392.7 Alcohol and head injury . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

2.7.1 General aspects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 412.7.2 Alcohol drinking as a risk factor for trauma . . . . . . . . . . . . . . . . . . . . . . 412.7.3 Alcohol drinking as a risk factor for head injury . . . . . . . . . . . . . . . . . . 422.7.4 Alcohol and trauma morbidity and mortality . . . . . . . . . . . . . . . . . . . . . 422.7.5 Alcohol and traumatic brain injury morbidity and mortality . . . . . . . . . 42

2.8 Identification of hazardous alcohol drinking . . . . . . . . . . . . . . . . . . . . . . . . . . 432.8.1 Tools for identifying hazardous alcohol drinking . . . . . . . . . . . . . . . . . . 44

2.8.1.1 Questionnaires and interviews . . . . . . . . . . . . . . . . . . . . . . . . . . 442.8.1.2 GGT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 442.8.1.3 AST & ALT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 442.8.1.4 CDT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 452.8.1.5 MCV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 452.8.1.6 BAC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46

2.8.2 Detecting hazardous alcohol drinking . . . . . . . . . . . . . . . . . . . . . . . . . . . 462.8.2.1 In patients with trauma . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 462.8.2.2 In patients with head injury . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46

2.9 Alcohol interventions as means of preventing injuries . . . . . . . . . . . . . . . . . . 473 Aims of the research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 484 Subject and methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

4.1 Subjects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 494.2 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50

4.2.1 Clinical examination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 504.2.1.1 Emergency room . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 504.2.1.2 Alcohol data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 514.2.1.3 Follow-up interviews . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

4.2.2 Laboratory procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 524.2.3 Statistical analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

5 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 545.1 Protein S100B as a predictor of PCSs in MHI patients (paper I) . . . . . . . . . . . 545.2 Effects of extracranial injuries on S100B levels (paper II) . . . . . . . . . . . . . . . 565.3 Binge drinking as a risk factor for head trauma (paper III) . . . . . . . . . . . . . . . 575.4 Laboratory markers of hazardous alcohol drinking in trauma patients

(paper IV) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 596 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60

6.1 PCSs in patients with mild head injury . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 606.2 Protein S100B as a marker of brain damage . . . . . . . . . . . . . . . . . . . . . . . . . . 616.3 Alcohol and trauma . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62

6.3.1 Hazardous alcohol drinking in patients with trauma . . . . . . . . . . . . . . . . 62

6.3.2 Alcohol as a risk factor for head trauma . . . . . . . . . . . . . . . . . . . . . . . . . 636.3.3 Identification of alcohol misuse in patients with trauma . . . . . . . . . . . . 64

6.4 Strengths and weaknesses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 647 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

References

1 Introduction

Head injury is a serious health burden, being a leading cause of death and disabilityamong teenagers and young adults, and resulting in enormous financial costs for society(Alexander 1995, Thorhill et al. 2000, Thompson et al. 2001). Alcohol is a major causeof traumas of this kind, and alcohol misuse and its consequences are currently one of themajor health problems world-wide (Lieber 1995, Hadfield et al. 2001).

Despite modern diagnostic and monitoring methods, it remains difficult to assess theextent of primary brain damage after head injury and the development of secondary dam-age (Raabe et al. 1999b). Most of the diagnostic process nowadays is based on neuroim-aging techniques such as computed tomography (CT) and magnetic resonance imaging(MRI), although in practice only CT is often available in the acute stage. CT is sensitivefor detecting fresh blood inside the skull, but has a low sensitivity for diffuse axonalinjury (DAI) (Bešenski 2002). There is a major need for supplementary diagnostic toolsin cases of acute head injury, e.g. a simple biochemical marker of brain damage.

The protein S100B is the most promising candidate for a biochemical marker of braindamage to date (Ingebrigtsen & Romner 2002), having been found to correlate with theextent of brain damage and the outcome after severe head injury (Raabe 1999b). It hasalso been shown to be elevated in approximately 20–38% of patients with mild headinjury (MHI) (Ingebrigtsen et al. 1995, Ingebrigtsen et al. 2000b). Its clinical utilityamong such patients is unresolved, however. S100B is also expressed in non-nervous tis-sues, and extracranial injuries may confound its use as a specific marker of brain damage(Haimoto et al. 1987, Anderson et al. 2001). The contribution of extracranial injuries tothe elevation in S100B levels observed after head injury also remains to be elucidated.

Most patients with MHI recover well (Binder 1997), although a small but significantproportion develop persistent post-concussion symptoms (PCSs) that may prevent opti-mal recovery and impair return to work and psychosocial functioning (Alexander 1995,Bernstein 1999). Routine follow-up of all patients with MHI is not warranted (Wade et al.1997), but it might be beneficial and cost-saving to follow up subjects who are likely todevelop PCSs. After identification of these patients, appropriate rehabilitation servicescould prevent the condition from becoming chronic and avoid financial costs and socialand human suffering. No methods have been proposed to date for the early detection ofpatients at risk of developing PCSs.

18

Alcohol and injuries are closely linked (Rivara et al. 1993a, Corrigan 1995, Porter2000). Approximately 50% of trauma patients have alcohol in their blood on admission,and even higher proportions have been reported in series consisting specifically ofpatients with head injuries (Brismar et al. 1983, Rivara et al. 1993a, Dikmen et al. 1995).Chronic alcohol misuse is also common in trauma patients, approximately 45% of whomhave been reported to be misusers (Nilssen et al. 1994). The prevalence of chronic alco-hol misuse among head injury patients has varied from 16% to 66% (Corrigan 1995).

Little is known of the different patterns of alcohol drinking in patients with trauma,and more attention should be paid to binge drinking in particular, because the dangers ofalcohol are not restricted to regular drinking. The social, health and economic costs ofacute alcohol-related problems may even exceed those due to chronic drinking(Chikritzhs et al. 2001). It would be important to know the prevalence of different pat-terns of alcohol drinking and their contribution to traumas, and whether a certain drink-ing pattern is associated specifically with head injuries.

There have been several reports of a positive effect of brief alcohol intervention asmeans of reducing alcohol intake and injury recurrence (Walsh et al. 1991, Fleming et al.1997, Gentilello et al. 1999), but there continues to be a lack of attention paid to alcoholmisuse in accident and emergency departments, so that patients with alcohol misuse tendto escape specific treatment. This is likely to be due to the lack of practical tools for iden-tifying the optimal target group for alcohol interventions. Laboratory tests such asgamma-glutamyl transpeptidase (GGT), aspartate aminotransferase (AST), carbohydrate-deficient transferrin (CDT) and mean corpuscular volume (MCV) of erythrocytes havebeen suggested to help the physician to detect alcohol misusers on admission (Mihas &Tavassoli 1992), but their clinical utility remains to be elucidated. In particular, noattempt has been made to correlate laboratory markers of alcohol consumption with pat-terns of alcohol consumption in trauma patients.

The present study was aimed at assessing the clinical utility of S100B as a specificmarker of brain damage and as a factor that predicts poor recovery after MHI, and also atelucidating the role of different drinking patterns in trauma patients and how alcoholdrinking increases the risk of head trauma. Methods were sought for identifying hazard-ous drinkers among trauma patients, in the hope that improved identification and assess-ment of brain damage and alcohol misuse may lead to a better understanding of the risksof injury and of a poor outcome after injury and may help us to develop treatments andpreventive medical and social policies.

2 Review of the literature

2.1 Epidemiology of head injury

The annual incidence of head injury is estimated to be between 150–300 per 100 000 ofpopulation (Annegers et al. 1980, Jennett & MacMillan 1981, Klauber et al. 1981, Krauset al. 1984, Tiret et al. 1990, Durkin et al. 1998, Marshall 2000). Men have approximate-ly twice as high a rate as women (Kraus et al. 1984, Lundar & Nestvold 1985), and thehighest age-specific incidence is found among young adults (Kraus 1993).

These incidence figures are based on studies of hospital discharges after diagnoses ofbrain injury or at least skull fracture (Klauber et al. 1981, Kraus 1993), and are likely tobe underestimated. Thornhill et al. (2000) reported that 20% of all head injury patientsadmitted to hospital were not recorded in the health service statistics, and found the totalincidence of such admissions to be 326 per 100 000 of population. McGuire et al. (1995)reported an incidence of 607 per 100 000 of population when all medically treated headinjury patients were taken into account, but according to an extensive survey by Fife(1987), approximately 10% of patients who reported having had a head injury were nevertreated for it medically. Finally, Segalowitz & Lawson (1995) found that approximatelyone third of high school and university students had a history of at least one head injury,and only one fifth of them had been admitted to hospital.

MHI explains approximately 80–90% of all head injuries, while moderate and severehead injuries explain the rest (McGuire et al. 1995, Thornhill et al. 2000). The annualincidence of fatal head injuries is consistently found to be between 17–30 per 100 000 ofpopulation (Kraus et al. 1984, Sosin et al. 1989, Tiret et al. 1990). Almost 60% of all fataltrauma cases admitted to hospital are due to head injury (Gennarelli et al. 1989), and traf-fic accidents account for approximately 60% of all severe head injuries (Sosin et al. 1989,Tiret et al. 1990), while falls and assaults are the most common causes of milder headinjuries (Thornhill et al. 2000).

Disability due to head injury is a poorly studied entity. Kraus (1993) estimated the rateof disability, i.e. severe or moderate disability according to the GOS (Jennet & Bond(1975), to be 33 to 45 per 100 000 of population/year, but a recent study from Glasgow

20

area has reported the annual incidence to be much higher, 154 per 100 000 of population(Thornhill et al. 2000).

The financial costs associated with head injury are enormous. Max et al. (1991)assessed the total costs per year in the US as exceeding $38 billion. Thompson and herco-workers (2001) suggested that the estimate presented by Max et al. (1991) is conserva-tive, capturing only part of the total costs, and estimated that the figure may well exceed$62 billion. In fact, the true financial costs due to head injuries are probably incalculable,as family members may develop physical and mental symptoms, for example, and theoverall financial impact is unknown (Thompson et al. 2001).

2.2 Classification of head injuries

Head injuries are traditionally classified as mild, moderate or severe (Vos et al. 2002), thepurpose being to evaluate the risk of complications and assess the probable outcome(Ingebrigtsen et al. 2000a, Thornhill et al. 2000). In practice, the classification is oftenperformed by assessing the level of impaired consciousness on the Glasgow coma scale(GCS), which is based on eye opening and motor and verbal responses, as presented inTable 1 (Teasdale & Jennett 1974). MHI is defined as a state with scores of 13 to 15,whereas moderate and severe head injuries are defined by scores of 9 to 12 and 3 to 8,respectively. In addition, the presence and duration of post-traumatic amnesia (PTA) areoften used as tools for classifying head injuries (Russell & Smith 1961, Kurup et al.2000), and a head CT scan-based classification has also been developed especially formoderate-to-severe head injuries (Marshall et al. 1991).

Table 1. The Glasgow Coma Scale.

Aspect of behaviour Response scoreEye-opening (E):

Spontaneous 4To voice 3To pain 2No response 1

Best motor response (M):Obeys commands 6Localizes 5Withdraws (flexion) 4Abnormal flexion (posturing) 3Extending (posturing) 2No response 1

Verbal response (V):Oriented conversation 5Confused, disoriented 4Inappropriate words 3Incomprehensible sounds 2

No response 1Glasgow Coma Scale score = E + M + V

21

The International Classification of Diseases (ICD) (1992) can also be used for classify-ing head injuries. This includes diagnoses indicating the clinical features of head andbrain injury, but does not classify them in terms of severity. Its clinical utility is thus lim-ited when assessing the possible outcome, for example. The ICD classification has provedto be practicable for epidemiological studies, however (Sosin et al. 1989). Head injuriescan also be classified in other ways, into impact versus deceleration/acceleration injuries(Elson & Ward 1994, Aikuisiän aivovammojen käypä hoito 2003), or penetrating versusclosed head injury (Bayston et al. 2000).

2.2.1 Mild head injury (MHI)

MHIs constitute a significant medical problem for several reasons. Firstly, approximately80 to 90% of all head injuries are mild (McGuire et al. 1995, Thornhill et al. 2000). Sec-ondly, 0.01 to 3% of patients initially assessed as having MHI will develop life-threaten-ing complications requiring immediate neurosurgical intervention (Dacey et al. 1986,Stein & Spettell 1995). Thirdly, great attention and financial resources have to be spenton the evaluation of these patients because of this minimal risk of possible intracranialcomplications (Stiell et al. 2001a). Fourthly, a small proportion of MHI patients do notachieve a good recovery (Thornhill et al. 2000) and may be left neuropsychologicallyimpaired (Binder et al. 1997) or develop long-lasting PCSs (Alexander 1995).

2.2.1.1 MHI without brain injury

Although there certainly exist head injuries without brain injury, the terms MHI and mildTBI are often used synonymously (de Kruijk et al. 2001b). This confusion is due to thefact that there are no specific measures available for detecting the presence of negligiblebrain damage or for separating head injury patients with mild TBI from those without(Bernstein 1999). In two current articles, head injury patients without TBI are character-ised as having a GCS score of 15 and no loss of consciousness (LOC), PTA, focal neuro-logical deficits, or physical symptoms such as headache, dizziness or vomiting (de Kruijket al. 2001b, Vos et al. 2002).

2.2.1.2 MHI with mild brain injury

MHI with mild brain injury applies to a subset of patients with MHI in whom brain dam-age is evident but mild. These patients have a GCS score of 13 to 15 on admission to hos-pital and may also have one or more symptoms and/or signs of TBI. The presence of PTA(duration < 24 hours) and unconsciousness (duration < 30 minutes) are such signs (Rimelet al. 1981, Alexander 1995, Bernstein 1999, Aikuisiän aivovammojen käypä hoito

22

2003), and the most frequent symptoms are headache, nausea, vomiting and dizziness (deKruijk et al. 2001a, de Kruijk et al. 2001b). Patients having a focal neurological deficit orstructural damage visible in a head CT scan or MRI are assessed as having more severeTBI (Williams et al. 1990, Alexander 1995, Aikuisiän aivovammojen käypä hoito 2003).

The terms commotio cerebri and cerebral concussion, which are often used in clinicalwork to denote mild TBI, were initially defined as referring to a state of functional distur-bance of the brain due to trauma without any tissue damage (Ward 1966). Controversyexists, however, over whether diffuse brain dysfunction can occur without structural dam-age, because it has become clear that brain damage is often present in patients with cere-bral concussion (Ommaya & Gennarelli 1974, Vos et al. 2002). It is currently suggested,therefore, that the severity of TBI forms a continuum from negligible to clinically signifi-cant, and there is simply less brain damage in mild TBI than in more severe cases (Alex-ander 1995, Ingebrigtsen 1998).

Mild TBIs can be further subclassified into several grades (Dacey et al. 1993, Inge-brigtsen et al. 2000, Kurup et al. 2000, Vos et al. 2002) in order to assess the risk ofintracranial complications and to predict the outcome (Ingebrigtsen et al. 2000, Vos et al.2002). These grades are also frequently used in sports medicine as a guide to when it isappropriate to return to action (Kurup et al. 2000).

2.2.2 Head injury with moderate brain injury

Russell & Smith (1961) proposed a four-grade scale of TBI severity based on the durationof PTA, moderate TBI being defined as a state with PTA between 1 and 24 hours. Afterthe introduction of the GCS, moderate TBI has been defined as a GCS score of 9 to 12 onadmission (Teasdale & Jennett 1974, Rimel et al. 1982). A current Finnish guidelinedefines moderate brain injury as a GCS score of 9 to 12 and/or PTA between 24 hoursand seven days (Aikuisiän aivovammojen käypä hoito 2003).

Moderate TBI patients may have focal deficits in clinical examination (Aikuisiän aivo-vammojen käypä hoito 2003), one third actually have structural lesions in the brain visi-ble on a head CT scan, and MRI may reveal some additional lesions (Rimel et al. 1982,Kelly et al. 1988, Stein & Spettell 1995, Ingebrigtsen et al. 2000a). Moderate TBIpatients without a lesion in the brain are considered to have diffuse axonal injury (DAI)(Rimel et al. 1982). Approximately one tenth of patients with moderate TBI need a neuro-surgical operation (Rimel et al.1982).

2.2.3 Head injury with severe brain injury

Patients are deemed to have severe TBI if they have an admission GCS score of 8 or less(Teasdale & Jennett 1974), although some who “talk and deteriorate” may have a GCSscore of over 8 initially or at some moment (Lobato et al. 1991). Severe TBI can also becharacterised by a duration of PTA of at least 7 days (Aikuisiän aivovammojen käypähoito 2003). Some investigators recognise a separate group of critical brain injuries (Vos

23

et al. 2002) consisting of patients who have a GCS score of 3 to 4 and evident brain stemdamage, characterised by diminished or no pupillary reactions and decerebrate motorposturing, or an absence of pupillary and motor reactions (Vos et al. 2002).

Approximately 5 to 30% of patients with severe TBI have a normal head CT scan ini-tially (Lobato et al. 1986, Eisenberg et al. 1990, Stein & Spettell 1995) with lesions fre-quently seen later, or else brain atrophy may develop during the subsequent months, indi-cating that the initial trauma type had been DAI (Lobato et al. 1986).

2.3 Clinical features of brain injury

2.3.1 Diffuse axonal injury (DAI)

Brain injuries are commonly divided into focal and diffuse injuries (Gennarelli 1993).Diffuse brain injury is a widespread or global disruption of brain tissue and often involvesonly microscopically visible lesions (Adams et al. 1989). Typical microscopic findingsare small focal brain lesions, clusters of microglia in the white matter, axonal bulbs, andaxonal swelling and degeneration (Adams et al. 1989). This is called diffuse axonal inju-ry (DAI) (Povlishock et al. 1983, Maxwell et al. 1997).

Strich originally reported two autopsy studies of closed head injury patients, all ofwhom had died in a comatose condition (Strich 1956, Strich 1961). Her postmortem find-ings showed that there were only a few lesions visible to the naked eye, mostly in the cor-pus callosum or in one or both of the superior cerebellar peduncles. The cerebral cortexappeared normal. In contrast, Strich (1961) found a widespread diffuse degeneration ofthe white matter in all cases, and concluded that this was due to mechanical stresses andstrains which lead the nerve-fibres to tear and degenerate.

More recently, DAI has been found to be a common type of brain damage both inexperimental studies of primates (Jane et al. 1985, Xiao-Sheng et al. 2000) and in studiesdealing with humans (Adams et al. 1982). DAI is often present in cases of severe TBI,Adams et al. (1989) reporting that 122 out of 434 non-missile head injury patients autop-sied had some evidence of DAI, although in 24 cases the DAI could be found only aftermicroscopic examination. Oppenheimer (1968) also found DAI in five autopsied mildTBI patients, although the majority of the evidence for an association between mild TBIand DAI has come from experimental studies (Gennarelli et al. 1982, Jane et al. 1985).Povlishock et al. (1983) demonstrated that mild TBI does not always disrupt the axonsmechanically but can lead to axonal swelling, it is currently thought that this non-disrup-tive axonal injury may result in secondary DAI some hours after injury, and that such asituation may be more common than was previously appreciated (Maxwell et al. 1997).

Although small lesions in the white matter, corpus callosum, or brain stem can bedetected in the head CT scan in some cases of more severe DAI (Zimmerman et al. 1978,Levi et al. 1990), CT is in general insensible for detecting DAI, whereas MRI may bemore sensitive (Bagley 1999). Thus Mittl et al. (1994) found some evidence of DAI in30% of MHI patients with normal CT scans, and it is likely that there are a lot of head

24

injury patients with DAI who do not have any detectable CT lesions (Adams et al. 1989,Bagley 1999), although neuroimaging may reveal diffuse brain atrophy during follow-up(Ito et al. 1997).

2.3.2 Cerebral contusion

Cerebral contusions are bruises and lacerations of the brain tissue which may either belimited to the cortex or extend into the white matter and typically include haemorrhage,infarction, necrosis and oedema (Adams 1975, Levin et al. 1988). Cerebral contusionshave traditionally been described as a result of coup (at the site of impact) or contracoupinjuries (opposite the site of impact, where the brain strikes the skull) (Jallo & Narayan2000). Cerebral contusions are frequent in patients with TBI, and are mostly found in thetemporal and frontal regions of the brain (Mattson & Levin 1990, Gentry et al. 1988a,Stiell et al. 2001b), the areas especially prone to injury being the anterior and inferiorportions of the temporal and frontal lobes, where the brain contacts the rough surfaces ofthe sphenoid wings, the petrous ridges, the cribriform plate and the planum sphenoidale(Bagley 1999), although cerebral contusions regularly occur in other regions of the brainas well (Hesselink et al. 1988). Cerebral contusions may swell and cause increased intrac-ranial pressure (Bullock et al. 1991).

Contusions are characterised in head CT scans by areas of hypodensity, hyperdensityand oedema (Bešenski 2002). If there are haemorrhagic components, CT may be slightlymore sensitive for detecting cerebral contusions than MRI (Hesselink et al. 1988),whereas MRI is more sensitive in depicting non-haemorrhagic contusions (Gentry et al.1988b, Hadley et al. 1988), contusions in the brain stem (Gentry et al. 1988b) and contu-sions at the subacute and chronic stages (Groswasser et al. 1987, Kelly et al. 1988).

The prognosis for patients with cerebral contusions varies according to the localiza-tion and size of the lesions. MHI patients with a small frontal lobe contusion may haveonly minor neuropsychological disturbances (Mattson & Levin 1990), whereas mortalityamong patients with multiple or extensive contusions may exceed 40%, which is the over-all mortality for severe TBI patients (Murray et al. 1999).

2.3.3 Intracerebral haematoma (ICH)

Traumatic ICH, bleeding in the brain parenchyma (Jallo & Narayan 2000), may developprimarily as a result of trauma or secondarily within the subsequent days after injury(Clifton et al. 1980, Soloniuk et al. 1986, Mertol et al. 1991). Cerebral contusions andtraumatic intracranial haematomas represent a continuum, so that if there is a clear bloodclot present this is called an intracranial haematoma (Jallo & Narayan 2000). There arealso non-traumatic ICHs, which may be spontaneous or due to bleeding from vascular orarterial abnormalities (Caplan 1988).

As with contusions, most traumatic ICHs are located in the temporal and frontal lobes(Adams 1975). Traumatic ICH is seldom found alone, but is often associated with contu-

25

sions, other haematomas and oedema (Caroli et al. 2001). Siddique et al. (2002) foundthat only 90 out of 5686 head injury patients had isolated ICH. A head CT scan is sensi-tive for detecting intracranial haematomas, and it also works as a tool for assessing theneed for surgical treatment (Bullock et al. 1989). An unfavourable outcome after ICH isassociated with the presence of other brain lesions, a prolonged increase in intracerebralpressure and more advanced age (Caroli 2001, Siddique et al. 2002), and mortality fol-lowing traumatic ICH still remains high, varying from 11 to 50% (Soloniuk et al. 1986,Mertol et al. 1991, Caroli et al. 2001).

2.3.4 Epidural haematoma (EDH)

The most common locations for EDH, bleeding between the skullbone and dura (Adams1975), are the temporal and frontal areas of the skull, often with the middle meningealartery as the source (Reale et al. 1984). The typical patient with EDH is a young adultmale (Reale et al. 1984, Miller et al. 1990). Skull fractures significantly increase the riskof EDH (Miller et al. 1990), which usually develops rapidly after trauma. Some patientsmay also have delayed EDHs, however, and most of the recurrent haematomas followingcraniotomy for traumatic intracranial mass are EDHs (Bullock et al. 1990).

The classic symptoms of EDH include dilatation of the pupil on the haematoma sideand hemiparesis of the opposite side. EDH is currently diagnosed from a head CT scan,by the typical appearance of a high-density, biconvex, lentiform, extra-axial mass, oftenbounded by suture lines (Bagley 1999). It may have underlying oedema, contusions, andother haematomas associated with it. Large EDHs may also have a mass effect, character-ised on a CT scan by sulcal effacement, midline shift and compression of the ventricularsystem, and in the most severe cases herniation occurs due to increasing intracerebralpressure (Adams 1975, Stieg & Kase 1998).

Since EDH may develop even after rather low-energy trauma (Stein & Spettell 1995),considerable attention and financial resources have been devoted to the identification ofcases that are likely to develop a haematoma after MHI. Ninety-eight per cent of head CTscans of MHI patients are negative for EDH (Stiell et al. 2001a). The primary therapy foracute EDH is urgent surgery (Reale et al. 1984, Miller et al. 1990), and the outcome isrelated to the severity of concomitant brain injuries, neurological status at the time of sur-gery, age and the interval between injury and surgery (Reale et al. 1984, Servadei 1997).Mortality due to EDH varies between 16 and 40% (Reale et al. 1984).

2.3.5 Subdural haematoma (SDH)

SDH, a collection of blood under the dura, is usually attributed to the rupture of smallbridging veins, ruptured vessels at the site of contusions, or tearing of small branches ofthe cerebral arteries, particularly the middle cerebral artery (Adams 1975). SDHs aredivided into acute, subacute and chronic types.

26

Acute SDH is present in approximately one fifth of patients with severe TBI, andalmost 60% of all primarily evacuated traumatic mass lesions among these patients areSDHs (Marshall et al. 1991a). The majority of patients with acute SDHs have concomi-tant focal lesions such as contusions and ICHs or DAI (Tandon 2001).

Subacute subdural haematoma is found between a few days and 2 weeks after injury,whereas chronic SDH may be found weeks or even months afterwards (Adams 1975).Fogelholm & Waltimo (1975) reported the incidence of chronic SDH to be 1.72 per100 000 of population, increasing with age to 7.35 per 100 000 of population at 70–79years. The most frequent additional risk factors for chronic SDH are coagulopathy andalcoholism (Chen & Levy 2000).

The diagnostic examination of choice for SDHs is a head CT scan. An acute SDH istypically visible as a high-density, crescentic collection with concave inner borders whichis not confined by the cranial sutures (Bagley 1999). Acute hyperdense blood becomesisodense within 4 to 7 days and hypodense within 1 to 2 weeks, after which the densityincreases again within 3 to 12 weeks because of fresh haemorrhage, until a reduction inX-ray attenuation is brought about during the subsequent healing process (Stoodley &Weir 2000). Subacute SDHs may occasionally be of mixed density. Isodense SDHs maybe difficult to diagnose on an unenhanced CT, and such haematomas are best detectedwith MRI (Stieg & Kase 1998).

Surgery is the primary treatment for SDHs, being indicated when the thickness of theblood clot exceeds 10 mm. Chronic SDHs can often be drained through a burrhole,whereas craniotomy is favoured in acute cases (Tandon 2001). Mortality from acute SDHis high. In a series of 1150 patients with severe head injuries, 101 had acute SDH, andmortality among these exceeded 66% (Wilberger et al. 1991). The extent of the underly-ing brain injury is nevertheless thought to be more important than the SDH clot itself indictating the outcome (Wilberger et al. 1991).

2.3.6 Subarachnoid haemorrhage (SAH)

SAH can occur in various clinical contexts, the most common of which are head traumaand spontaneous rupture of an intracranial aneurysm (Stieg & Kase 1998). TraumaticSAH results from the disruption of small subarachnoidal vessels or direct extension intothe subarachnoidal space by a contusion or haematoma.

Traumatic SAH is usually diagnosed by means of a head CT scan, and it is oftenpresent in patients with severe TBI, the proportion varying from 33% to 39% (Eisenberget al. 1990, Kakarieka et al. 1995, Servadei et al. 2002). The presence of traumatic SAHis often associated with a poor outcome. Eisenberg et al. (1990) reported that head injurypatients with SAH had a two-fold increase in the risk of death relative to those withoutSAH, and other investigators have arrived at similar findings, suggesting that the pooroutcome may be attributed to vasospasm due to the haemorrhage (Kararieka et al. 1995).Servadei et al. (2002) have recently reported a similar incidence of SAH among patientswith head injury, a similar risk of death and a generally poor outcome, but showed inaddition that death among patients with traumatic SAH was related to the severity of theinitial mechanical brain damage rather than to the effects of delayed vasospasm and sub-

27

sequent secondary brain damage. Patients with traumatic SAHs more often had ICHs andmultiple parenchymal lesions than did those without SAH (Servadei et al. 2002).

2.3.7 Secondary brain injury

Traumatic brain injuries can be divided into primary and secondary types. Secondarybrain injury is important because much can be done to minimise it, whereas primary inju-ry is unavoidable (Bullock et al. 1996, Maas et al. 1997).

Secondary brain injury may occur at any time after the initial impact, and can be eithersystemic or intracranial (Maas et al. 1997). The causes of intracranial secondary braininjury are infection, cerebral ischaemia, and seizures (Maas et al. 1997, Lindsay & Bone2001), while the systemic insults are hypoxaemia, hypotension, hypercapnia, hypocap-nia, hyperthermia, hyperglycaemia, hypoglycaemia, and hyponatraemia (Maas et al.1997).

Cerebral ischaemia, which is the leading cause of secondary brain damage after TBI(Graham et al. 1989, Maas et al. 1997), may be due to systemic hypoxia or hypotensionor impaired cerebral perfusion due to increased intracranial pressure (Maas et al. 1997).Mass lesions, brain swelling and oedema are all able to raise intracranial pressure.Increased intracranial pressure may not only lead to cerebral ischaemia but also to tonsil-lar or tentorial herniation (Lindsay & Bone 2001). Brain swelling is thought to be due toincreased intravascular blood within the brain, whereas brain oedema refers to a specificsituation in which there is an increase in extravascular brain water (Gennarelli 1993).Mass lesions such as EDHs, SDHs and ICHs may develop secondarily within minutes tohours after trauma, but also within the next few days (Bullock et al. 1990, Mertol et al.1991). Infections may develop after a dural tear, which provides a route for infection intothe brain (Bayston et al. 2000). Meningitis and cerebral abscess are the clinical features ofsecondary infection.

2.3.8 Skull and facial bone fracture

Brain injuries are often associated with skull fractures, either linear or depressed, or frac-tures of the facial bones. The incidence of skull fractures in MHI patients varies from 3%to 11% (Dacey et al. 1986, Stiell et al. 2001b), while Rimel et al. (1982) reported that24% of their patients with moderate TBI had a skull fracture. Correspondingly, Shapiro etal. (2001), reviewing all trauma admissions during an 11-year period at a level I traumacentre, found that 11% of the patients had a facial bone fracture.

Skull or facial bone fracture does not directly mean brain injury. Williams et al. (1990)even suggested that TBI patients with a depressed skull fracture without an underlyinglesion can be classified as mild. Patients with a skull fracture have an increased risk ofintracranial haematoma, however (Mendelow et al. 1983), and it has been estimated thatapproximately one out of every four head injury patients with a skull fracture has an

28

intracranial haematoma and half of the patients who develop a haematoma have a fracture(Ingebrigtsen et al. 2000a).

Basal skull fractures can be suspected if the patient has nasal cerebrospinal rhinorroea,bilateral periorbital haematoma, subconjunctival haemorrhage, bleeding through a torntympanic membrane, or bruising over the mastoid (Lindsay & Bone 2001). Most skulland facial bone fractures are diagnosed by skull radiography or a head CT scan, is the lat-ter currently being a modality of choice if a skull fracture is suspected (Hofman et al.2000, Salvolini 2002, Vos et al. 2002).

2.4 Diagnosing brain injury in patients with head injury

2.4.1 Clinical examination

A clinical neurological examination forms the basis for assessing head injury patients onadmission, as this can reveal brain injury and help the physician to estimate the outcome(Maas et al. 1997, Vos et al. 2002).

Although typical symptoms such as headache, neck pain, nausea, dizziness and vomit-ing symptoms are frequent among TBI patients shortly after trauma, they are not specificto brain injury, as they may be due to a mixture of brain injury, peripheral vestibularinjury and injury to the soft and bony tissues of the head or neck (de Kruijk et al. 2001b).

The relationship between amnesia and unconsciousness in patients with TBI isdepicted in Figure 1. The assessment and follow-up of impaired consciousness and amne-sia plays an important role in such cases as these features are known to correlate with theseverity of TBI (Russel & Smith 1961, Murray et al. 1999, van der Naalt et al. 1999). TheGCS (Table 1) is currently the most widely used tool among clinicians and investigatorsfor assessing coma and impaired consciousness (Teasdale & Jennett 1974), but as it wasinitially developed for patients with severe head injury, there is a wide heterogeneity inthe severity of TBI among patients with an admission GCS score of 13 to 15 (Williams etal. 1990, Ingebrigtsen et al. 2000a). It has been suggested that the length of PTA may cor-relate better with the outcome among moderate-to-mild TBI patients (van der Naalt et al.1999), and there are few validated questionnaires available to assess the length of PTA(Fortuny et al. 1980, Shores et al. 1986, King et al. 1997). In any case, both GCS andPTA have their limitations, since they are vulnerable to confounding factors such as alco-hol intoxication (Galbraith et al. 1976, Wrightson & Gronwall 1981, Jagger et al. 1984).Absence of the pupillary light reflex and/or oculocephalic reflex was also found to beassociated with a poor outcome after severe head trauma (Attia et al. 1998). Expandingintracranial lesions such as EDH may be revealed by the absence of a pupillary lightreflex (Lindsay & Bone 2001). Basal skull fractures can also be identified by clinicalexamination if the patient has nasal cerebrospinal fluid rhinorroea, bilateral periorbitalhaematoma, subconjunctival haemorrhage, bleeding through a torn tympanic membrane,or bruising over the mastoid.

29

Fig. 1. Relationship between amnesia and loss of consciousness in patients with TBI.

2.4.2 Skull radiography

Skull radiography can detect linear skull fractures, depressed fractures, fluid in the para-nasal sinuses and air inside the skull. The skull fractures diagnosed in this way carry anincreased risk of intracranial haematoma (Mendelow et al. 1983), but skull radiographydoes not detect intracranial haematoma as such, nor does it reveal brain damage, and ithas been concluded in a recent meta-analysis that skull radiography is of little value forthe initial assessment of patients with head injury (Hoffman et al. 2000).

2.4.3 Computed tomography (CT)

A head CT scan is considered the gold standard for the detection of intracranial abnor-malities in head injury patients (Vos et al. 2002), as it enables extra-axial haematomassuch as EDHs, SDHs, and SAHs, intra-axial lesions such as cerebral contusions and someDAIs, and secondary lesions such as brain oedema and herniation to be detected (Bešens-ki 2002).

A head CT scan has several important purposes when examining patients with headinjury. Firstly, it is an objective method for documenting scalp, bone and brain damage(Bagley 1999), secondly, initial management decisions can be based on the CT findings(Stieg & Kase 1998), thirdly, the presence of CT abnormalities and their type can assist inassessing the severity of brain injury (Marshall et al. 1991b), and fourthly, acute head CT

30

findings provide a baseline from which to monitor changes in TBI over time (Bešenski2002).

If all head trauma patients with an initial GCS score of 13 to 15 are examined, thepresence of traumatic intracranial findings in the head CT scan varies from 2 to 11%(Moran et al. 1994, Nagy et al. 1999, Haydel et al. 2000, Livingston et al. 2000, Stiell etal. 2001b), but if we take those with an admission GCS score of 9 to 12, the presence ofintracranial abnormality exceeds 30% (Rimel et al. 1982), and for those with a GCS scoreof 8 or less it may even exceed 94% (Eisenberg et al. 1990).

A normal head CT scan on admission has been found in two large multicentre studiesto be good for excluding the need for neurosurgical intervention (Livingston et al. 2000,Shackford et al. 1992). The series considered by Shackford et al. (1992) included 2766head injury patients with an admission GCS at or above 13, among whom none of thosewho had a normal neurological examination and normal head CT scan needed neurosurgi-cal intervention (negative predictive value 100%). Livingston et al. (2000) studied 2152head injury patients with an admission GCS score of 14 to 15 and reported that a normalinitial head CT scan excluded the need for a neurosurgical operation with an accuracy of99.7%. They therefore concluded that MHI patients with a normal initial head CT scancan safely be discharged from the emergency department without observation (Living-ston et al. 2000).

On the other hand, the use of a head CT scan is not efficient, since almost 80% of thescans ordered on account of MHI are negative (Stiell et al. 2001a). It is regarded as man-datory for patients with moderate-to-severe head injury (Ingebrigtsen et al. 2000a), butthe indications for imaging in the case of patients who are initially assessed as havingMHI are controversial.

Two large prospective studies have validated a set of clinical criteria for identifyingMHI patients who need to undergo head CT scanning on admission (Stiell et al. 2001b,Haydel et al. 2000). Haydel et al. (2000) studied 1411 consecutive MHI patients with anadmission GCS score of 15 and found that if all those with risk factors (headache, vomit-ing, seizure, PTA, drug or alcohol intoxication, or age over 60) were imaged, all the trau-matic intracranial abnormalities could be identified (sensitivity 100%, 95% CI 95 to100%). On their criteria, 78% of MHI patients should undergo a head CT scan (Haydel etal. 2000).

Stiell et al. (2001b) reported two sets of clinical criteria for ordering head CT scans forMHI patients, based on a series of 3121 consecutive patients with an admission GCSscore of 13 to 15. The first set of criteria included five risk factors for predicting the needfor neurosurgical intervention (GCS score <15 at 2 h after injury, suspected open ordepressed skull fracture, any sign of basal skull fracture, two or more episodes of vomit-ing, and age at or over 65). These enabled all such complications to be identified (sensi-tivity 100%, 95% CI 92 to 100%), and implied that only 32% of MHI patients would needto undergo a head CT scan. The second set of criteria, which included two risk factors(amnesia before impact > 30 minutes and a dangerous mechanism of injury, such as apedestrian struck by a motor vehicle, an occupant ejected from a motor vehicle, a fallfrom >five stairs), predicted 98% (95% CI 96 to 99%) of all intracranial lesions depictedby a CT scan. The use of these risk factors would lead to the ordering of a CT in 54% ofcases (Stiell et al. 2001b). It may be concluded that not all MHI patients need to undergo

31

head CT scanning and that imaging resources can be directed to those with significantrisk factors, leading to a reduction in total costs (Stiell et al. 2001b, Haydel et al. 2000).

A head CT scan can also be used as a tool for predicting the outcome after head injury.Marshall et al. (1991b) reported a scale based on CT findings that, together with age andthe best motor GCS score, predicted mortality with a sensitivity of 59% and a specificityof 87%. The CT is negative in some patients with severe head injury, however, the pro-portion being 6% according to Eisenberg et al. (1990), although Lobato et al. (1986) hadreported earlier that approximately one half of such patients develop lesions during fol-low-up and that brain atrophy, associated with a poor outcome, often developed withinthe next few months (Lobato et al. 1986). The findings of Lobato et al. (1986) and Eisen-berg et al. (1990) may be explained by the fact that head CT scanning is insensitive fordetecting DAI, and a more sensitive imaging method, such as MRI, is warranted for suchpatients (Bešenski 2002).

2.4.4 Magnetic resonance imaging (MRI)

MRI is more sensitive than a head CT scan for all types of traumatic brain lesions exceptskull fractures and SAHs (Bešenski 2002), but it has some limitations. Firstly, the equip-ment is not as common as for CT, and secondly, the scanning time is longer and prob-lems may be experienced in imaging an unstable trauma patient and organising monitor-ing outside the magnetic resonance field.

Most studies dealing with MRI and TBIs are small and concentrate on severe braininjuries using images obtained with only T1 and T2-weighted sequences. Acute MRI hasbeen found to depict traumatic intracranial pathology in 25 to 57% of MHI patients and in10% of those with a normal head CT scan (Ingebritsen et al. 1999, Voller et al. 1999,Hofman et al. 2001). In cases of more severe head injuries the presence of intracraniallesions in MRI has varied from 60 to 92%, but the image were often obtained at thechronic stage (Hadley et al. 1988, Levin et al. 1988, Fiser et al. 1998).

DAI is the most common type of brain injury seen in MRI of TBI patients, but headCT scans detected only 19% of non-haemorrhagic DAIs (Gentry et al. 1988a). For thehaemorrhagic subset of DAIs, however, MRI was not superior to CT (Gentry et al.1988b). Cortical contusions without haemorrhagic components are also better depicted byMRI, but if haemorrhage is present, CT may reveal all the contusions (Gentry et al.1988b, Hesselink et al. 1988). If a contusion is located in the brain stem, MRI can detectit more sensitively (Gentry et al. 1988b). SDHs are also best detected with MRI (Gentryet al. 1988b, Kelly et al. 1988) and finally, the method is superior to a head CT scan fordetecting all types of brain injury in the subacute and chronic stages (Groswasser et al.1987, Kelly et al. 1988).

Despite its better sensitivity, MRI does not alter the clinical management of imagedTBI patients in the acute stage (Fiser et al. 1998, Kelly et al. 1988). An acute MRI is rec-ommended, however, if the CT is normal but the presence of TBI is suspected, and if apatient is to be imaged in the subacute and chronic stages (Aikuisiän aivovammojenkäypä hoito 2003).

32

New magnetic resonance-based imaging methods have been developed within the lastfew years, e.g. diffusion-weighted imaging, perfusion-weighted imaging, proton magneticresonance spectroscopy and fluid-attenuated inversion-recovery sequences (Ashikaga etal. 1997, Cecil et al. 1998, Ingebrigtsen 1999, Garnett et al. 2001, Sinson et al. 2001,Hammound & Wasserman 2002). Although these have not been validated in large groupsof TBI patients, their use may be indicated in certain situations, at least in the chronicstages, because they may reveal some undetected small traumatic lesions (Aikuisiän aivo-vammojen käypä hoito 2003).

2.4.5 Single photon emission computed tomography (SPECT)

Unlike static neuroimaging techniques such as CT or MRI, SPECT produces functionalrather than structural images of the brain. Using radionuclides, the cerebral blood flow isimaged and reconstructed as a visual data set which also indirectly describes the cerebralmetabolism (Umile et al. 1998).

SPECT is generally more sensitive than CT and MRI for detecting abnormalities afterTBI (Ichise et al. 1994, Ito et al. 1997, Hofman et al. 2001). Ichise et al. (1994) examined29 patients with different degrees of head injury and found SPECT abnormalities in 66%,whereas MRI detected abnormalities in 45% and CT in 34%. SPECT detects more lesionsin the acute stage than in the later stages (Abdel-Dayem et al. 1998), and the majority ofSPECT abnormalities are found in the frontal and temporal regions of the brain (Abdel-Dayem et al. 1998). Some investigators have also found a correlation between neuropsy-chological tests and SPECT abnormalities (Ichise et al. 1994, Hoffman et al. 2001).

SPECT has many limitations. Firstly, there are limited resources for performing it rou-tinely, and secondly, although it can detect many lesions which cannot be seen on CT orMRI, it remains to be elucidated whether these really represent traumatic lesions (Davalos& Bennett 2002). Depression may reduce perfusion and cause false positive lesions inSPECT, for example (Umile et al. 1998).

2.4.6 Electroencephalography (EEG)

The use of EEG for the acute assessment of TBI patients has become rare. EEG describesthe functional response of the brain to insult, EEG abnormalities are often seen in headinjury patients. Geets & Zegher (1985) studied 400 MHI patients and found abnormali-ties in 27%, whereas Voller et al. (1999) found MRI abnormalities in 25% of MHIpatients but none showed EEG abnormalities. Thatcher et al. (1991) examined 162 closedhead injury patients, most of whom had moderate-to-severe head injuries, and found thatEEG was better at predicting the outcome than the GCS score or head CT scans (Thatcheret al. 1999). However, Attia & Cook (1998), in their review of all studies dealing withEEG and the outcome in patients with TBI, concluded that it is still unclear how muchincremental information on clinical examination is added by EEG for predicting the out-come.

33

2.4.7 Biochemical markers

Despite significant progress in cerebral monitoring, it is still difficult to quantify theextent of primary and ongoing secondary brain injury after head trauma. Thus, there is amajor need for a supplementary diagnostic test to follow CT and MRI, such as a bio-chemical marker of brain damage (Ingebrigtsen & Romner 2002).

Bakay & Ward (1983) defined an ideal biochemical serum marker of brain damage asone with high sensitivity to brain injury and high specificity for the brain tissue. It shouldalso be released into the serum rapidly and only after irreversible brain damage, and itshould have adequate clinical relevance.

Numerous potential biochemical markers of brain damage have been studied inpatients with head injury, including creatine kinase BB (Nordby & Urdal 1985), myelinbasic protein (Thomas et al. 1978), neuron specific enolase (Raabe et al. 1999a) and pro-tein S100B (Townend et al. 2002). Glial fibrillary acidic protein has also been suggestedas a potential marker of brain damage, but there are no studies dealing with it in the con-text of TBI (van Geel et al. 2002). The most promising markers are neuron specific eno-lase and protein S100B.

2.4.7.1 Neuron specific enolase (NSE)

NSE is a dimeric protein with a molecular mass of 78 kDa, the and isoforms of which aremainly expressed in neurons and peripheral neuroendocrine tissues (Cooper 1994). NSEis located in the cytoplasm of the neurons, and it is considered a possible marker of neu-ron damage. There are many limitations on its use, however. It is expressed in erythro-cytes, for example, and is known to increase in serum due to haemolysis (Johnsson 1996).Serum values of more than 10 µg/L are considered pathological in studies dealing withhead injuries (Woertgen et al. 1997).

de Kruijk et al. (2001c) measured NSE in 104 MHI patients, but failed to find any sig-nificant association between the marker and the length of PTA, reported LOC, or acutesymptoms due to trauma. They did not report the number of patients above the cut-offlevel of 10 µg/L, but the 90th percentile of NSE was 14.4 in the patients with MHI and13.3 in healthy controls (de Kruijk et al. 2001c).

Woertgen et al. (1997) reported that all their patients with severe head injury had ele-vated levels of NSE, but still had to admit that NSE neither predicted the outcome(Woertgen et al. 1997, Raabe et al. 1999a) nor correlated with head CT scan findings(Raabe 1998). Herrmann et al. (2001) found a trend between increased concentrations ofNSE and neurospychological deficits at two weeks after trauma, but not at six months, inpatients with moderate-to-mild head injury, which led Ingebritsen & Romner (2002) toconclude that despite some promising findings, the sensitivity of NSE is inadequate andits clinical value as a marker of traumatic brain damage is questionable.

34

2.4.7.2 Protein S100B

Protein S100 is a dimeric protein with a molecular mass of 21 kDa (Zimmer et al. 1995).It was first described by Moore (1965) and its name is derived from its solubility in 100%saturated ammonium sulphate at neutral pH. Its two isoforms, ββ and αβ, referred to asS100B, have been studied as a biochemical marker of acute brain damage (Ingebrigtsen& Romner 2002).

S100B is expressed in most human tissues, but the highest concentrations are found inthe astroglial cells of the central nervous system (Hidaka et al. 1983, Kato & Kimura1985, Haimoto et al. 1987, Zimmer DB & van Eldik 1987). The second highest tissueconcentration has been found in adipose tissue, where it is consistently approximately onethird to one fourth of that found in brain tissue (Hidaka et al. 1982, Zimmer & van Eldik1987). The median serum value of S100B in healthy people has been found to be 0.01 µg/L with an upper 97.5% cut-off level of 0.13 µg/L (Anderson et al. 2001). Excretion intourine is the major route for its elimination and its half-life in serum is thought to be short,between 25 and 113 minutes (Usui et al. 1989, Jönsson et al. 2000, Ytrebø et al. 2001).

Damage to glial cells and a reduced integrity of the blood brain barrier may lead toincreased serum S100B concentrations (Nygaard et al. 1997), and consistent with this,elevated S100B levels have been found in patients with TBIs of varying severity. Inge-brigtsen et al. (1995) initially reported that S100B was at or above 0.50 µg/L in 20% oftheir MHI patients who had normal head CT scans on admission, while in a later studythat included 182 MHI patients 35% had S100B values at or above 0.20 µg/L (Ingebrigt-sen et al. 2000b).

Elevated S100B levels in patients with MHI have been found to be associated withprolonged hospitalisation and the presence of neuropsychological dysfunctions (Inge-brigtsen et al. 1997, Waterloo et al. 1997, Ingebrigtsen et al. 1999, Ingebrigtsen et al.2000b, Herrmann et al. 2001), and also to predict the presence of intracranial lesions seenon a head CT scan in cases initially assessed as having MHI (Ingebrigtsen et al. 1999,Ingebrigtsen et al. 2000b, Biberthaler et al. 2001). Even more importantly, normal or lowS100B has been shown to exclude the presence of intracranial abnormalities in a head CTscan with a high accuracy, since a cut-off level of 0.20 µg/L has been found to predict anormal CT scan with an accuracy of 99%.

In patients with moderate-to-severe head injury, increased S100B values have beenassociated with lower GCS scores shortly after trauma (Herrmann et al. 2000, Romner etal. 2000), traumatic lesion size as seen in a head CT scan (Raabe et al. 1998, Romner etal. 2000) and ongoing secondary brain injury (Raabe et al. 1999b, Woertgen et al. 1997).They have also been found to predict a poor outcome in these patients. Woertgen et al.(1997), examined 30 patients with severe head injury (GCS score of 8 or less), and foundthat an unfavourable outcome on discharge was associated with higher S100B values thanwas a favourable outcome, and later the same authors included 44 patients with severehead injury in their analysis and found that the S100B serum levels on admission pre-dicted the outcome better than either the GCS or the Marshall Computed TomographicClassification (Woertgen et al. 1999). Raabe et al. (1999b) included 84 patients withsevere head injury in their series and found that 58% of the patients who died had a peakS100B value of at least 2 µg/L shortly after the trauma. S100B was also found to be anindependent prognostic factor for survival, even when age, GCS score, intracranial pres-

35

sure and head CT scan findings were taken into account (Raabe et al. 1999b). Townend etal. (2002), studying 118 patients with an initial GCS score of 4 to 15, found that a serumS100B concentration above 0.32 µg/L predicted an unfavourable outcome at one monthpost injury with a sensitivity of 93% and a specificity of 72%, while a good recovery atthe same point (extended GOS ≥7) was best predicted with a cut-off of 0.27 µg/L, givinga sensitivity of 76% and a specificity of 69% (Townend et al. 2002). In conclusion,S100B measured shortly after trauma has been shown to correlate with the severity ofTBI and predict the outcome even when other prognostic factors such as the GCS score,CT findings, intracranial pressure level and age are taken into account.

Anderson et al. (2001) reported on a series of 17 trauma patients with severe extracra-nial injuries without evident head injury, all of whom had elevated S100B levels onadmission (0.19 to 10.2 µg/L). Their observations suggest that the S100B measured inserum could also be of extracranial origin, but further research is required to demonstratethe extents to which different types of extracranial trauma contribute to the elevation inS100B.

2.4.8 Neuropsychological assessment

Neuropsychological and neurobehavioural dysfunctions often complicate the outcomeafter TBI. In fact, outcome studies dealing with severe TBIs have shown that chronic dis-ability and the burden imposed on the family are primarily attributable to neuropsycho-logical and neurobehavioural deficits, whereas motor and sensory deficits are less conse-quential (Levin 1993).

Neuropsychological assessment plays an important role in determining the extent andseverity of the sequelae of head trauma, and the healing process after TBI can also be fol-lowed up with neuropsychological tests. A recent Finnish guideline recommends neurop-sychological assessment for all TBI patients who are still suspected of having some neu-ropsychological or neurobehavioural deficits at one month post injury (Aikuisiän aivo-vammojen käypä hoito 2003). The results of neuropsychological tests are not specific toTBI, however, but are vulnerable to many confounding variables such as alcohol abuseand psychotropic drug effects (Mearns & Lees-Haley 1993). Thus, a comprehensiveassessment performed by a professional neuropsychologist is the key factor in obtaining areliable test result (Mearns & Lees-Haley 1993, Aikuisiän aivovammojen käypä hoito2003).

In patients with MHI, information processing speed, memory and attention are themost common impaired functions, and deficits are often seen during the first weeks aftertrauma (Gronwall & Wrightson 1981, Levin et al. 1987, Stuss et al. 1989). In a recentmeta-analysis, the prevalence of impaired neuropsychological performance among thisgroup of patients was found to vary from 1.4 to 7.4%, and the most common single deficitwas attributed to the measures of attention (Binder et al. 1997). The authors concluded,however, that neuropsychological assessment is likely to have a positive predictive valueof less than 50% in patients with mild head injury, and that clinicians should exercise con-siderable caution before diagnosing TBI based on the findings (Binder et al. 1997).

36

The prevalence of neuropsychological impairments increases sharply in patients withmore severe head injury. Herrmann et al. (2001), studying patients with mild-to-moder-ate head injury, found that 74% showed impairment at two weeks post injury and 69%still displayed disorders at six months. Typical dysfunctions were attributed to measuresof attention, executive functions and memory performance (Herrmann et al. 2001). Neu-rospychological impairments practically always complicate the outcome for patients withsevere head injury (Levin 1993). Recovery is rapid within the first year after trauma(Levin et al. 1987, Levin 1993), but subtle changes in memory, attention and behaviouraladaptation may continue for up to 2 years after injury, although the patient’s overall cog-nitive capacity may already have reached a plateau (Stuss et al. 1985, Levin 1993).

2.5 Treatment of head injury

Although there is very little firm scientific evidence on which to base the acute manage-ment of head injury patients (Young A & Willatts 1998, Thornhill et al. 2000), currentevidence suggests that the aims must differ according to degree of the injury (Maas et al.1997, Ingebrigtsen 2000a). The management of those who are initially evaluated ashaving MHI should be focused on assessing the risk of traumatic intracranial haematomaand preventing persistent PCSs (Williams et al. 1990, Stein & Spettell 1995, Ingebrigts-en et al. 2000a, Viola et al. 2000, Servadei et al. 2001, Vos et al. 2002), whereas the man-agement of patients with moderate-to-severe head injury must be directed at minimisingsecondary brain damage (Menon 1999).

Three recent guidelines summarising the acute management of patients with MHImaintain that this should be based on the risk for intracranial complications (Ingebrigtsenet al. 2000a, Servadei et al. 2001, Vos et al. 2002). In the scheme shown in Table 2, MHIpatients are divided into three risk groups: low-risk patients, having an admission GCSscore of 15 without LOC, PTA or additional risk factors, medium-risk patients, having anadmission GCS score of 15 with a history of brief LOC or amnesia, and high-riskpatients, who initially have a GCS score <15. If any of the listed additional risk factors ispresent, the patient must be included in the high-risk group (Table 2). If a head CT scanreveals intracranial pathology with or without a need for intervention, then consultation ortransfer to a neurosurgical centre is recommended. The observation stage can also be per-formed in a local trauma or medical centre. Instructions should always be given topatients who are sent home (Ingebrigtsen et al. 2000a).

37

Table 2. Complication rates and management of MHI patients by risk of intracranial les-ions.

Inconsistent findings exist as to whether a routine follow-up and counselling should begiven for every patient with MHI and whether this can prevent persistent PCSs. Somereports support reassurance and counselling (Relander et al. 1972, Alves et al. 1993, Pon-sford et al. 2002), whereas others do not (Wade et al. 1997). Bed rest after head traumahas not been shown to prevent PCSs (de Kruijk et al. 2002b).

Moderate-to-severe head injuries have to be treated in trauma centres, since this hasbeen shown to reduce overall mortality among such patients (The Brain Trauma Founda-tion 2000a). Two recent guidelines summarising the acute management of severe headinjury patients (Bullock et al. 1996, Maas et al. 1997) both point out that managementshould be focused on preventing secondary brain damage. The preventable causes of suchinjury can be divided into systemic ones such as hypoxaemia, hypotension, hypercapnia,hypocapnia, hyperthermia, hyperglycaemia, hypoglycaemia and hyponatraemia, andintracranial ones such as mass lesions, brain swelling, oedema, seizures and infection(Maas et al. 1997). The systemic causes can be monitored and treated by maintaining thephysiological conditions of the body (Menon 1999), while there are five interventionsavailable to treat and prevent intracranial causes, namely hyperventilation, mannitoladministration, cerebrospinal fluid drainage and the use of barbiturates or corticoids, butall of these have failed to reduce death or disability in controlled trials (Roberts et al.1998). There is general agreement, however, that cerebral perfusion pressure has to beadequate in patients with head injury, between 60 and 70 mmHg, the accepted methodsfor maintaining it being ones which either lower the intracranial pressure or raise themean arterial blood pressure (Menon 1999). It is accepted that all the aforementionedinterventions except for corticoids either lower the intracranial pressure or raise the meanarterial pressure and thus maintain an adequate cerebral perfusion pressure (Maas et al.1997). The indications for surgical treatment in cases of head injury are also generallyaccepted. If an extra-axial haematoma of thickness over 10mm or an intra-axial hae-matoma volume of over 25–30 ml is observed, or if a clot appears to have raised theintracranial pressure, surgery is indicated (Maas et al. 1997).

The use of prophylactic anticonvulsants in the acute stage is likely to prevent theoccurrence of early post-traumatic seizures, but there is no reliable evidence to supporttheir use for preventing late post-traumatic seizures (The Brain Trauma Foundation.2000b). Early enteral feeding should be favoured, since it seems to be associated with atrend towards a better outcome in terms of survival and disability (Yanagawa et al. 2002).

Severity category Approximate risk of lesions (%) ManagementIntracranial Surgical

Low risk Almost zero Almost zero Send home with instructionsMedium risk <10 1–3 CT; If none, observe for 12 hHigh risk 15–20 3–6 CT mandatory; observe for 12 hAdditional risk factors: prolonged unconsciousness or amnesia, anticoagulation or haemophilia, demonstratedor suspected skull fracture, post-traumatic seizures, neurological deficit, multi-injuries, age > 60 years, vomit-ing ≥ 2 episodes. Modified from Ingebrigtsen et al. 2000a, Haydel et al. 2000, Servadei et al. 2001, Vos et al.2002, Stiell et al. 2001b

38

2.6 Outcome after head injury

The Glasgow Outcome Scale (GOS) introduced by Jennett & Bond (1975) for assessingthe outcome after head injury divides patients into five groups: dead, persistent vegeta-tive state, severely disabled, moderately disabled and good recovery. This is a widelyused method to describe the overall outcome after head injury, but as it is likely to beinsensitive to mild disabilities, the authors later introduced an amendment called theextended GOS, in which the three highest categories of the original version are dividedinto better and worse levels (Jennett et al. 1981). Although the GOS and its amended ver-sion have been found to correlate with the emotional and cognitive consequences inpatients with severe head injury (Wilson et al. 2000), there has been a need for a practi-cal outcome scale, especially for MHI patients, who often suffer from subjective symp-toms rather than objectively measurable deficits. Consequently, King et al. (1995) intro-duced the Rivermead Post Concussion Symptoms Questionnaire, which is now widelyused for assessing PCSs attributable to mild head trauma (Ingebrigtsen et al. 2000b).

2.6.1 Mild head injury

Recovery from MHI is usually good and disability unusual. Binder (1997) concludedafter reviewing 17 studies dealing with 2660 MHI patients (GCS ≥ 13) that 86% returnedto their previous work and 7–8% had persistent PCSs. In fact, the prevalence of disabilityafter MHI is clearly attributable to the definition of what is called “mild”. Thornill et al.(2000) studied head injury patients with an admission GCS of 13 to 15 and found thatthey were moderately or severely disabled (GOS: severely disabled or moderately dis-abled) just as often as those who were initially assessed as having a more severe headinjury (GCS ≤ 13). Thus one should not pool all MHI patients with an admission GCSscore of 13 to 15 together but define those with intracranial lesions in CT or MRI or witha PTA over 24 hours as having at least moderate brain injury (Williams et al. 1990,Aikuisiän aivovammojen käypä hoito 2003).

MHI patients have also been reported to have approximately a 1.5 times higher risk ofepileptic seizures than healthy persons, a situation that can last for up to 4 years aftertrauma (Annegers et al. 1998).

2.6.1.1 Post-concussion symptoms (PCSs)

MHI may result in a subjective symptom complex known as PCSs (Alexander 1995,Bernstein 1999), to be found both in children and adults (Lundar & Nestvold 1985, Inge-brigtsen et al. 1998). The most common PCSs are headache, fatigue, dizziness, depres-sion, poor concentration and irritability (Szymanski & Linn 1992). Most patients recoverfrom these symptoms within a year, but a significant minority, approximately 7–15%,

39

may have persistent symptoms lasting for more than a year after the trauma (Alexander1995, Binder 1997).

PCSs were assessed earlier with the investigators’ own checklists, but there is now astandardised questionnaire, known as the Rivermead Post Concussion Symptoms Ques-tionnaire, which includes the sixteen most commonly reported PCSs (King et al. 1995).There is no consensus as to whether one symptom or a certain number of symptoms last-ing long enough should be referred to as the post-concussion syndrome (King 1997), noris there any consensus as to whether or not the symptoms should be divided into sub-groups such as physical symptoms, cognitive complaints and behavioural and affectivesymptoms (Bernstein 1999).

The aetiology of PCSs remains unclear. Demographic, psychogenic and organic fac-tors have all been suggested as playing a role in their development (Lishman 1988,Gasquoine 1997), and there is some evidence that age could increase their risk (Ruther-ford et al.1978). Similarly, females have been thought to develop PCSs more often thanmales (Rutherford 1977, Bazarian et al. 1999). Psychogenic and behavioural factors suchas depression, emotional distress, alcohol or drug abuse and litigation have also beenthought to increase the risk of PCSs (Miller 1961, Lishman 1988, Bohnen & Jolles 1992,Karzmark et al. 1995, Cattelani et al. 1996, King 1996). Organic factors have been poorlystudied, however. PTA as a marker of the severity of TBI has been assessed in relation toPCSs (King 1996, Bazarian et al. 1999), and traumatic brain damage could be an underly-ing factor, at least in some patients. Mittl et al. (1994) found that 30% of their patientswho had been initially assessed as having MHI with normal CT had abnormalities com-patible with DAI in MRI, while Ingebrigtsen et al. (2000b) found a tendency for anincreased frequency of PCSs in patients with an elevated serum S100B concentration.The relationship between PCSs and neuropsychological deficits is unclear. Althoughsome investigators have found higher prevalences of specific deficits such as attentionand reaction time tests in neuropsychological tests performed on patients suffering fromPCSs, there have been a number of patients with PCSs without any deficit (Bohnen &Jolles 1992, Bazarian et al. 1999).

In conclusion, PCSs relatively often complicate recovery after MHI, and in a small butsignificant group of patients the symptoms may be persistent. The clinical picture of PCSsreflects a multifactorial state in which organic, psychological and socioeconomic factorsmay play a role (Lishman 1988). Individual symptoms may substitute for each other overtime and thus the association of a single symptom with damage in a certain region of thebrain may be difficult to establish (Bohnen & Jolles 1992). The role of brain injury in thedevelopment of PCSs has been poorly studied, however. In any case, the prevention ofPCSs has been raised as one of the main aims in the management of MHI patients (Inge-brigtsen et al. 2000a).

2.6.2 Moderate-to-severe head injury

Moderate-to-severe head injury often leads to death or chronic disability. The GOS scoresreported after severe head injury, defined by an admission GCS score ≤8, in four largestudies are summarised in Table 3.

40

Table 3. Glasgow Outcome Scale (GOS) at six months post injury. Findings in four largesurveys of the outcome after severe traumatic brain injury.

Murray et al. (1999) also reported data on 273 patients with moderate head injury (admis-sion GCS score 8 to 12) and found the GOS distribution at six months post injury to be48% good recovery, 21% moderate disability, 14% severe disability, 1% vegetative stateand 16% dead. Similar findings have also been reported at three months (Rimel et al.1982) and one year after moderate head injury (Thornhill et al. 2000). Thus the contribu-tions of patients with severe and moderate head injury to the classes of severe disabilityand moderate disability do not differ much, whereas the occurrence of death is more fre-quent in those with severe head injury and a good recovery in those with moderate headinjury.There are a number of prognostic factors which may influence the outcome in patientswith moderate-to-severe head injury. Increasing age, lower GCS motor score, absence ofpupillary light reflex or oculocephalic reflex, and CT findings especially marks ofincreased intracranial pressure or SAH, have been consistently found to be associatedwith a poor outcome (Choi et al. 1988, Marshall et al. 1991a, Attia & Cook 1998, Signo-rini et al. 1999, Wardlaw et al. 2002), as also have multiple injuries (Gennarelli et al.1989), haemostatic abnormalities (Olson et al. 1989), female gender (Farace & Alves2000), previous head injury (Thornhill et al. 2000), apolipoprotein ε4 genotype (Teasdaleet al. 1997), alcohol abuse (Corrigan 1995) and lower socioeconomic status (Binder1997, Wagner et al. 2000).

Patients with moderate-to-severe head injury have an increased risk of post-traumaticepilepsy (Annegers et al. 1998), this being approximately 7% within 1 year after a closedsevere head injury and 12% within 5 years, or approximately 1% and 1.5%, respectively,after a closed moderate head injury (Annegers et al. 1980, Annegers et al. 1998). Thepresence of contusion, SDH, or skull fracture marks a specific type of trauma thatincreases the risk of later seizures. In the case of both moderate and severe head injuries,the increased risk of late seizures lasts for up to approximately ten years after the trauma(Annegers et al. 1998).

Variable EBIC core data survey - severe

cases

International databank - full

series

USA traumatic coma databank

UK four centres study

Sample size 481 2959 746 976GOS

Dead 40% 49% 36% 39%Vegetative 4% 2% 5% 1%Severe disability 16% 13% 16% 17%Moderate disability 19% 15% 16% 16%Good recovery 21% 20% 27% 24%Unspecified 3%

Modified from Murray et al. (1999)

41

2.7 Alcohol and head injury

2.7.1 General aspects

Alcoholism and hazardous alcohol drinking have an enormous impact on the productivi-ty and quality of life in Western countries. It was estimated that the total annual costs ofalcohol misuse in the US in 1985 exceeded $86–100 billion (Rice et al. 1991, Angell &Kassirer 1994), and ten years later the total costs were estimated to have increased to$175.9 billion (Rice 1999). Thirty three per cent of all economic costs attributed tomotor-vehicle crashes, for example, have been estimated to be associated with alcohol(CDC 1993).

In Finland, the annual direct costs arising from alcohol misuse exceed 550–705 millioneuros. If all the direct medical care expenditure, indirect costs and the value of lost pro-ductivity, including persons who die prematurely due to alcohol misuse, the total annualcosts are estimated to reach 2443–4563 million euros (Päihdetilastollinen vuosikirja2003).

2.7.2 Alcohol drinking as a risk factor for trauma

Acute hospitalisations are often associated with alcohol, so that Charalambous (2002)estimated that 2–40% of all accident and emergency department attendances are of thiskind, and alcohol is especially linked to traumas, approximately one half of all traumapatients being BAC-positive on admission and more than a third having a BAC at orabove 100mg/dL (Rivara et al. 1993a). A high incidence of acute alcohol abuse has alsobeen found among adolescents, to the extent that Porter (2000) reported that 32% of aseries of 2724 trauma patients aged 12–24 years were BAC-positive on admission.

There is direct evidence that alcohol intoxication increases the risk of trauma (Hon-kanen et al. 1976, Cherptitel et al. 1995, Rivara et al. 1997, McLeod et al. 1999, Li et al.2001, Borges et al. 2004). Cherpitel et al. (1995) reported that the risk of injury wasfound to increase with an average daily consumption of one drink and with a frequency ofconsuming five or more drinks on one day more often than twice a year. The case-controlstudy by McLeod et al. (1999) showed in a multivariate analysis after adjustment fordemographic variables that the consumption of more than 60 grammes of ethanol (purealcohol) in a 6-hour period produced an odds ratio of 3.4 (95% CI 1.8–6.4) on sustainingan injury. Alcohol intoxication has also been found to be associated with other forms ofrisk behaviour for sustaining an injury, such as suicidal ideation and violence (Field et al.2001).

42

2.7.3 Alcohol drinking as a risk factor for head injury

Alcohol ingestion is often involved in head injuries, the proportion of BAC positives hav-ing been found to be over 50% in prospective head trauma series (Rimel et al. 1982, Bris-mar et al. 1983, Corrigan 1995, Dikmen et al. 1995), while the number of patients withBAC at or above 100mg/dL has been reported to vary from 35 to 40% (Rimel et al. 1981,Dikmen et al. 1995). Head injury patients are also often chronic alcohol misusers.Corrigan (1995) found in a review of seven studies that the prevalence of chronic alcoholabuse varied from 16% to 66%. Bias was evident in the case of the lowest rates, becausethe findings were based on chart reviews (Corrigan 1995). Brismar et al. (1983) foundthat psychiatric examination revealed alcohol dependence in 43 out of 100 consecutivehead injury patients. Chronic alcohol misuse among head injury patients also seems to beassociated with many forms of risk behaviour which may compound the risk of head inju-ry. These include lower educational attainment, problems with the law, lower perceivedsocial support, and greater prevalence of other substance abuse (Cherner et al. 2001).

2.7.4 Alcohol and trauma morbidity and mortality

Ward et al. (1982) reported that mortality after trauma was significantly lower amongthose with a positive BAC on admission than in sober counterparts, and suggested thatalcohol may have some protective effects on the consequences of injury (Ward et al.1982). Patients who died on the scene were not included in the series. It is known, how-ever, that many accidental deaths occur before hospital admission (Klauber et al. 1981),so that the finding of Ward et al. (1982) may be biased. Waller et al. (1986) later foundafter studying more than 1 million drivers involved in motor vehicle crashes that thedrinking drivers more often suffered serious injury or death than the sober ones, and simi-lar findings have also been reported in bicycling injuries (Li et al. 2001). Thus alcoholingestion is more likely to potentiate the primary injury than to provide protection fromits consequences. Alcohol intoxication on admission has also been shown to result in anincreased use of invasive diagnostic and therapeutic procedures (Gurney et al. 1992,Jurkovich et al. 1992, Melnick et al. 2000), and there is evidence that alcohol abusersoften suffer recurrent traumas (Rivara et al. 1993b).

2.7.5 Alcohol and traumatic brain injury morbidity and mortality

The pathophysiological changes associated with acute and chronic alcohol exposure inthe setting of TBI are complex. Alcohol intoxication can cause haemodynamic and respi-ratory depression, blood-brain barrier disruption and haemostatic impairments (Zink et al.1993, Kelly 1995). Kelly et al. (1997a) found in an experimental setting involving braininjury in rats that high-dose acute alcohol ingestion (3g/kg) increased mortality after trau-ma, whereas better recovery rates were recorded in the low (1g/kg) and moderate (2.5g/

43

kg) alcohol dose groups. Theoretically, alcohol may also be neuroprotective, since itcauses inhibition of N-methyl-D-aspartate (NMDA) receptors (Kelly 1995). On the otherhand, chronic alcohol exposure results in up-regulation of NMDA receptors and down-regulation of -aminobutyric acid receptors, and this adaptation may result in hyperexcit-ability during withdrawal and thus exacerbate TBI (Lovinger 1993, Kelly 1995). There isno clear experimental evidence that chronic alcohol administration exacerbates TBI, how-ever. Neither 6 weeks nor three months of alcohol ingestion in rats was found to have anysignificant exacerbating effects on neurological motor deficits or sizes of morphologicalbrain lesions (Shapira et al. 1997, Masse et al. 2000).

The contribution of acute alcohol intoxication to the outcome after human head injuryremains to be elucidated. There are studies in which alcohol intoxication on admissionhas been found to predict either a favourable (Kraus et al. 1989) or an unfavourable out-come (Brooks et al. 1989, Sparadeo & Gill 1989, Kelly et al. 1997b, Bombardier &Thurber 1998), and studies in which alcohol has not been found to correlate with the out-come at all, or else the correlation has been lost when other prognostic factors have beentaken into account (Choi et al. 1988, Ruff et al. 1990, Signorini et al. 1999, Wagner et al.2000). In contrast, chronic alcohol misuse has quite consistently been found to be associ-ated with a poor outcome. Corrigan (1995) concluded after reviewing seven studies deal-ing with this relationship that only one had failed to find such an association. Separateassociations between alcohol abuse and more severe primary injury, higher mortality rateand poorer neuropsychological and overall outcome have also been reported (Corrigan1995).

2.8 Identification of hazardous alcohol drinking

Hazardous alcohol drinking often means frequent drinking aimed at intoxication. Theidentification of alcohol misuse has been based on a clinical history, specific question-naires and various laboratory tests or combinations of tests (Ewing 1984, Skinner et al.1986, Ross et al. 1990, Mihas & Tavassoli 1992, Hartz et al. 1997). In the case of traumapatients, the results of clinical histories, questionnaires and laboratory tests of alcoholmisuse have been compared with clinical diagnoses of alcohol dependence (Soderstromet al. 1997, Ryb et al. 1999), or with each other (Rivara et al. 1993, Nilssen et al. 1994,Dikmen et al. 1995, Bombardier et al. 1997b, Cherner et al. 2001), but there have beenno studies dealing with the identification of different patterns of alcohol drinking.

44

2.8.1 Tools for identifying hazardous alcohol drinking

2.8.1.1 Questionnaires and interviews

There are a number of questionnaires designed to identify and assess alcohol misuse anddependence, of which the most common are the Michigan Alcoholism Screening Test(MAST) (Selzer 1971), the self-administered Short Michigan Alcoholism Screening Test(SMAST) (Selzer et al. 1975), the Alcohol Dependence Scale (Skinner & Allen 1982),the CAGE questionnaire (Ewing 1984), Self-Administered Alcoholism Screening Test(SAAST) (Hurt et al. 1980) and the Alcohol Use Disorders Identification Test (AUDIT)(Saunders et al. 1993). Most of these have been validated with materials consisting ofalcoholics and non-alcoholics, and they have been shown to distinguish between themwith high sensitivity and specificity (Davis et al. 1987, Ross et al. 1990).

There is also a structured interview for assessing the amount of alcohol drinking andits pattern. The Timeline Followback technique is a retrospective survey which includesmemory aids to enhance patient recall and is thought to allow the collection of reliableinformation on a period of at least 12 months (Sobell & Sobell 1995, Allen et al. 1997).

2.8.1.2 GGT

Gamma-glutamyl transpeptidase (GGT) is a membrane-bound glycoprotein found in themembrane fractions of many tissues, including liver, kidney, brain, spleen, pancreas andheart (Mihas & Tavassoli 1992). Chronic alcohol consumption leads to elevated serumGGT values, and the sensitivity of GGT for detecting alcohol misuse has been reported tovary between 34 and 85% (Sillanaukee 1996). The suggested mechanisms for theincrease of GGT in serum due to alcohol ingestion include induction of hepatic GGT,increased permeability of hepatic plasma membranes and hepatocellular injury. Anincrease in GGT calls for heavy drinking for several days or weeks (Rosman 1992, Sil-lanaukee 1996). Values above 80 U/L (men) and 50 U/L (women) are considered indica-tive of alcohol misuse. The half-life of GGT in serum is approximately 2–3 weeks.

Elevated serum GGT may also be due to non-alcoholic conditions, however. Non-alco-holic liver diseases, diabetes, obesity, heart failure, the use of anticonvulsants and severetrauma are such conditions (Sillanaukee 1996).

2.8.1.3 AST & ALT

Aspartate aminotransferase (AST) and alanine aminotransferase (ALT) are enzymes thatare abundant in the liver, and AST is also found in the heart, muscle, brain, pancreas, kid-ney and lungs (Mihas & Tavassoli 1992, Rosman 1992). These enzymes may increase inserum after alcohol-induced liver damage, but also in cases of non-alcoholic liver diseas-

45

es, muscle disorders and myocardial infarction (Sillanaukee 1996). AST values above 50U/L (men) and 35 U/L (women) are considered indicative of alcohol misuse. The sensi-tivity of AST for detecting alcohol misuse has varied between 15 and 69%, but the speci-ficity has been low (Sillanaukee 1996). ALT is not as useful as AST as a marker of alco-hol misuse, but a ratio of AST to ALT of over two may be useful, and may indicate alco-holic hepatitis (Mihas & Tavassoli 1992).

2.8.1.4 CDT

Transferrin is the iron transport protein, which includes two globular domains and sever-al polysaccharide side chains (Mihas & Tavassoli 1992). When transferrin is exposed toalcohol, abnormal side chains result and carbohydrate-deficient transferrin (CDT) is pro-duced (Lesch & Walter 1996).

CDT is a promising marker of excessive alcohol use. Serum CDT was found to beincreased after a daily intake of 50–80 grammes of pure alcohol for at least one week andwas normalised slowly during abstinence, with a half-life of approximately 15 days(Stibler 1991). Values above 20 U/L (men) and 26 U/L (women) are considered indicativeof alcohol misuse.

After pooling approximately 2500 individuals from relevant studies, Stibler (1991)found that CDT had a sensitivity of 82% and a specificity of 97% for detecting alcoholmisuse. Meerkerk et al. (1998), however, reported that in studies dealing with more gen-eral populations the sensitivity of CDT varied from 12% to 45% and the specificity from87% to 96%. False positive findings have been reported at least in patients with primarybiliary cirrhosis, chronic hepatitis C, hepatic malignancies, a genetic variation in transfer-rin and certain rare inborn glycoprotein disorders (Sillanaukee 1996). Common chronicdiseases such as hypertension, asthma/bronchitis, diabetes, angina pectoris, depressionand disorders of the digestive tract have not been found to cause unspecificity (Meerkerket al. 1998).

2.8.1.5 MCV

The mean corpuscular volume (MCV) of erythrocytes may be elevated as a result ofexcessive alcohol consumption, but also in response to many other conditions such as liv-er disease, folate deficiency and hypothyroidism (Sillanaukee 1996). Values above 96 fLare considered indicative of alcohol misuse. MCV returns to normal values within 3months after the beginning of abstinence. Its sensitivity for alcohol misuse is relativelylow, since MCV detects only 30–40% of subjects with a drinking problem (Mihas &Tavassoli 1992).

46

2.8.1.6 BAC

Blood alcohol concentration (BAC) can be measured directly from a blood sample or beestimated from breath, saliva or urine measurements. Urine measurements often showhigher concentrations than measurements in blood, whereas breath measurements give anaccurate estimate of the real BAC (Bendtsen et al. 1999).

BAC of at least 150mg/dL without gross evidence of intoxication, 300mg/dL at anytime or 100mg/dL in routine examination are first-level criteria for alcoholism (CriteriaCommittee, National Council on Alcoholism, New York 1972).

2.8.2 Detecting hazardous alcohol drinking

2.8.2.1 In patients with trauma

CAGE has been found to be the best questionnaire for predicting chronic alcohol depen-dence in trauma patients (Nilssen et al. 1994, Soderstrom et al. 1997), where it predictedlife-time alcohol dependence with a sensitivity of 84% and a specificity of 90% (Soder-strom et al. 1997). By contrast, laboratory tests such as GGT, AST, MCV and serumosmolality seem to lack sensitivity and specificity for detecting alcohol dependence intrauma patients. Ryb et al. (1999) compared the above four tests in 684 male traumapatients and found relative low sensitivities (0.51, 0.43, 0.27, and 0.74, respectively) andspecificities (0.78, 0,74, 0.89, and 0.70, respectively) for detecting current alcohol depen-dence. BAC on admission nevertheless appeared to be associated with recent alcoholdependence better than any of the conventional biochemical markers (Ryb et al. 1999).

2.8.2.2 In patients with head injury

Although the prevalence of hazardous alcohol drinking may be high in patients with headinjury, the identification of alcohol misuse among such patients has not been studiedgreatly (Corrigan 1995). Dikmen et al. (1995) found that 45% of patients with head inju-ry reported 2 or more items on the SMAST, a level considered indicative of problemdrinking. Bombardier et al. (1997b) studied 50 head injury patients admitted to an in-patient rehabilitation programme and found that MAST showed alcohol problems in 74%of these cases, while Brismar et al. (1983) found that 43 out of 100 consecutive headinjury patients were alcohol-dependent, and GGT was elevated in 37%, AST in 35% andALT in 28% of the total series. No relationship between alcohol dependence and bio-chemical markers was reported (Brismar et al. 1983).

47

2.9 Alcohol interventions as means of preventing injuries

Although alcohol is thought to be a significant cause of trauma, there still continues to bea lack of attention to hazardous drinking in accident and emergency departments (Soder-strom & Cowley 1987, Schermer et al. 2003), although these could be optimal places forcommencing alcohol interventions and thus reducing its detrimental effects (Charalam-bous 2002).

Alcohol interventions are aimed at reducing alcohol drinking and alcohol-relatedeffects such as violence, injuries and illnesses. There are many intervention protocols, buta brief alcohol intervention is thought to be the most practicable in accident and emer-gency departments (Dunn et al. 1997). Holder et al. (1991) found a negative relationshipbetween the effectiveness and costs of alcohol interventions, which further supports theuse of inexpensive, easy methods of alcohol counselling.

Evidence is accumulating that brief interventions are effective in reducing alcoholdrinking and may thereby also reduce the risk of alcohol-related accidents (Antti-Poika Iet al. 1988, Walsh et al. 1991, Fleming et al. 1997, Holder et al. 2000). Two recent meta-analyses have concluded that alcohol interventions may be associated with reductions insuicide attempts, domestic violence, drinking-related injuries and injury hospitalizations,and with reductions in deaths ranging from 27% to 65% (Dinh-Zarr et al. 1999, Dinh-Zarret al. 2000).

In a study of 2524 consecutive trauma patients, including 1153 who were screened aspositive with respect to alcohol problems, Gentilello et al. (1999) randomised 366patients to a counselling group (a brief alcohol intervention) and 396 to a control group. Asignificant decrease in alcohol consumption was found in the former at 12 months,accompanied, interestingly, by significant reductions in injuries requiring emergency ortrauma centre admissions (47%) or hospital admissions (48%) within three years of theoriginal trauma (Gentilello et al. 1999).

Even though hazardous alcohol drinking is a major problem affecting patients withhead injury, there are no studies dealing with a brief alcohol intervention as a means ofreducing alcohol consumption in this group. On the other hand, the consumption of alco-hol has been found to decrease after the occurrence of head injury even without structuredinterventions (Kreutzer et al. 1991, Dikmen et al. 1995). Dikmen et al. (1995) found thatalcohol consumption had decreased most compared with the pre-injury level at 1 month,after which it increased significantly, but the patients were not followed-up for more than12 months. The reduction in alcohol consumption early after trauma may be due to lackof access to alcohol (still hospitalised), non-structured advice given by health care provid-ers, decreased tolerance of alcohol, or rethinking in the aftermath of a major trauma (Dik-men et al. 1995).

3 Aims of the research

The purpose of the present research was:1. to find simple clinical measures that predict the development of PCSs in patients with

mild head injury and to validate serum protein S100B as a prognostic marker. (I)2. to evaluate the clinical utility of protein S100B as a marker of brain damage in patie-

nts with multitrauma, and to find out whether its usefulness is confounded by extrac-ranial injuries. (II)

3. to study patterns of alcohol drinking in trauma patients and assess whether there arespecific risk patterns for head trauma. (III)

4. to investigate how hazardous alcohol drinkers among patients with head injuries andother types of trauma can best be identified on admission. (IV)

4 Subject and methods

The work was carried out in the Department of Neurology, University of Oulu, during theyears 1998–2004. The protocol was approved by the Ethics Committee of the MedicalFaculty, University of Oulu, and carried out according to the principles of the Declara-tion of Helsinki. All the patients or their relatives gave informed consent before inclusionin the series.

4.1 Subjects

The series comprised 385 consecutive acute trauma patients aged 16 to 49 years admittedto the emergency department of Oulu University Hospital between June 1998 and July2000, comprising 224 having head injuries with or without extracranial injuries and 161patients having extracranial injuries only. Patients admitted more than 6 hours after thetrauma event were excluded. None of the patients reported suffering from cancer, multi-ple sclerosis, stroke, or treated epilepsy.

Paper I involved 172 MHI patients, of whom 129 were men and 43 women. These rep-resented the 199 out of the 224 head injury patients who fulfilled the following criteria forMHI: Glasgow Coma Scale (GCS) score of 13–15 on admission, loss of consciousnessfor ≤ 30 minutes, if at all, and no focal deficits in a neurological examination performedon admission. Twenty-seven of these 199 MHI patients had to be excluded because theywere not reached, so that no interview to determine post-concussion symptoms (PCSs)could be performed.

Paper II covered a total of 379 trauma patients, of whom 224 were head injury patientsand 155 patients with extracranial injuries only. Fifty-four of the head injury patients hadsimultaneous extracranial injuries. Six patients with extracranial injuries only were omit-ted because they reported PTA, so that the presence of TBI could not be convincinglyexcluded. Additionally, we examined whether the exposure of healthy individuals to high+Gz forces without actual impact on tissues could increase serum S100B values. For thispurpose, we recruited eight flight crew members of the Finnish Air Force, seven of whomperformed two flight missions seated in the back seat of a BA Hawk Mk 51 A aircraft and

50

wearing an extended coverage anti-G suit. The detailed characteristics of the flight mis-sions and management of these subjects have been described elsewhere (Siitonen 2000).

Paper III included all the 385 trauma patients, and paper IV included those who wereinterviewed for patterns of alcohol drinking (n=349).

4.2 Methods

4.2.1 Clinical examination

4.2.1.1 Emergency room

The patients included in the series were carefully examined on admission by emergencyroom physicians according to a structured checklist designed for this purpose. Clinicaldata, including neurological examinations, the presence of clinical symptoms such asunconsciousness, amnesia, dizziness, headache, neck pain, nausea and vomiting, and thetype and extent of the head and extracranial injuries were carefully recorded. A history ofinfections (HIV/AIDS, hepatitis B and C) and other diseases was also recorded. The lev-el of alcohol intoxication was assessed clinically and graded as none, slight, or severeintoxication. Venous blood samples were obtained from each patient for biochemicaldeterminations.

The injured body parts of each patient were divided into six categories: head, spine,thorax, abdomen and upper and lower extremities. The presence of injury was recordedwhen there was distinct physical evidence of trauma as assessed by the emergency roomphysician. The Injury Severity Score (ISS) was also used as an index of trauma severity(Baker et al. 1974, Copes et al. 1988), and the GCS to assess the level of consciousness(Teasdale & Jennett 1974).

The physician decided on admission which of the additional examinations (imagingmodalities) were needed, and the patients were subsequently treated according to the hos-pital’s routine protocols. Sixty-six head injury patients underwent a head CT scan within24 hours from the trauma, and the resulting scans were evaluated twice, first as part of theroutine clinical evaluation and later in a re-evaluation of all the scans by a trained neuro-radiologist. Thirty-three patients were further examined by skull radiography.

The patients with head injury discussed in paper II were classified on the basis of theirsymptoms and signs as follows: 1. 35 patients having head injury without brain injury, 2.165 patients having head injury with mild brain injury, and 3. 24 patients having headinjury with moderate-to-severe brain injury. The groups fulfilled the following diagnosticcriteria:1. Head injury without amnesia, unconsciousness, headache, dizziness or vomiting.2. Head injury with amnesia and/or unconsciousness (duration ≤ 24 hours), headache,

dizziness or vomiting

51

3. Head injury with amnesia and/or unconsciousness (duration > 24 hours), intracraniallesion visible in the head CT scan or focal deficits in the neurological examination.

The extracranial injuries were divided in paper II into two groups: small (soft tissuecontusions, wounds, sprains, luxations and small fractures) and large (large fractures andabdominal injuries). Vertebral, pelvic, femoral and brachial fractures and fractures of bothforearm bones were assessed as large and all other extracranial fractures as small.

4.2.1.2 Alcohol data

Blood alcohol concentration (BAC) on admission was determined either from exhaled air(n=206) or from serum samples (n=179). Personal interviews were conducted with 349/385 (91%) of the patients by the same interviewer throughout using a structured question-naire to gather data on their alcohol consumption. The interviews were performed blindto any knowledge of the determinations of BAC levels or other biochemical markers ofalcohol consumption. The interviews took place within 2–6 weeks of the injury.

The history of alcohol consumption included the following information: how manydrinks of alcohol (one standard drink = 12 g of ethyl alcohol, corresponding to one beer,one glass of table wine or four centilitres of proof spirit) the patient had consumed during(i): 24 hours, and (ii): one week preceding the injury. Daily alcohol consumption during aperiod of one year prior to the trauma was also assessed using a time-line follow backtechnique (Sobell & Sobell 1995, Allen et al. 1997). Based on the resulting data, thepatients were classified into dependent drinkers, binge drinkers, light-to-moderate drink-ers and non-drinkers. The dependent drinkers were those who showed clinical evidence ofpathological alcohol use, social impairment and tolerance/withdrawal. Their daily alco-hol consumption had exceeded 80 g. Binge drinking (i.e. heavy episodic drinking) wasdefined as an intake of 6 or more (men) or 4 or more (women) standard drinks of alcoholin one session. Binge drinkers were further divided into frequent binge drinkers, whoreported this type of drinking at least once a month, and infrequent binge drinkers, whoreported that it occurred 1–11 times per year. Light-to-moderate drinkers did not drink forintoxication, but consumed 1–2 standard drinks per day either daily or less frequently.Non-drinkers had not drunk any alcohol during the year preceding the injury, andincluded both life-long abstainers and ex-drinkers. The dependent drinkers and frequentbinge drinkers together made up the group referred to as hazardous drinkers.

4.2.1.3 Follow-up interviews

The follow-up interviews were carried out twice, a face-to-face interview within 2–6weeks of the trauma, but not before full recovery from amnesia, and a second, telephoneinterview 8–30 months after the injury.

The first interview included questions on the trauma, education and employment sta-tus, previous diseases, use of medicines, lifestyle factors and a modified Rivermead Post

52

Concussion Symptoms Questionnaire (King et al. 1995) in which the answers weredichotomized to yes or no. Questions were also added regarding decreased alcohol toler-ance and the occurrence of panic attacks. The presence and duration of post-traumaticamnesia (PTA) at the time of injury was assessed according to the Rivermead Post-Trau-matic Amnesia Protocol (King et al. 1997), and educational status was recorded in termsof the number of years of schooling. The patient was also asked whether he or she hadinsurance that covered the accident. Previous diseases such as TBIs, psychiatric illnessesand contacts and HIV, hepatitis or other liver diseases were also ascertained, and medica-tions were carefully recorded.

The second interview was a telephone interview (8–30 months after the injury), whichaimed at ensuring that the patients with PCSs really fulfilled our criteria (I). This wasdone because the first interview had been performed 2–6 weeks after trauma, and accord-ing to our definition of PCSs, symptom(s) had to last at least one month post injury beforethe criteria were fulfilled.

4.2.2 Laboratory procedures

Venous blood samples were obtained from each trauma patient as soon as possible afteradmission, but no later than six hours from trauma. The samples were centrifuged andstored at –20ºC until analysed, except that the mean corpuscular volume (MCV) of eryth-rocytes was determined as soon as possible after admission. Protein S100B, serum gam-ma-glutamyl transpeptidase (GGT), aspartate aminotransferase (AST) and carbohydrate-deficient transferrin (CDT) were measured for each patient, whereas MCV was deter-mined in 302/385 cases (78%).

The concentration of protein S100B was measured using a commercially availablemonoclonal two-site immunoluminometric assay (LIA-mat® Sangtec® 100; AB SangtecMedical, Bromma, Sweden) with a detection limit of 0.02 µg/L. Blood alcohol concentra-tions were measured on a Vitros 250 clinical chemistry analyser (Johnson & JohnsonClinical Diagnostics, Rochester, New York), and the ALCO-SENSOR III (IntoximetersInc., St. Louis, Mo) was used for the breath analyses. Serum CDT was measured with acompetitive radioimmunoassay after microcolumn separation (CDTect, AxisShield, Oslo,Norway), and the measurements of MCV, GGT and AST were carried out using standardclinical chemical methods.

Normalised S100B values for paper I were calculated according to a half-life of 120minutes as follows: normalised S100B = 2 (time from trauma to blood sampling in minutes / 120minutes) x measured S100B. The following cut-off values were used in the analyses for thediagnostic characteristics of the laboratory tests: in paper I: S100B 0.50 µg/L, in paper IV:MCV >96 fL for women and men, GGT >50 U/L for women and >80 U/L for men, AST>35 U/L for women and >50 U/L for men and CDT >26 U/L for women and >20 U/L formen.

53

4.2.3 Statistical analysis

Categorial variables were compared by means of Fisher’s exact 2-tailed test or the Pear-son χ2 test. Continuous variables were compared among groups using Student’s t-test, theMann-Whitney U-test, or the Kruskall-Wallis test. Univariate associations of continuousvariables were tested by means of Spearman’s rank correlation coefficients (rs). Receiveroperating characteristic (ROC) curves for the S100B values were also drawn in paper Ifor detecting PCSs patients with and without S100B normalisation, and sensitivities,specificities, positive and negative predictive values, and 95% confidence intervals (CIs)were calculated for each marker of alcohol consumption in paper IV for detecting hazard-ous alcohol drinking.

Odds ratios (OR) and 95% confidence intervals (CI) before and after adjustment forpossible confounding variables were calculated by logistic regression for paper I, and thehypotheses were tested and 95% CI determined using standard error estimates for thelogistic coefficients. Stepwise multiple logistic regression (p<0.1 for entry limit andp>0.15 for removal limit) was used to test the significant independent risk factors forPCSs. In paper III, relative risk (RR) estimates and 95% CIs were calculated before andafter adjustment for confounding factors, and in paper II, a two-way analysis of variancewas used to assess the effects of head injuries and other types of trauma on S100B values.

All the analyses in papers I and II were performed using the Statistical Package for theSocial Sciences (SPSS, version 10.0, for Windows, SPSS Inc, Chicago, IL USA), and inpapers III and IV using the CIA statistical software (Gardner & Altman 1989) and SPSS.

5 Results

5.1 Protein S100B as a predictor of PCSs in MHI patients (paper I)

A total of 172/199 (86%) MHI patients were included in this series, 129 men and 43women, with a mean age of 31.3 years (SD 10, range 10–49) (I; Table 2). Thirty-seven ofthem (22%) reported PCSs, i.e. at least one PCS lasting at least one month following thetrauma event. The most common PCSs were depression, dizziness, fatigue and headache(I; Table 1). Thirty of these patients (81%) had suffered from at least two symptoms, andthe mean number of symptoms amounted to 5.0 (SD 3.6, range 1–14). Thirty-nine of the172 patients (23%) underwent a head CT scan on admission, and eight had traumaticintracranial lesions visible. Of the 164 patients who did not have any evidence of intracra-nial lesions, 32 (20%) reported PCSs.

Significant risk factors for PCSs emerging in univariate analysis were the presence ofskull or facial bone fracture, elevated serum protein S100B (0.50 µg/L), the presence ofdizziness and headache on admission, psychiatric illness in childhood, loss of conscious-ness and PTA (I; Table 2). There was also a correlation between the number of symptomsand the S100B values (Spearman correlation, p < 0.001), but no statistically significantcorrelation between PCSs and age (as a continuous variable), psychiatric illness in adult-hood, use of psychotropic drugs, current smoking, prior use of illicit drugs, or neck painon admission.

The ROC curves for S100B in the 172 MHI patients with and without S100B normali-sation are presented in Figure 2. Serum protein S100B with a cut-off level of 0.50 µg/Lshowed a high specificity for PCSs (93%) but a rather low sensitivity (27%). The sensi-tivity with a cut-off value of 0.20 µg/L was 68% and the specificity 67%, but the negativepredictive value was as high as 88%.

55

Fig. 2. Comparison of ROC curves for S100B in 172 MHI patients before and after normalisa-tion of S100B values (to correspond to the time of onset of head injury). Sensitivities and speci-ficities at different S100B cut-off levels are shown. Figure from paper I.

Skull fracture, elevated protein S100B (≥ 0.50 µg/L), dizziness and headache on admis-sion and age were identified as independent risk factors for PCSs (Table 4). The modelremained unchanged after normalisation of the S100B values. This model gave ORs of6.1 (1.9–19.6) for skull fracture, 6.0 (2.3–15.2) for protein S100B, 3.0 (1.1–7.9) for dizzi-ness, 2.3 (0.9–5.9) for headache and 1.05 (1.00–1.10) for age.

Table 4. Multivariate ORs for PCSs in mild head injury patients (n=172). Table frompaper I.

Loss of consciousness, post-traumatic amnesia, presence of extracranial injury, prior headinjury, employment status, insurance, psychotropic drugs, current heavy drinking, smok-ing and prior use of illicit drugs were not independent risk factors for PCSs.

Variable OR 95% CI p valueSkull fracture 8.0 2.6–24.6 0.001Protein S100B ≥ 0.50 µg/L 5.5 1.6–18.6 0.007Dizziness on admission 3.1 1.2–8.0 0.021Headache on admission 2.6 1.0–6.5 0.043Age 1.05 1.01–1.10 0.027ORs represent comparisons between patients with and without at least one symptom lasting for one month or more and are adjusted for sex, educational status and the other variables listed in the table.

56

5.2 Effects of extracranial injuries on S100B levels (paper II)

A total of 379 trauma patients were included in this series, of whom 224 had head trau-ma and 155 had extracranial injuries only. Fifty-four (24%) of the 224 head traumapatients also had extracranial injuries.

The median serum concentration of protein S100B among the patients with headtrauma was 0.17 µg/L (0.07–0.36, Q1–Q3), whereas the corresponding value for thosewithout head trauma was 0.07 µg/L (0.03–0.13, Q1–Q3) (Mann-Whitney U-test: p <0.001). When the head trauma patients were classified into those with no brain injury,mild brain injury and moderate-to-severe brain injury, the median levels (n, Q1–Q3) ofS100B were 0.10 µg/L (35, 0.06–0.18), 0.16 µg/L (165, 0.07–0.32) and 1.27 µg/L (24,0.32–3.3), respectively (Kruskall-Wallis test: p < 0.001). The median S100B concentra-tion (n, Q1–Q3) was 0.35 µg/L (7, 0.20–0.64) for the patients with large extracranial inju-ries only and 0.07 µg/L (148, 0.03–0.12) for those with small extracranial injuries (Mann-Whitney U-test: p < 0.001).

The effects of the severity of brain injury and the size of extracranial injury on S100Blevels (logarithmic modification of S100B values) are illustrated in Figure 3. Both braininjury and extracranial injury independently increased the serum S100B concentrations inthe trauma patients, since two-way analysis of variance showed that the two variables andtheir interaction were statistically significant (p < 0.001).

A comparison of S100B cut-off levels showed that there were only a few head traumapatients without brain injury or patients with extracranial injuries alone above the highestcut-off level (0.50 µg/L) (II; Table 4), whereas the head trauma patients with moderate-to-severe brain injury exceeded this cut-off in 67% of cases. Likewise, 61% (136/224) of thehead trauma patients and 26% (41/155) of those with extracranial injuries only hadS100B above 0.13 µg/L (Pearson χ2 test: p < 0.001).

Fig. 3. Effects of the severity of brain injury and the size of extracranial injury on S100B levels.Logarithmic modification of S100B concentrations: Ln (1 + S100B). Figure from paper II.

57

Of the different types of extracranial injuries, large extracranial injuries such as largefractures and abdominal injuries elevated S100B values significantly (II; Table 3), where-as soft tissue contusions, wounds, sprains, luxations and small fractures (i.e. small extrac-ranial injuries) did so only slightly (II; Table 3). We did not find any significant increasein serum S100B in healthy individuals after exposure to high +Gz forces, the medianS100B level (Q1–Q3) in blood samples taken 5 and 60 minutes after the 22-minute flightmissions with a maximal acceleration of + 6 Gz being 0.00 µg/L (0.00–0.03).

5.3 Binge drinking as a risk factor for head trauma (paper III)

Eight per cent of the trauma patients were found to be dependent drinkers, 61% frequentbinge drinkers and 17% infrequent binge drinkers (Table 5). Thus a total of 78% of thepatients reported binge-type drinking. The light-to-moderate drinkers and non-drinkersrepresented 8% and 6% of the population, respectively. The dependent drinkers tended tobe older than the patients with other drinking patterns, and they also had the highestBACs on admission (Table 5). The proportion of women among the dependent drinkersand frequent binge drinkers was found to be low.

Table 5. Characteristics of the patients classified according to the history of alcohol con-sumption (n = 349).

Dependent drinking and frequent binge drinking were found to be common patternsamong the head trauma patients (Table 6), in whom they occurred significantly moreoften than in those with other types of trauma (77% versus 59%, RR, 2.38; 95% CI, 1.50to 3.77).

Fifty-one per cent of all the patients had alcohol in their blood on admission and mostof the intoxicated subjects (86%) had reached a level of 100 mg/dL. When the patientswere classified according to the type of trauma, the incidence of elevated BAC was sig-nificantly higher in the head trauma patients than in those with other types (65% versus32%, RR, 3.92; 95% CI, 2.55 to 6.03). The relative risk of sustaining a traumatic head

Pattern of dringing Patientsn (%)

Womenn (%)

Age (years)Mean ±

SD

BAC (mg/dL)

on admissionMean ± SD

Clinicallyintoxicated on

admissionn (%)

Alcohol consump-tion during the pre-ceding year, g/day

Mean ± SD,women/men

Dependent drinkers 26 (8) 0 38 ± 9 210 ± 160 20 (77) 137±45 †Frequent binge drinkers

214 (61) 51 (24) 30 ± 10 120 ± 110 124 (58) 16±15 / 27±20

Infrequent binge drinkers

59 (17) 25 (42) 33 ± 10 20 ± 70 5 (8) 6±3 / 6±6

Light-to-moderate drinkers

28 (8) 14 (50) 31 ± 12 0 ± 10 1 (4) 2±1 / 4±4

Non-drinkers 22 (6) 7 (32) 30 ± 12 0 0 0 / 0BAC, blood alcohol concentration. † Men only

58

injury increased sharply with increasing BAC, being significantly higher than that forother types of trauma above the level of 150 mg/dL (Figure 4). The RRs were as follows:0 mg/dL (0.51; 95% CI 0.42–0.63), 1–99 mg/dL (0.67; 95% CI 0.32–1.38), 100–149 mg/dL (0.84; 95% CI 0.40–1.76), 150–199 mg/dL (1.71; 95% CI 0.93–3.17) and > 199 mg/dL (4.87; 95% CI 2.82–8.40).

Table 6. Drinking patterns of the trauma patients interviewed (n=349), by type of trauma.

Fig. 4. Crude relative risks for head injury at various level of blood alcohol concentration(BAC). Above the level of 150 mg/dL, the risk for head injury was greater than the risk for othertypes of injury. The bars indicate the 95 % confidence intervals at different BACs, and if thebar does not cross the level one, it is statistically significant.

The main causes of trauma under the influence of alcohol were assaults, falls on theground and biking accidents, so that these together accounted for 68% (135/198) of allthe trauma patients who had alcohol in their blood. Ninety-four per cent of the assaultvictims were BAC-positive on admission and most of their traumas (92%) were headinjuries (III; Table 3). The patients injured in falls on the ground and biking accidents alsofrequently had alcohol in their blood, 60% and 61%, respectively. The BAC-positive sub-jects sustained head traumas significantly more often both in falls on the ground (RR,3.50; 95% CI, 1.32 to 9.26) and bicycling injuries (RR, 8.26; 95% CI, 1.48 to 45.45) than

Pattern of drinking All

N=349

Head trauma

N=192

Other trauma

N=157Dependent drinkers 26 (8%) 24 (12%) 2 (1%)Frequent binge drinkers 214 (61%) 124 (65%) 90 (57%)Infrequent binge drinkers 59 (17%) 27 (14%) 32 (21%)Light-to-moderate drinkers 28 (8%) 8 (4%) 20 (13%)Non-drinkers 22 (6%) 9 (5%) 13 (8%)

59

their sober counterparts (III; Table 3), but the BAC-negative bicyclists had injuries totheir extremities (50%, 9/18) significantly more often than the BAC-positive ones (7%, 2/28) (RR, 12.99; 95% CI, 2.35 to 71.43).

5.4 Laboratory markers of hazardous alcohol drinking in trauma patients (paper IV)

Sixty-nine per cent (240/349) of the trauma patients interviewed reported hazardous alco-hol drinking (i.e. they were dependent drinkers or frequent binge drinkers). BAC (blood/breath alcohol) proved to be the most sensitive marker of hazardous drinking and a high-ly specific one (Table 7), so that 137 (57%) of the hazardous alcohol drinkers had a read-ing above 100 mg/dL. When a cut-off level of > 0 mg/dL was used, the sensitivity foridentifying hazardous drinkers increased to 68% (95% CI, 61 to 73%) with a positive pre-dictive value of 96% (95% CI, 92 to 98%). Thus, 96% of the BAC-positive traumapatients proved to be hazardous alcohol drinkers.

We further analysed the usefulness of various combinations of biochemical markers.BAC (>0 mg/dL) together with CDT was the most sensitive combination, correctly iden-tifying 73% of the target population, but even though both CDT and GGT slightlyimproved the sensitivity when combined with BAC, the additional effect did not reachsignificance.

Table 7. Sensitivities, specificities, positive predictive values (PPV) and negative predic-tive values (NPV) of markers of alcohol consumption for detecting hazardous alcoholdrinking (including dependent drinkers and frequent binge drinkers) among the traumapatients (n=349). Table modified from paper IV.

Screening test Sensitivity Specificity PPV NPVBAC >0 mg/dL, (95% CI)

68% (61 to 73%) 94% (87 to 97%) 96% (92 to 98%) 57% (49 to 64%)

BAC >100mg/dL 57% (51 to 63%) 94% (89 to 98%) 96% (91 to 98%) 50% (43 to 57%)GGT 11% (8 to 16%) 97% (92 to 99%) 90% (74 to 97%) 33% (28 to 39%)MCV 18% (13 to 24%) 94% (88 to 98%) 88% (75 to 95%) 34% (28 to 40%)CDT 33% (28 to 40%) 88% (81 to 93%) 86% (78 to 92%) 38% (32 to 44%)AST 17% (13 to 22%) 94% (89 to 98%) 87% (74 to 94%) 34% (29 to 40%)BAC = blood alcohol concentration (100 mg/dL = 22 mmol/L); cut-off values: CDT (20/26 U/L, men/women), GGT (80/50 U/L, men/women), AST (50/35 U/L, men/women), MCV (96 fL, men and women). MCV was measured for 288 (83%) of the patients interviewed (n = 349).

6 Discussion

6.1 PCSs in patients with mild head injury

The reported incidence rates of PCSs in patients with MHI vary widely, but a significantminority, approximately 7–15%, may have persistent symptoms lasting more than a yearafter trauma (Alexander 1995, Binder 1997). In our series, 22% of the MHI patients hadPCSs one month after the trauma, and if those with intracranial lesions visible in CT areexcluded, the corresponding figure is 20%.

Skull or facial bone fracture, elevated serum protein S100B (≥ 0.50 µg/L), dizzinessand headache on admission were found to be independent predictors of PCSs in the MHIpatients. Elevated serum protein S100B was a significant predictor even when the patientswith intracranial lesions verified by a head CT scan were excluded. The associationbetween elevated protein S100B serum concentrations on admission shortly after headinjury and PCSs strongly suggests an organic aetiology for PCSs. Serum S100B was ele-vated on admission in 27% of the MHI patients who developed PCSs, but only in 7 % ofthose who did not.

Previous studies have also observed an association between elevated S100B and PCSsamong MHI patients (Ingebritsen et al. 2000b, de Kruijk et al. 2002a), but the associa-tion was even more distinct in the present case, possibly because we had a rather low pro-portion of PCSs patients (22%). Ingebrigtsen et al. (2000b) identified a relatively highproportion of PCSs among MHI patients (62%), but their timing and methods of inter-view differed from ours. In a recent study, de Kruijk et al. (2002a) found that elevatedS100B and NSE values on admission predicted more severe PCSs at six months postinjury, and those who had headache, dizziness, or nausea shortly on admission alsoreported more severe PCSs (de Kruijk et al. 2002a). Our results are in good accordancewith these findings, since we also found headache and dizziness on admission to predictthe development of PCSs, although we also found skull or facial bone fracture to be anindependent and important risk factor for PCSs.

Our results do not support the notion of an association between heavy alcohol drink-ing or illicit drug use and PCSs, although it has been suggested that alcohol misuse couldplay a role in their development (Lishman 1988), neither did we find significant associa-

61

tions between other possible risk factors such as gender, prior disease, and factors predis-posing to malingering (insurance, unemployment, etc.) and PCSs. Litigation is also oftenthought to be a risk factor for PCSs (Bernstein 1999, Mickeviciene et al. 2002), but ourfindings do not support this hypothesis, although they do not exclude the possibility thatlitigation may exacerbate the symptoms.

6.2 Protein S100B as a marker of brain damage

Serum protein S100B has recently been an object of increasing attention in the identifica-tion of brain damage because of its high brain tissue specificity. Its clinical utility withhead trauma patients has not yet been fully elucidated, however.

Acute traumatic brain injury has been found to cause a leakage of S100B into theserum (Ingebrigtsen et al. 1999, Raabe et al. 1999b, II), and the severity of brain injuryhas been found to correlate positively with S100B values, the highest levels beingreported in patients with moderate-to-severe injuries (Raabe et al. 1999b, Romner et al.2000, II). Brain lesions found in head CT scans have also been shown to be associatedwith elevated S100B (Raabe et al. 1999b, Ingebrigtsen et al. 2000b). Romner et al.(2000), studying 278 acute head injury patients with minor, moderate and severe injuries,found that S100B was elevated (>0.20 µg/L) in 92% of those with intracranial lesions intheir head CT scans and 34% of those without. In fact, serum S100B with a cut-off of0.20 µg/L was extremely good at excluding intracranial pathology visible in a head CTscan, since its negative predictive value was as high as 99%.

A positive correlation has also been reported between elevated S100B values and poorrecovery in patients with head injury (Raabe et al. 1999a, Woertgen et al. 1999, Townendet al. 2002). Townend et al. (2002) studied 119 such patients (with an initial GCS score of4–15) and found that a serum S100B concentration of > 0.32 µg/L predicted severe dis-ability (extended GOS < 5) at one month after trauma with a sensitivity of 93%, a speci-ficity of 72%, and a negative predictive value of 99%. Thus, patients with a concentra-tion of S100B > 0.32 µg/L and poor recovery at one month post injury must be extremelyrare.

S100B levels have been shown to be elevated in 20–42% of patients with MHI (Inge-brigtsen et al. 2000b, de Kruijk et al. 2001c, de Kruijk et al. 2002a, I, II). In addition, apoor neuropsychological outcome has been reported to be associated with increasedS100B concentrations in MHI patients (Waterloo et al. 1997, Herrman et al. 2001), andS100B levels have been shown to be higher in patients who will develop PCSs than inthose who will not (Ingebrigtsen et al. 2000b, de Kruijk et al. 2002a). Furthermore, weshowed in paper I that S100B was an independent and good predictor of PCSs. When acut-off value of 0.50 µg/L was used, the sensitivity, specificity and negative predictivevalue for PCSs were 27%, 93% and 82%, respectively, while if a lower cut-off value(0.20 µg/L) was used, the corresponding figures were 68%, 67% and 88%. The MHIpatients with S100B values of < 0.20 µg/L shortly after trauma rarely developed PCSs.

One of our objectives was to study the effects of extracranial injuries on S100B levelsin serum (II). We showed that not only brain damage but also extracranial injuries, includ-

62

ing small ones, increased serum S100B values in patients with multitrauma, althoughonly large extracranial injuries elevated S100B values significantly, as reported by others(Andersson et al. 2001). Although small extracranial injuries (i.e. wounds, soft tissue con-tusions, distensions, luxations and small fractures) also increased serum S100B values,the effect was not found to be clinically significant (II). Thus it is evident that extracranialinjuries cause some false positive findings related to S100B, at least if a large extracra-nial injury is present. However, normal S100B values (< 0.20 µg/L) predict the absence ofintracranial damage and good recovery after trauma with a high accuracy. Thus, S100Bcan be used as a marker to exclude intracranial pathology (Romner et al. 2000), poor out-come (Townend et al. 2002) and the development of PCSs (I) in patients with headtrauma.

6.3 Alcohol and trauma

6.3.1 Hazardous alcohol drinking in patients with trauma

There is no doubt that alcohol and injuries are closely linked to each other. Approximate-ly one half of all trauma patients are BAC-positive on admission, and more than a thirdhave a BAC at or above 100mg/dL (Rivara et al. 1993a, III). The proportion of BAC-pos-itive cases among patients with head injury has been found to be over 50%, and the pro-portion of those with a BAC at or above 100mg/dL has been shown to vary from 35 to40% (Rimel et al. 1982, Brismar et al. 1983, Corrigan 1995, Dikmen et al. 1995, III).

The methods used to detect heavy alcohol drinking and alcoholism in previous studiesof trauma patients have varied greatly (Corrigan 1995, Nilssen et al. 1994), the lowestincidences, 16% (Rimel et al. 1982), 25% (Sparadeo & Gill 1989) and 36% (Wong et al.1993), having been found in studies where the assessments were based on retrospectivechart reviews, while the highest ones, 66% (Kreutzer et al. 1991) and 58% (Kreutzer et al.1990), have been reported from rehabilitation centres. Nilssen et al. (1994) reported anapproximately 45% incidence of alcohol abuse in trauma patients based on screening withthe CAGE and SMAST questionnaires.

We found that binge-type drinking was the predominant pattern among the traumapatients. Patterns of alcohol consumption have so far not received much attention. We fur-ther showed, however, that a total of two-thirds of our trauma patients reported hazardousdrinking (i.e. they were dependent drinkers or frequent binge drinkers), and hazardousdrinking was found to be more frequent in the patients with head injury that in those withother types of trauma (III).

The high prevalence of binge drinkers among trauma patients is alarming. It suggeststhat more attempts should be made to reduce this type of drinking. If alcohol consumptionis assessed in terms of grammes of ethanol consumed per day, most binge drinkers will beclassified as moderate drinkers and not as hazardous ones (Naimi et al. 2003), but it isprecisely this pattern of binge drinking that is closely associated with alcohol-impaireddriving (Liu et al. 1997), and there is growing evidence that the social, health and eco-

63

nomic costs of acute alcohol-related problems may even exceed those arising from thechronic effects (CDC 1993, Chikritzhs et al. 1993). The highest prevalence of bingedrinking has been found among young adults, but it is also a frequent and favoureddrinking pattern among adolescents (American Academy of Pediatrics Committee onSubstance Abuse 1995, Naimi et al. 2003). Binge drinking is an increasing problemworldwide (Mäkelä et al. 2001, Naimi et al. 2003), and the present findings suggest that itis a substantial, and growing medical hazard, and that more interventions should beemployed in attempts to prevent it.

Fleming et al. (1997) found that brief advice from a physician reduced binge drinkingby approximately 40% in a randomised controlled trial, and the reduction was found tolast for at least 12 months. Thus, if alcohol interventions are to be used, they should befocused on this group, because most of them are young adults and not heavy drinkers, andtherefore the intervention could come before the patients have reached a stage of severealcohol dependence.

6.3.2 Alcohol as a risk factor for head trauma

Previous reports have emphasized the high prevalence of alcohol intoxication amongtrauma patients, but not the high frequency of binge drinkers. Our data emphasise theadverse consequences of binge drinking, which may cause severe intoxication in individ-uals who are not regular drinkers and who do not have increased alcohol tolerance. Thusthe more frequent binge drinking is, the greater is the risk of trauma. Accordingly, thelower incidence of trauma observed here in infrequent binge drinkers than in frequentones is apparently due to their fewer occasions of alcohol intoxication (III). WhereMäkelä et al. (2001) found that the mean frequency of binge drinking in Finland wasapproximately 11 times a year, 61% of all our trauma patients reported at least 12 suchepisodes a year prior to injury. The high binge drinking rate among trauma patients sup-ports the view that this type of drinking is a more significant risk factor for trauma thanhas been previously acknowledged.

We showed that the risk of sustaining a head trauma increased as a function of increas-ing blood alcohol level, and significantly so above the level of 1.5‰ (III). The most com-mon causes of injury among the BAC-positive patients were assaults, falls on the groundand bicycling injuries. Alcohol is likely to cause deleterious effects on psychomotor skillsand the preventive mechanisms to respond to situational hazards which, in turn, mayfavour the occurrence of head trauma. Consistently, BAC-positive patients injured byfalls on the ground or in bicycling accidents injured their head significantly more oftenthan did their sober counterparts (III). In contrast, sober bicyclists more often avoidedhead trauma and injured their extremities (III).

McLeod et al. (1999) reported that consuming more than 60 grammes of alcohol in a6-hour period conferred an odds ratio of 3.4 (95% CI 1.8 to 6.4) for sustaining an injuryin general, and increased risks of more severe injuries under the influence of alcohol havealso been found in certain specific traumas such as bicycling injuries (Olkkonen & Hon-kanen 1990, Li et al. 2001), motor vehicle crashes (Waller et al. 1986), motor vehicle-

64

related pedestrian injuries (Rivara et al. 1997), assaults (Martin & Bachman 1997, Kyria-cou et al. 1998) and injuries sustained in falls (Honkanen et al. 1976).

6.3.3 Identification of alcohol misuse in patients with trauma

A positive effect of a brief alcohol intervention as a means of reducing alcohol intake andthe adverse health effects has been reported in several studies (Walsh et al. 1991, Flem-ing et al. 1997, Gentilello et al. 1999, Dinh-Zarr et al. 2000), but there still continues tobe a lack of attention to hazardous alcohol drinking in accident and emergency depart-ments (Soderstrom & Cowley 1987, Charalambous 2002). One apparent reason for this isthat hazardous drinkers usually remain unrecognised by the clinicians. A simple and inex-pensive method for the early identification of hazardous drinkers would thus be of theutmost importance.

Ryb et al. (1999) reported that the BAC test is the best detector of alcohol dependencein trauma patients, while GGT, AST and MCV are of little value as screening tests. Wealso found that the conventional biochemical markers (GGT, MCV, CDT, and AST)lacked sensitivity and specificity, especially for detecting binge drinkers, but were of theopinion that the BAC test, performed either on exhaled air or on a blood sample, is thebest method of detecting hazardous alcohol drinking (i.e. dependent drinking and frequentbinge drinking) among trauma patients (IV). BAC not only identifies acute alcohol drink-ing but also provides a good estimate of chronic patterns of hazardous drinking. BAC (>0mg/dL) on admission was found to be a sensitive (68%) and specific (94%) marker ofhazardous drinking, and 96% of all the BAC-positive trauma patients turned out to behazardous drinkers. Thus, the BAC test can be recommended as a primary screening toolto guide patients for alcohol interventions without subjecting them to an unacceptabledegree of stigma as problem drinkers and before they have reached a stage of severedependence.

6.4 Strengths and weaknesses

The material was collected from the emergency department of a hospital which is theonly trauma centre for Oulu, a city with approximately 120,000 inhabitants. The patientswere recruited consecutively seven days a week and 24 hours a day, but with certainexclusion criteria applied.

Firstly, patients representing the age range 16–49 were recruited. This was done,because the highest rates of morbidity, mortality and persistent functional and psychologi-cal impairment due to trauma are known to occur in this group. Secondly, many commondiseases such as cardiovascular and neurological diseases are frequent in the elderly, andlittle is known about their effect on S100B levels. Additionally, life-style factors such asalcohol drinking habits are likely to differ between our material and cases of children orelderly people, and as our material includes acute trauma patients, the results of seriesderived from out-patient clinics may differ from ours.

65

The interviews were always performed by the same person. The first interview to placeface to face, and the second, which occurred 8 to 30 months post injury was a telephoneinterview aimed at ensuring that the PCSs had lasted long enough to fulfil our criteria (i.e.one month post injury). The long interval between trauma and the interview may havecaused underestimation of certain symptoms, particularly impairment of cognitive func-tions.

Of the 385 patients, 349 (91%) were interviewed within 2–6 weeks of the trauma(including alcohol data) and PCSs data was collected from 86% of the MHI patients (172/199). Thus, the percentage of patients who were interviewed was high, higher than inmany other studies concerning PCSs or alcohol drinking (Ingebrigtsen et al. 1998, Ryb etal. 1999, Ingebrigtsen et al. 2000b).

In conclusion, the material represents a good spectrum of consecutive Finnish traumapatients in a general hospital, excluding children and elderly people. Because the resultsconcerning protein S100B and the alcohol literature from other countries is quite compa-rable to ours in this respect, it can be assumed that the results can also be generalisedworldwide.

7 Conclusions

1. PCSs often complicate recovery after mild head injury. Skull or facial bone fractures,elevated serum protein S100B and headache or dizziness on admission are associatedwith an increased risk of developing PCSs. These simple criteria may help the physi-cian on duty to identify patients for counselling and follow-up.

2. Serum protein S100B is a feasible supplementary method for detecting traumaticbrain damage. In particular, normal levels of S100B after head injury predict normalhead CT scan findings, good overall recovery and the absence of PCSs in the follow-up.

3. In addition to brain injury, extracranial injuries also increase serum S100B levels andmay cause false-positive findings. S100B is not a specific marker of brain injury.

4. Alcohol intoxication indicative of hazardous alcohol drinking (i.e. dependent drin-king and frequent binge drinking) is a common finding in trauma patients, and morecommon in head injury patients than in those with other types of trauma. Binge drin-king is the predominant pattern of alcohol drinking in trauma patients, and particu-larly in head trauma patients.

5. The risk of sustaining a head injury significantly increases with increasing BAC, andis significant if BAC exceeds 150 mg/dL.

6. Hazardous alcohol drinkers, i.e. dependent drinkers and frequent binge drinkers, canbest be identified by means of a BAC test on admission. The conventional laboratorymarkers of alcohol consumption such as CDT, GGT, AST, or MCV do not offer anysignificant benefit in patients with trauma.

7. BAC should be measured, either from exhaled air or blood, in the case of all thetrauma patients on admission. If there is an option for brief alcohol intervention, allBAC-positives should be guided for such counselling.

8. The results point to the usefulness of the BAC test as a marker of alcohol misuse andprotein S100B as a marker of brain damage in patients with trauma.

References

Abdel-Dayem HM, Abu-Judeh H, Kumar M, Atay S, Naddaf S, El-Zeftawy H & Luo J-Q (1998)SPECT brain perfusion abnormalities in mild or moderate traumatic brain injury. Clin Nucl Med23: 309–317.

Adams JH (1975) The neuropathology of head injuries. In: Vinken PJ & Bruyn GW (eds) Handbookof Clinical Neurology 23: 35–65. North-Holland Publ. Co., Amsterdam.

Adams JH, Doyle D, Ford I, Gennarelli TA, Graham DI & Mclellan DR (1989) Diffuse axonal injury:definition, diagnosis and grading. Histopathology 15:49–59.

Adams JH, Graham DI, Murray LS & Scott G (1982) Diffuse axonal injury due to nonmissile headinjury in humans: An analysis of 45 cases. Ann Neurol 12: 557–563.

Aikuisiän aivovammojen käypä hoito (2003) Suomen Neurologinen Yhdistys, Suomen Neurokirur-ginen Yhdistys, Suomen Fysiatriyhdistys, Suomen Vakuutuslääkärien Yhdistys ja Suomen Neu-ropsykologinen Yhdistys. Duodecim 7: 654–81.

Alexander MP (1995) Mild traumatic brain injury: pathophysiology, natural history, and clinicalmanagement. Neurology 45: 1253–1260.

Allen JP, Fertig JB, Litten RZ, Sillanaukee P & Anton RF (1997) Proposed recommendations for re-search on biochemical markers for problematic drinking. Alcohol Clin Exp Res 21: 244–247.

Alves W, Macciocchi SN & Barth JT (1993) Postconcussive symptoms after uncomplicated mildhead injury. J Head Trauma Rehabil 8: 48–59.

American Academy of Pediatrics Committee on Substance Abuse (1995) Alcohol use and abuse: apediatric concern. Pediatrics 95: 439–442.

Anderson RE, Hansson L-O, Nilsson O, Dijlai-Merzoug R, Settergren G (2001) High serum S100Blevels for trauma patients without head injuries. Neurosurgery 48: 1255–1260.

Angell M & Kassirer JP (1994) Alcohol and other drugs–toward a more rational and consistent poli-cy. N Engl J Med 331: 537–539

Annegers JF, Grabow JD, Groover RV, Laws ER Jr, Elveback LR & Kurland LT (1980) Seizures af-ter head trauma: a population study. Neurology 30: 683–689.

Annegers JF, Grabow JD, Kurland LT & Laws ER (1980) The incidence, causes, and secular trendsof head trauma in Olmsted County, Minnesota, 1935–1974. Neurology 30: 912–919.

Annegers JF, Hauser WA, Coan SP & Rocca WA (1998) A populations-based study of seizures aftertraumatic brain injuries. N Engl J Med 338: 20–24.

Antti-Poika I, Karaharju E, Roine R & Salaspuro M (1988) Intervention of heavy drinking—a pro-spective and controlled study of 438 consecutive injured male patients. Alcohol Alcohol 23: 115–121.

68

Ashikaga R, Araki Y & Ishida O (1997) MRI of head injury using FLAIR. Neuroradiology 39: 239–242.

Attia J & Cook DJ (1998) Prognosis in anoxic and traumatic coma. Crit Care Clin 14: 497–511.Bagley LJ (1999) Imaging of neurological emergencies: Trauma, hermorrhage, and infarction. Semin

Roentgenol 34: 144–159.Bakay RA & Ward AA Jr (1983) Enzymatic changes in serum and cerebrospinal fluid in neurological

injury. J Neurosurg 58: 27–37.Baker SP, O’Neill B, Haddon W & Long WB. (1974) The injury severity score: a method for describ-

ing patients with multiple injuries and evaluating emergency care. J Trauma 14: 187–196.Bayston R, de Louvois J, Brown EM, Johnston RA, Lees P & Pople IK (2000) Use of antibiotics in

penetrating craniocerebral injuries. “infection in neurosurgery” Working party of British Societyfor Antimicrobial Chemotherapy. Lancet 355: 1813–1817.

Bazarian JJ, Wong T, Harris M, Leahey N, Mookerjee S & Dombovy M (1999) Epidemiology andpredictors of post-concussive syndrome after minor head injury in an emergency population.Brain Inj 13: 173–189.

Bendtsem P, Hultberg J, Carsllon M & Jones AW (1999) Monitoring ethanol exposure in a clinicalsetting by analysis of blood, breath, saliva, and urine. Alc Clin Exp Res 23: 1446–1451.

Bernstein DM (1999) Recovery from mild head injury. Brain Inj 13: 151–172.Bešenski N (2002) Traumatic injuries: imaging of head injuries. Eur Radiol 12: 1237–1252.Biberthaler P, Mussack T, Wiedemann E, Gilg T, Soyka M, Koller G, Pfeifer KJ, Linsenmaier U,

Mutschler W, Gippner-Steppert C & Jochum M (2001) Elevated serum levels of S-100B reflectthe extent of brain injury in alcohol intoxicated patients after mild head trauma. Shock 16: 97–101.

Binder LM (1997) A review of mild head trauma. Part II: Clinical implications. J Clin Exp Neuropsy-chol 19: 432–457.

Binder LM, Rohling ML & Larrabee GL (1997) A review of mild head trauma. Part I: meta-analyticreview of neuropsychological studies. J Clin Exp Neuropsychol 19: 421–431.

Bohnen N & Jolles J (1992) Neurobehavioral aspects of postconcussive symptoms after mild headinjury. J Nerv Ment Dis 180: 683–692.

Bombardier CH Edhe D & Kilmer J (1997a) Readiness to change alcohol drinking habits after trau-matic brain injury. Arch Phys Med Rehabil 78: 592–596.

Bombardier CH, Kilmer J & Ehde D (1997b) Screening for alcoholism among persons with recenttraumatic brain injury. Rehabil Psychol 42: 259–271.

Bombardier CH & Thurber CA (1998) Blood alcohol level and early cognitive status after traumaticbrain injury. Brain Inj 12: 725–734.

Borges G, Cherpitel CJ, Mondragon L, Poznyak V, Peden M & Gutierrez I (2004) Episodic alcoholuse and risk of nonfatal injury. Am J Epidemiol 159: 565–571.

Brismar B, Engström A & Rydberg U (1983) Head injury and intoxication: a diagnostic and thera-peutic dilemma. Acta Chir Scand 149: 11–14.

Brooks N, Symington C, Beattie A, Campsie L, Bryden J & McKinlay W (1989) Alcohol and otherpredictors of congnitive recovery after severe head injury. Brain Inj 3: 235–246.

Bullock R, Chesnut RM, Clifton G, Ghajar J, Marion DW, Narayan RK, Newell DW, Pitts LH, Ros-ner MJ & Wilberger JW (1996) Guideline for the management of severe head injury. Eur J EmergMed 2: 109–127.

Bullock R, Golek J & Blake G (1989) Traumatic intracerebral hematoma--which patients should un-dergo surgical evacuation? CT scan features and ICP monitoring as a basis for decision making.Surg Neurol 32: 181–187.

Bullock R, Hannemann CO, Murray L, Graham M & Teasdale M (1990) Recurrent hematomas fol-lowing craniotomy for traumatic intracranial mass. J Neurosurg 72: 9–14.

Bullock R, Maxwell WL, Graham DI, Teasdale GM & Adams JH (1991) Glial swelling followinghuman cerebral contusion: an ultrastructural study. J Neurol Neurosurg Psychiatry 54: 427–434.

69

Caroli M, Locatelli M, Campanella R, Balbi S, Martinelli F & Arienta C (2001) Multiple intracraniallesions in head injury: clinical considerations, prognostic factors, management, and results in 95patients. Surg Neurol 56: 82–88.

Cattelani R, Gugliotta M, Maravita A & Mazzucchi A (1996) Post-concusive syndrome: paraclinicalsigns, subjective symptoms, cognitive functions and MMPI profiles. Brain Inj 10: 187–195.

Cecil KM, Hills EC, Sandel ME, Smith DH, McIntosh TK, Mannon LJ, Sinson GP, Bagley LJ, Gros-mman RI & Lenkinski RE (1998) Proton magnetic resonance spectroscopy for detection of axonalinjury in the splenium of the corpus callosum of brain-injured patients. J Neurosurg 88: 795–801.

CDC (1993) Economic impact of motor-vehicle crashes–United States, 1990. MMWR 42: 443–448.Charalambous MP (2002) Alcohol and the accident and emergency department: a current review. Al-

cohol Alcoholism 37: 307–312.Chen JCT & Levy ML (2000) Causes, epidemiology, and risk factors of chronic subdural hematoma.

Neurosurg Clin N Am 11: 399–406.Cherner M, Temkin NR, Machamer JE & Dikmen SS (2001) Utility of a composite measure to detect

problematic alcohol use in persons with traumatic brain injury. Arch Phys Med Rehabil 82: 780–786.

Cherpitel CJ, Tam T, Mdanik L, Caetano R & Greenfield T (1995) Alcohol and non-fatal injury inthe U.S. general population: a risk function analysis. Accid Anal Prev 27: 651–661.

Chikritzhs TN, Jonas HA, Stockwell TR, Heale PF & Dietze PM (2001) Mortality and life-years lostdue to alcohol: a comparison of acute and chronic causes. Med J Aust 174: 281–284.

Choi SC, Narayan RK, Anderson RL & Ward JD (1988) Enhanced specificity of prognosis in severehead injury. J Neurosurg 69: 381–385.

Clifton GL, Grossman RG, Makela ME, Miner ME, Handel S & Sadhu V (1980) Neurological courseand correlated computerized tomography findings after severe closed head injury. J Neurosurg 52:611–624.

Cooper EH (1994) Neuron-specific enolase. Int J Biol Markers 9: 205–210.Copes WS, Champion HR, Sacco WJ, Lawnick MM, Keast SL & Bain LW (1988) The injury severity

score revisited. J Trauma 28: 69–77.Corrigan JD (1995) Substance abuse as a mediating factor in outcome from traumatic brain injury.

Arch Phys Med Rehabil 76: 302–309.Criteria for the diagnosis of alcoholism by the Criteria Committee, National Council on Alcoholism

(1972) Am J Psychiatry 129: 127–135.Dacey RG Jr, Alves WM, Rimel RW, Winn HR & Jane JA (1986) Neurosurgical complications after

apparently minor head injury. J Neurosurg 65: 203–210.Dacey RG Jr, Vollmer D & Dikmen SS (1993) Mild head injury. In: Cooper PR (ed) Head Injury 8:

159–182. Williams&Wilkins, Baltimore.Davalos DB & Bennett TL (2002) A review of the use of single-photon emission computed tomog-

raphy as a diagnostic tool in mild traumatic brain injury. Appl Neuropsychol 9: 92–105.Davis LJ Jr, Hurt RD, Morse RM & O’Brien PC (1987) Discriminant analysis of the Self-Adminis-

tered Alcoholism Screening Test. Alcohol Clin Exp Res 11: 269–273.Dikmen SS, Machamer JE, Donovan DM, Winn HR & Temkin NR (1995) Alcohol use before and

after traumatic head injury. Ann Emerg Med 26: 167–176.Dinh-Zarr T, Diguiseppi C, Heitman E & Roberts I (1999) Preventing injuries through interventions

for problem drinking: a systematic review of randomized controlled trials. Alcohol Alcohol 34:609–621.

Dinh-Zarr T, DiGuiseppi C, Heitman E & Roberts I (2000) Interventions for preventing injuries inproblem drinkers. Cochrane Database Syst Rev 2: CD0011857.

Dunn C, Donovan D, Gentilello LM (1997) Practical guidelines for performing alcohol interventionsin trauma centers. J Trauma 42: 299–304.

70

Durkin MS, Olsen S, Barlow B, Virella A & Connolly ES Jr (1998) The epidemiology of urban pedi-atric neurological trauma: evaluation of, and implications for, injury prevention programs. Neu-rosurgery 42: 300–310.

Eisenberg HM, Gary HE Jr, Aldrich EF, Saydjari C, Turner B, Foulkes MA, Jane JA, Marmarou A,Marshall LF & Young HF (1990) Initial CT findings in 753 patients with severe head injury. Areport from the NIH Traumatic Coma Data Bank. J Neurosurg 73: 688–698.

Elson LM & Ward CC (1994) Mechanisms and pathophysiology of mild head injury. Semin Neurol14: 8–18.

Ewing JA (1984) Detecting alcoholism. The CAGE questionnaire. JAMA 252: 1905–1907.Facco E (1999) The role of EEG in brain injury. Intensive Care Med 25: 872–877.Farace E & Alves WM (2000) Do women fare worse: a metaanalysis of gender differences in trau-

matic brain injury outcome. J Neurosurg 93: 539–545.Field CA, Claassen CA & O’Keefe G (2001) Association of alcohol use and other high-risk behaviors

among trauma patients. J Trauma 50: 13–19.Fiser SM, Johnson SB & Fortune JB (1998) Resource utilization in traumatic brain injury: The role

of magnetic resonance imaging. Am Surg 64: 1088–1093.Fleming MF, Lawton Barry K, Baier Manwell L, Johnson K & London R (1997) Brief physician ad-

vice for problem alcohol drinkers. A randomized controlled trial in community-based primarycare practices. JAMA 277: 1039–1045.

Fogelholm R & Waltimo O (1975) Epidemiology of chronic subdural haematoma. Acta Neurochir(Wien) 32: 247–250.

Fortuny LAI, Briggs M, Newcombe F, Ratcliff G & Thomas C (1980) Measuring the duration of posttraumatic amnesia. J Neurol Neurosurg Psychiatry 43: 377–379.

Galbraith S, Murray WR, Patel AR & Knill-Jones R (1976) The relationship between alcohol andhead injury and its effect on the conscious level. Br J Surg 63: 128–130.

Gardner MJ & Altman DG (1989) Statistics with confidence. British Medical Journal, London.Gardnett MR, Cadoux-Hudson T & Syles P (2001) How useful is magnetic resonance imaging in pre-

dicting severity and outcome in traumatic brain injury? Curr Opin Neurol 14: 753–757.Gasquoine PG (1997) Postconcussion symptoms. Neuropsychology Rev 7: 77–85.van Geel WJ, de Reus HP, Nijzing H, Verbeek MM, Vos PE & Lamers KJ (2002) Measurement of

glial fibrillary acidic protein in blood: an analytical method. Clin Chim Acta 326: 151–154.Geets W & Zegher F (1985) EEG and brainstem abnormalities after cerebral concussion. Acta Neurol

Belg 85: 277–283.Gennarelli TA (1993) Cerebral concussion and diffuse brain injuries. In: Coorper PR (ed) Head Injury

7: 137–158.Gennarelli TA, Champion HR, Sacco WJ, Copes WS & Alves WM (1989) Mortality of patients with

head injury and extracranial injury treated in trauma centers. J Trauma 29: 1193–1202.Gennarelli TA, Thibault LE, Adams JH, Graham DI, Thompson CJ & Marcincin RP (1982) Diffuse

axonal injury and traumatic coma in the primate. Ann Neurol 12: 564–74.Gentilello LM, Rivara FP, Donovan DM, Jurkovich GJ, Daranciang E, Dunn CW, Villaveces A, Co-

pass M & Ries RR (1999) Alcohol interventions in a trauma center as a means of reducing the riskof injury recurrence. Ann Surg 230: 473–483.

Gentilello LM, Rivara FP, Donovan DM, Villaveces A, Daranciang E, Dunn CW & Ries RR (2000)Alcohol problems in women admitted to a level I trauma center: a gender-based comparison. JTrauma 48: 108–114.

Gentry LR, Godersky JC & Thompson B (1988a) MR imaging of head trauma: Reviev of the distri-bution and radiopathologic features of traumatic lesions. Am J Neuroradiol 9: 101–110.

Gentry LR, Godersky JC, Thompson B & Dunn VD (1988b) Prospective comparative study of inter-mediate-fiel MR and CT in the evaluation of closed head trauma. Am J Neuroradiol 9: 91–100.

71

Graham DI, Ford I, Adams JH, Doyle D, Teasdale GM, Lawrence AE & McLellan DR (1989) Ischae-mic brain damage is still common in fatal non-missile head injury. J Neurol Neurosurg Psychiatry52: 346–350.

Gronwall D & Wrightson P (1981) Memory and information processing capacity after closed headinjury. J Neurol Neursurg Psychiatry 44: 889–895.

Groswasser Z, Reider-Groswasser I, Soroker N & Machtey Y (1987) Magnetic resonance imaging inhead injured patients with normal late computed tomography scans. Surg Neurol 27: 331–337.

Gurney JG, Rivara FP, Mueller BA, Newell DW, Copass MK & Jurkovich GJ (1992) The effects ofalcohol intoxication on the initial treatment and hospital course of patients with acute brain injury.J Trauma 33: 709–713.

Hadfield RJH, Mercer M & Parr MJA (2001) Alcohol and drug abuse in trauma. Resuscitation 48:25–36.

Hadley DM, Teasdale GM, Jenkins A, Condon B, MacPherson P, Pattersson J & Rowan JO (1988)Magnetic resonance imaging in acute head injury. Clin Radiol 39: 131–139.

Haimoto H, Hosoda S & Kato K (1987) Differential distribution of immunoreative S100-α and S100-β proteins in normal nonnervous human tissues. Lab Invest 57: 489–498.

Hammoud DA & Wasserman BA (2002) Diffuse axonal injuries: pathophysiology and imaging. Neu-roimaging Clin N Am 12: 205–216.

Hartz A, Guse C & Kajdacsy-Balla A (1997) Identification of heavy drinkers using a combination oflaboratory tests. J Clin Epidemiol 50: 1357–1368.

Haydel MJ, Preston CA, Mills TJ, Luber S, Blaudeau E & DeBlieux PM (2000) Indications for com-puted tomography in patients with minor head injury. N Engl J Med 343: 100–105.

Herrmann M, Curio N, Jost S, Grubich C, Ebert AD, Fork ML & Synowitz H (2001) Release of bio-chemical markers of damage to neuronal and glial brain tissue is associated with short and longterm neuropsychological outcome after traumatic brain injury. J Neurol Neurosurg Psychiatry 70:95–100.

Herrmann M, Jost S, Kutz S, Ebert AD, Kratz T, Wunderlich MT & Synowitz H (2000) Temporalprofile of release of neurobiochemical markers of brain damage after traumatic brain injury is as-sociated with intracranial pathology as demonstrated in cranial computerized tomography. J Neu-rotrauma 17: 113–122.

Hesselink JR, Dowd CF, Healy ME, Hajek P, Baker LL & Luerssen TG (1988) MR imaging of braincontusions: A comparative study with CT. Am J Neuroradiol 9: 269–278.

Hidaka H, Endo T, Kawamoto S, Yamada E, Umekawa H, Tanabe K & Hara K (1983) Purificationand characterization of adipose tissue S-100b protein. J Biol Chem 258: 2705–2709.

Hofman PA, Nelemans P, Kemerink GJ & Wilmink JT (2000) Value of radiological diagnosis ofskull fracture in the management of mild head injury: meta-analysis. J Neurol Neurosurg Psychi-atry 68: 416–422.

Hofman PA, Stapert SZ, van Kroonenburgh MJ, Jolles J, de Kruijk JR & Wilmik JT (2001) MR im-aging, single-photon emission CT, and neurocognitive performance after traumatic brain injury.Am J Neuroradiol 22: 441–449.

Holder HC, Gruenewald PJ, Ponicki WR, Treno AJ, Grube JW, Saltz RF, Voas RB, Reynolds R, Dav-is J, Sanchez L, Gaumont G & Roeper P (2000) Effect of community-based interventions on high-risk drinking and alcohol-related injuries. JAMA 284: 2341–2347.

Holder H, Longabaugh R, Miller WR & Rubonis AV (1991) The cost effectiveness of treatment foralcoholims: a first approximation. J Stud Alcohol 52: 517–540.

Honkanen R, Ertama L, Kuosmanen P, Linnoila M & Visuri T (1976) A case-control study on alcoholas a risk factor in pedestrian accidents. Mod Probl Pharmacopsychiatry 11: 46–50.

Hurt RD, Morse RM & Swenson WM (1980) Diagnosis of alcoholism with a self-administered alco-holism screening test: results with 1,002 consecutive patients receiving general examinations.Mayo Clin Proc 55: 365–370.

72

International Statistical Classification of Diseases and Related Health Problems (1992) Tenth Revi-sion. Geneva.

Ichise M, Chung D-G, Wang P, Wortzman G, Gray BG & Franks W (1994) Technetium-99m-HM-PAO SPECT, CT and MRI in the evaluation of patients with chronic traumatic brain injury: A cor-relation with neuropsychological performance. J Nucl Med 35: 217–226.

Ingebrigtsen T (1998) Aspects on the management of minor head injury. Thesis, University of Trom-sø, Institute of Clinical Medicine, Department of Neurosurgery.

Ingebrigtsen T & Romner B (2002) Biochemical serum markers of traumatic brain injury. J Trauma52: 798–808.

Ingebrigtsen T, Romner B & Kock-Jensen C (2000a) Scandinavian guidelines for initial managementof minimal, mild, and moderate head injuries. J Trauma 48: 760–766.

Ingebrigtsen T, Romner B, Kongstad P & Langbakk B (1995) Increased serum concentrations of pro-tein S-100 after minor head injury: a biochemical serum marker with prognostic value? J NeurolNeurosurg Psychiatry 59: 103–104.

Ingebrigtsen T, Romner B, Marup-Jensen S, Dons M, Lundqvist C, Bellner J, Alling C & BørgesenSE (2000bbThe clinical value of serum S-100 protein measurements in minor head injury: a Scan-dinavian multicentre study. Brain Inj 14: 1047–1055.

Ingebrigtsen T, Romner B & Trumpy JH (1997) Management of minor head injury: the value of earlycomputed tomography and serum protein S-100 measurements. J Clin Neurosci 4: 29–33.

Ingebrigtsen T, Waterloo K, Jacobsen EA, Langbakk B & Romner B (1999) Traumatic brain damagein minor head injury: Relation of serum S-100 protein measurements to magnetic resonance im-aging and neurobehavioral outcome. Neurosurgery 45: 468–476.

Ingebrigtsen T, Waterloo K, Marup-Jensen S, Attner E & Romner B (1998) Quantification of post-concussioin symptoms 3 months after minor head injury in 100 consecutive patients. J Neurol 245:609–612.

Ito H, Ishii K, Onuma T, Kawashima R & Fukuda H (1997) Cerebral perfusion changes in traumaticdiffuse brain injury; IMP SPECT studies. Ann Nucl Med 11: 167–172.

Jagger J, Fife D, Vernberg K & Jane JA (1984) Effect of alcohol intoxication on the diagnosis andapparent severity of brain injury. Neurosurgery 15: 303–306.

Jallo JI & Narayan RK (2000) Craniocerebral trauma. In: Bradley GB, Daroff RB, Fenichel GM &Marsden CD (eds) Neurology in Clinical Practice 56: 1055–1111.

Jane JA, Steward O & Gennarelli T (1985) Axonal degeneration induced by experimental noninva-sive minor head injury. J Neurosurg 62: 96–100.

Jennett B & Bond M (1975) Assessment of outcome after severe brain damage. Lancet 1: 480–484.Jennett B & MacMillan R (1981) Epidemiology of head injury. Br Med J 282: 101–104.Jennett B, Snoek J, Bond MR & Brooks N (1981) Disability after severe head injury: observations on

the use of the Glasgow Coma Scale. J Neurol Neurosurg Psychiatry 44: 285–293.Johnsson P (1996) Markers of cerebral ischemia after cardiac surgery. J Cardiolthorac Vasc Anesth

10: 120–126.Jurkovich GJ, Rivara FP, Gurney JG, Seguin D, Fligner CL & Copass M (1992) Effects of alcohol

intoxication on the initial assessment of trauma patients. Ann Emerg Med 21: 704–708.Jönsson H, Johnsson P, Höglund P, Alling C & Blomquist S (2000) Elimination of S100B and renal

function after cardiac surgery. J Cardiothorac Vasc Anesth 14: 698–701.Kakarieka A, Braakman R & Schakel EH (1995) Subarachnoidal haemorrhage after head injury. Cer-

ebrovasc Dis 5: 403–406.Karzmark P, Hall K & Englander J (1995) Late-onset post-concussion symptoms after mild brain in-

jury: the role of premorbid, injury-related, environmental, and personality factors. Brain Inj 9: 21–26.

Kato K & Kimura S (1985) S100a0 (αα) protein is mainly located in the heart and striated muscles.Biochim Biophys Acta 842: 146–150.

73

Kelly AB, Zimmerman RD, Snow RB, Gandy SE, Heier LA & Deck MDF (1988) Head trauma:Comparison of MR and CT-Experience in 100 patients. Am J Neuroradiol 9: 699–708.

Kelly DF (1995) Alcohol and head injury: an issue revisited. J Neurotrauma 12: 883–890.Kelly DF, Lee SM, Pinanong PA & Howda DA (1997a) Paradoxical effects of acute ethanolism in

experimental brain injury. J Neurosurg 86: 876–882.Kelly MP, Johnson CT, Knoller N, Drubach DA & Winslow MM (1997b) Substance abuse, traumatic

brain injury and neuropsychological outcome. Brain Inj 6: 391–402.King N (1997) Mild head injury: Neuropathology, sequelae, measurement and recovery. Br J Clin

Psychol 36: 161–184.King NS (1996) Emotional, neuropsychological, and organic factors: their use in the prediction of

persistent postconcussion symptoms after moderate and mild head injury. J Neurol NeurosurgPsychiatry 61: 75–81.

King NS, Crawford S, Wenden FJ, Moss NEG & Wade DT (1995) The Rivermead Post ConcussionSymptoms Questionnaire: a measure of symptoms commonly experienced after head injury andits reliability. J Neurol 242: 587–592.

King NS, Crawford S, Wenden FJ Moss NEG, Wade DT & Caldwell FE (1997) Measurement ofpost-traumatic amnesia: how reliable is it? J Neurol Neurosurg Psychiatry 62: 38–42.

Klauber MR, Barrett-Connor E, Marshall LF & Bowers SA (1981) The epidemiology of head injury.A prospective study of an entire community–San Diego County, California, 1978. Am J Epidemi-ol 113: 500–509.

Kraus JF (1993) Epidemiology of head injury. In: Cooper PR (ed) Head Injury 1: 1–25. Williams &Wilkins, Baltimore.

Kraus JF, Black MA, Hessol N, Ley P, Rokaw W, Sullivan C, Bowers S, Knowlton S & Marshall L(1984) The incidence of acute brain injury and serious impairment in a defined population. Am JEpidemiol 119: 186–201.

Kraus JF, Morgenstern H, Fife D, Conroy C & Nourjah P (1989) Blood alcohol tests, prevalence ofinvolvement, and outcomes following brain injury. Am J Public Health 79: 294–299.

Kreutzer JS, Doherty KR, Harris JA & Zasler ND (1990) Alcohol use among persons with traumaticbrain injury. J Head Trauma Rehabil 5: 9–20.

Kreutzer JS, Wehman PH, Harris JA, Burns CT & Young HF (1991) Substance abuse and crime pat-terns among persons with traumatic brain injury referred for supported employment. Brain Inj 5:177–187.

de Kruijk JR, Leffers P, Menheere PPCA, Meerhoff S, Rutten J & Twijnstra A (2002a) Prediction ofpost-traumatic complaints after mild traumatic brain injury: early symptoms and biochemicalmarkers. J Neurol Neurosurg Psychiatry 73: 727–732.

de Kruijk JR, Twijnstra A, Meerhoff S & Leffers P (2001a) Management of mild traumatic brain in-jury: lack of consensus in Europe. Brain Inj 15: 117–123.

de Kruijk JR, Twijnstra A & Leffers P (2001b) Diagnostic criteria and differential diagnosis of mildtraumatic brain injury. Brain Inj 15: 99–106.

de Kruijk JR, Leffers P, Meerhoff S, Rutten J & Twijnstra A (2002b) Effectiveness of bed rest aftermild traumatic brain injury: a randomised trial of no versus six days of bed rest. J Neurol Neuro-surg Psychiatry 73: 167–172.

de Kruijk JR, Leffers P, Menheere PPCA, Meerhoff S & Twijnstra A (2001c) S-100B and neuron-specific enolase in serum of mild traumatic brain injury patients. A comparison with healthy con-trols. Acta Neurol Scand 103: 175–179.

Kurup V, McLatchie G & Jennett B (2000) Head injury. In: Mclatchie G, Harries M, Williams C &King J (eds) ABC of sports medicine 2: 7–12. BMJ Publishing Group, London.

Kyriacou DN, McCabe F, Anglinn D, Lapesarde K & Winer MR (1998) Emergency department-based study of risk factors for acute injury from domestic violence against women. Ann EmergMed 31: 502–506.

74

Lesch OM & Walter H (1996) New ‘state’ markers for the detection of alcoholism. Alcohol Alcohol31: 59–62.

Levi L, Guilburd JN, Lemberger A, Soustiel JF & Feinsod M (1990) Diffuse axonal injury: analysisof 100 patients with radiological signs. 27: 429–32.

Levin HS (1993) Neurobehavioural sequelae of closed head injury. In: Cooper PR (ed) Head Injury22: 525–551. Williams & Wilkins, Baltimore.

Levin HS, Mattis S, Ruff RM, Eisenberg HM, Marshall LF, Tabaddor K, High WM Jr & FrankowskiRF (1987) Neurobehavioral outcome following minor head injury: a three-center study. J Neuro-surg 66: 234–243.

Levin HS, Williams D, Crofford MJ, High WM Jr, Eisenberg HM, Amparo EH, Guinto FC Jr, Ka-lisky Z, Handerl SF & Goldman AM (1988) Relationship of depth of brain lesions to conscious-ness and outcome after closed head injury. J Neurosurg 69: 861–866.

Li G, Baker SP, Smialek JE & Soderstrom CA (2001) Use of alcohol as a risk factor for bicyclinginjury. JAMA 285: 893–896.

Lindsay KW & Bone I (2001) Head Injury. In: Lindsay KW & Bone I (ed) Neurology and Neurosur-gery Illustrated 214–233. Harcourt Publishers Ltd.

Liu S, Siegel Pz, Brewer RD, Mokdad AH, Sleet DA & Serdula M (1997) Prevalence of alcohol-im-paired driving. Results from a national self-reported survey of health behaviors. JAMA 227: 122–125.

Lieber CS (1995) Medical disorders of alcoholism. N Engl J Med 333: 1058–1065.Lishman WA (1988) Physiogenesis and psychogenesis in the ‘post-concussional syndrome’. Br J

Psychiatry 153: 460–469.Livingston DH, Lavery RF, Passannante MR, Skurnick JH, Baker S, Fabian TC, Fry DE & Malang-

oni MA (2000) Emergency department discharge of patients with a negative cranial computed to-mography scan after minimal head injury. Ann Surg 232: 126–132.

Lobato RD, Rivas JJ, Gomez PA, Castaneda M, Canizal JM, Sarabia R, Cabrera A & Munoz MJ(1991) Head-injured patients who talk and deteriorate into coma. Analysis of 211 cases studiedwith computerized tomography. J Neurosurg 75: 256–61.

Lobato RD, Sarabia R, Rivas JJ, Cordobes F, Castro S, Muñoz MJ, Cabrera A, Barcena A & LamasE (1986) Normal computerized tomography scans in severe head injury. Prognostic and clinicalmanagement implications. J Neurosurg 65: 784–789.

Lovinger DM (1993) Exitotoxity and alcohol-related brain damage. Alcohol Clin Exp Res 17: 19–27.Lundar T & Nestvold K (1985) Pediatric head injuries caused by traffic accidents. A prospective

study with 5-year follow-up. Child’s Nerv Syst 1: 24–28.Maas AIR, Dearden M, Teasdale GM, Braakman R, Cohadon F, Iannotti F, Karimi A, Lapierre F,

Murray G, Ohman J, Persson L, Servadei F, Stocchetti N & Unterberg A (1997) EBIC-Guidelinesfor management of severe head injury in adults. Acta Neurochir (Wien) 139: 286–294.

Marshall LF (2000) Epidemiology and cost of central nervous system injury. Clin Neurosurg 46:105–112.

Marshall LF, Gautille T, Klauber M, Eisenberg H, Jane JA, Luerssen TG, Marmarou A & Foulkes M(1991a) The outcome of severe closed head injury. J Neurosurg 75: 28–36.

Marshall LF, Marshall SB, Klauber MR, van Berkum Clark M, Eisenberg HM, Jane JA, LuerssenTG, Marmarou A & Foulkes MA (1991b) A new classification of head injury based on computer-ized tomography. J Neurosurg 75: 14–20.

Martin SE & Bachman R (1997) The relationship of alcohol to injury in assault cases. Recent DevAlcohol 13: 41–56.

Masse J, Billings B, Dhillon HS, Mace D, Hicks R, Barron S, Kraemer PJ, Dendle P & Prasad RM(2000) Three months of chronic ethanol administration and the behavioral outcome of rats afterlateral fluid percussion brain injury. J Neurotrauma 17: 421–430.

75

Mattson AJ & Levin HS (1990) Frontal lobe dysfunction following closed head injury. J Nerv MentDis 178: 282–291

Max W, MacKenzie EJ & Rice DP (1991) Head injury: costs and consequences. J Head Trauma Re-habil 6: 76–91.

Maxwell WL, Povlishock JT & Graham DL (1997) A Mechanistic analysis of nondisruptive axonalinjury: a review. J Neurotrauma 14: 419–440.

McGuire SA, Gronseth GS & Adams A (1995) Incidence of traumatic head injury in a defined pop-ulation. Neurology 45 (Suppl 4): A328.

McLeod R, Stockwell T, Stevens M & Phillips M (1999) The relationship between alcohol consump-tion patterns and injury. Addiction 94: 1719–1734.

Mearns J & Lees-Haley PR (1993) Discriminating neuropsychological sequelae of head injury fromalcohol-abuse-induced deficits: A review and analysis. J Clin Psychol 49: 714–720.

Meerkerk G-J, Njoo KH, Bongers MB, Trienekens P & van Oers JAM (1998) The specificity of theCDT assay in general practice: The influence of common chronic diseases and medication on theserum CDT concentration. Alcohol Clin Exp Res 22: 908–913.

Melnick DM, Maio RF, Blow FC & Hill EM (2000) Does acute alcohol ingestion and/or chronic al-cohol use among motor vehicle crash victims lead to a more comlicated hospital course? AnnuProc Assoc Adv Automot Med 44: 429–443.

Mendelow AD, Teasdale G, Jennett B, Bryden J, Hessett C & Murray G (1983) Risks of intracranialhaematoma in head injured adults. Br Med J 287: 1173–1176.

Menon DK (1999) Cerebral protection in severe brain injury: physiological determinants of outcomeand their optimisation. Br Med Bull 55: 226–258.

Mertol T, Guner M, Acar U, Atabay H & Kirisogly U (1991) Delayed traumatic intracerebral hemato-ma. Br J Neurosurg 5: 491–498.

Mickeviciene D, Schrader H, Nestvold K, Surkiene D, Kunickas R, Stovner LJ & Sand T (2002) Acontrolled historical cohort study on the post-concussion syndrome. Eur J Neurol 9: 581–587.

Mihas AA & Tavassoli M (1992) Laboratory markers of ethanol intake and abuse: a critical appraisal.Am J Med Sci 303: 415–428.

Miller H (1961) Accident neurosis. Br Med J 1: 919–925.Miller JD, Murray LS & Teasdale GM (1990) Development of a traumatic intracranial hematoma af-

ter a “minor” head injury. Neurosurgery 27: 669–673.Mittl RL Jr, Grossman RI, Hiehle JF Jr, Hurst RW, Kauder DR, Gennarelli TA & Alburger GW

(1994) Prevalence of MR evidence of diffuse axonal injury in patients with mild head injury andnormal head CT findings. Am J Neuroradiol 15: 1583–1589.

Moran SG, McCarthy MC, Uddin DE & Poeltstra RJ (1994) Predictors of positive CT scans in thetrauma patient with minor head injury. Am Surg 60: 533–535.

Moore BW (1965) A soluble protein characteristic of the nervous system. Biochem Biophys ResCommun 19: 739–744.

Murray GD, Teasdale GM, Braakman R, Cohadon F, Dearden M, Iannotti F, Karimi A, Lapierre F,Maas A, Ohman J, Persson L, Servadei F, Stocchetti N, Trojanowski T and Uterberg A (1999) TheEuropean brain injury consortium survey of head injuries. Acta Neurochir (Wien) 141: 223–236.

Mäkelä P, Fonager K, Hibell B, Nordlund S, Sabroe S & Simpura J (2001) Episodic heavy drinkingin four Nordic countries: a comparative Survey. Addiction 96: 1575–1588.

van der Naalt J, van Zomeren SH, Sluite WJ & Minderhoud JM (1999) One year outcome in mild tomoderate head injury: the predictive value of acute injury characteristics related to complaints andreturn to work. J Neurol Neurosurg Psychiatry 66: 207–213.

Nagy KK, Joseph KT, Krosner SM, Robers RR, Lelie CL, Dufty K, Smith RF & Barret J (1999) Theutility of head computed tomography after minimal head injury. J Trauma 46: 268–270.

Naimi TS, Brewer RD, Mokdad A, Denny C, Serdula MK & Marks JS (2003) Birge drinking amongUS adults. JAMA 298: 70–75.

76

NIH Consensus Development Panel on Rehabilitation of Persons With Traumatic Brain Injury. Con-sensus conference. Rehabilitatioin of persons with traumatic brain injury (1999) JAMA 282: 974–983.

Nilssen O, Ries RK, Rivara FP, Gurney JG & Jurkovich GJ (1994) The CAGE questionnaire and theShort Michigan Alcohol Screening Test in trauma patients: comparison of their correlations withbiological alcohol markers. J Trauma 36: 784–788.

Nordby HK & Urdal P (1985) Creatine kinase BB in blood as index of prognosis and effect of treat-ment after severe head injury. Acta Neurochir (Wien) 76: 131–136.

Nygaard Ø, Langbakk B & Romner B (1997) Age- and sex-related changes of S-100 protein concen-trations in cerebrospinal fluid and serum in patients with no previous history of neurological dis-order. Clin Chem 43: 541–543.

Olkkonen S & Honkanen R (1990) The role of alcohol in nonfatal bicycle injuries. Accid Anal Prev22: 89–96.

Olson JD, Kaufman HH, Moake J, O’Gorman TW, Hoots K, Wagner K, Brown CK & GildenbergPL (1989) The incidence and significance of hemostatic abnormalities in patients with head inju-ries. Neurosurgery 24: 825–832.

Ommaya AK & Gennarelli TA (1974) Cerebral concussion and traumatic unconsciousness. Correla-tion of experimental and clinical observations on blunt head injuries. Brain 97: 633–654.

Oppenheimer DR (1968) Microscopic lesions in the brain following head injury. J Neurol NeurosurgPsychiatry 31: 229–306.

Ponsford J, Willmott C, Rothwell A, Cameron P, Kelly AM, Nelms R & Curran C (2002) Impact ofearly intervention on outcome following mild head injury in adults. J Neurol Neurosurg Psychia-try 73: 330–332.

Porter RS (2000) Alcohol and injury in adolescents. Pediatr Emerg Care 16: 316–320.Povlishock JT, Becker DP, Cheng CLY & Vaughan GW (1983) Axonal change in minor head injury.

J Neuropathol Exp Neurol 42: 225–242.Päihdetilastollinen vuosikirja (2003) SVT Sosiaaliturva 2003: 3, Stakes.Raabe A, Grolms C, Keller M, Döhnert J, Sorge O & Seifert V (1998) Correlation of computed tom-

ography findings and serum brain damage markers following severe head injury. Acta Neurochir(Wien) 140: 787–792.

Raabe A, Grolms C & Seifert V (1999a) Serum markers of brain damage and outcome prediction inpatients after severe head injury. Br J Neurosurg 13: 56–59.

Raabe A, Grolms C, Sorge O, Zimmermann M & Seifert V (1999b) Serum S-100B protein in severehead injury. Neurosurgery 45: 477–483.

Reale F, Delfini R, Mencattini G (1984) Epidural hematomas. J Neurosurg Sci 28: 9–16.Relander M, Troupp H & Björkesten G (1972) Controlled trial of treatment for cerebral concussion.

Br Med J 4: 777–779. Rimel RW, Giordani B, Barth JT & Jane JA (1982) Moderate head injury: Completing the clinical

spectrum of brain trauma. Neurosurgery 11: 344–351.Rimel RW, Giordani B, Barth JT, Boll TJ & Jane JA (1981) Disability caused by minor head injury.

Neurosurgery 9: 221–228.Rice DP (1999) Economic costs of substance abuse, 1995. Proc Assoc Am Physicians 111: 119–125.Rice DP, Kelman S & Miller LS (1991) Estimated of economic costs of alcohol and drug abuse and

mental illness 1985 and 1988. Public Health Rep 106: 280–292.Rivara FP, Grossman DC & Cummings P (1997) Injury prevention. First of two parts. New Engl J

Med 337: 543–548.Rivara FP, Jurkovich GJ, Gurney JG, Seguin D, Fligner CL, Ries R, Raisys V & Copass M (1993a)

The magnitude of acute and chronic alcohol abuse in trauma patients. Arch Surg 128: 907–913.Rivara FP, Koepsell TD, Jurkovich GJ, Gurney JG & Soderberg R (1993b) The effects of alcohol

abuse on readmission for trauma. JAMA 270: 1962–1964.

77

Roberts I, Schierhout G & Alderson P (1998) Absence of evidence for effectiveness of five interven-tions routinely used in the intesive care management of severe head injury: a systematic review. JNeurol Neurosurg Psychiatry 65: 729–733.

Romner B, Ingebrigtsen T, Kongstad P & Borgesen SE (2000) Traumatic brain damage: serum S-100protein measurements related to neuroradiological findings. J Neurotrauma 17: 641–647.

Rosman AS (1992) Utility and evaluation of biochemical markers of alcohol consumption. J SubstAbuse 4: 277–297.

Ross HE, Gavin DR & Skinner HA (1990) Diagnostic validity of the MAST and the Alcohol Depend-ence Scale in the assessment of DSM-III alcohol disorders. J Stud Alcohol 51: 506–513.

Ruff RM & Grant I (1999) Postconcussional disorder: Background to DSM-IV and future consider-ations. In: Varney NR & Roberts RJ (eds) The evaluation and treatment of mild traumatic braininjury. Lawrence Erlbaum Associates, New Jersey, pp. 315–325.

Ruff RM, Marshall LF, Klauber MR, Blunt BA, Grant I, Foulkes MA, Eisenberg H, Jane J & Mar-marou A (1990) Alcohol abuse and neurological outcome of the severely head injured. J HeadTrauma Rehabil 5: 21–31.

Rutherford WH, Merrett JD, McDonald JR (1977) Sequelae of concussion caused by minor head in-juries. Lancet 1: 1–4.

Rutherford WH, Merrett JD, McDonald JR (1978) Symptoms at one year following concussion fromminor head injuries. Injury 10: 225–230.

Russell WR & Smith A (1961) Post-traumatic amnesia in closed head injury. Arch Neurol 5: 16–29.Ryb GE, Soderstrom CA, Kufera JA, Dischinger PC & Shiu M (1999) Use of blood alcohol concen-

tration and laboratory tests to detect current alcohol dependence in trauma center patients. J Trau-ma 47: 874–879.

Salvolini U (2002) Traumatic injuries: Imaging of facial injuries. Eur Radiol 12: 1253–1261.Saunders JB, Aasland OG, Babor TF, de la Fuente JR & Grant M (1993) Development of the Alcohol

Use Disorders Identification Test (AUDIT): WHO Collaborative Project on Early Detection ofPersons with Harmful Alcohol Consumption--II. Addiction 88: 791–804.

Segalowitz SJ & Lawson S (1995) Subtle symptoms associated with self-reported mild head injury28: 309–319.

Selzer ML (1971) The Michigan alcoholism screening test: the quest for a new diagnostic instrument.Am J Psychiatry 127: 1653–1658.

Selzer ML, Vinokur A & van Rooijen L (1975) A self-administered Short Michigan AlcoholismScreening Test (SMAST) J Stud Alcohol 36: 117–126.

Servadei F, Murray GD, Teasdale GM, Dearden M, Iannotti F, Lapierre F, Maas A Jr, Karimi A,Ohman J, Persson L, Stocchetti N, Trojanowski T & Unterberg A (2002) Traumatic subarachnoidhemorrhage: Demographic and clinical study of 750 patients from the European Brain Injury con-sortium Survey of head injuries. Neurosurgery 50: 261–269.

Servadei F (1997) Prognostic factors in severely head injured adult patients with epidural haemato-ma’s. Acta Neurochir (Wien) 139: 273–278.

Servadei F, Teasdale G & Merry G (2001) Defining acute mild head injury in adults: a proposal basedon prognostic factors, diagnosis, and management. J Neurotrauma 18: 657–664.

Shackford SR, Wald SL, Ross SE, Cogbill TH, Hoyt DB, Morris JA, Mucha PA, Pachter HL, Suger-man HJ, O’Malley K, Strutt PJ, Winchell RJ, Rutherford E, Rhodes M, Koslow M & DeMaria EJ(1992) The clinical utility of computed tomographic scanning and neurologic examination in themanagement of patients with minor head injuries. J Trauma 33: 385–394.

Shapiro AJ, Johnson RM, Miller SF & McCarthy MC (2001) Facial fractures in a level I trauma cen-tre: the importance of protective devices and alcohol abuse. Injury 32: 353–256.

Shapira Y, Lam AM, Paez A, Artru AA, Laohaprasit V & Donato T (1997) The influence of acuteand chronic alcohol treatment on brain edema, cerebral infarct volume and neurological outcomefollowing experimental head trauma in rats. J Neurosurg Anesthio l 9: 118–127.

78

Schermer CR, Gentilello LM, Hoyt DB, Moore EE, Moore JB, Rozycki GS & Peliciano DV (2003)National survey of trauma surgeons’ use of alcohol screening and brief intervention. J Trauma 55:849–856.

Shores EA, Marosszeky JE, Sandanam J & Batchelor J (1986) Preliminary validation of a clinicalscale for measuring the duration of post-traumatic amnesia. Med J Aust 144: 569–572.

Siddique MS, Gregson BA, Fernandes HM, Barnes J, Treadwell L, Wooldridge TD & Medelow AD(2002) Comparative study of traumatic and spontaneous intracerebral hemorrhage. J Neurosurg96: 86–89.

Signorini DF, Andrews PJD, Jones PA, Wardlaw JM & Miller JD (1999) Predicting survival usingsimple clinical variables: a case study in traumatic brain injury. J Neurol Neurosurg Psychiatry 66:20–25.

Siitonen S (2000) Effects of in-flight Gz acceleration on military aviators using modern anti-G gar-ments. Thesis. Univ Kuopio, Kuopio University Publications D. Medical Sciences 204.

Sillanaukee P (1996) Laboratory markers of alcohol abuse. Alcohol Alcohol 31: 6113–616.Sinson G, Bagley LJ, Cecil KM, Torchia M, McGowan JC, Lenkinski RE, McIntosh TK & Grossman

RI (2001) Magnetization transfer imaging and proton MR spectroscopy in the evaluation of axonalinjury: correlation with clinical outcome after traumatic brain injury. Am J Neuroradiol 22: 143–151.

Skinner HA & Allen BA (1982) Alcohol dependence syndrome: measurement and validation. J Ab-norm Psychol 91: 199–209.

Skinner HA, Holt S, Sheu WJ & Israel Y (1986) Clinical versus laboratory detection of alcohol abuse:the alcohol clinical index. Br Med J 292: 1703–1708.

Sobell LC & Sobell MB (1995) Alcohol consumption measures. Assessing alcohol problems: AGuide for Clinicians and Researchers. NIAAA Treatment Handbook Series 4: 55–73. NIH Publi-cation No. 95–3745. Government Printing Office, Washington, D.C.

Soderstrom CA, Cole FJ Jr & Porter JM (2001) Injury in America: the role of alcohol and otherdrugs–an EAST position paper prepared by the Injury Control and Violence Prevention Commit-tee. J Trauma 50: 1–12.

Soderstrom CA & Cowley RA (1987) A national alcohol and trauma center survey. Arch Surg 122:1067–1071.

Soderstrom CA, Smith GS, Kufera JA, Dischinger PC, Hebel JR, McDuff DR, Gorelick DA, Ho SM,Kerns TJ & Read KM (1997) The accuracy of the CAGE, the Brief Michigan Alsoholism Screen-ing Test, and the Alcohol Use Disorders Identification Test in screening trauma center patients foralcoholism. J Trauma 43: 962–969.

Soloniuk D, Pitts LH, Lovely M & Bartkowski H (1986) Traumatic intracerebral hematomas: timingof appearance and indications for operative removal. J Trauma 26: 787–794.

Sosin DM, Sacks JJ & Smith SM (1989) Head injury-associated deaths in the united states from 1979to 1986. JAMA 262: 2251–2255.

Sparadeo FR & Gill DG (1989) Effects of prior alcohol use on head injury recovery. J Head TraumaRehabil 4: 75–82.

Stein CS & Spettell C (1995) The head injury severity scale (HISS): a practical classification ofclosed-head injury. Brain Inj 9: 437–444.

Stibler H (1991) Carbohydrate-deficient transferrin in serum: a new marker of potentially harmful al-cohol consumption reviewed. Clin Chem 37: 2029–2037.

Stieg PE & Kase CS (1998) Intracranial hemorrhage: diagnosis and emergency management. NeurolClin 16: 373–390.

79

Stiell IG, Lesiuk H, Wells GA, Coyle D, McKnight RD, Brison R, Clement C, Eisenhauer MA,Greenberg GH, MacPhail I, Reardon M, Worthington J, Verbeek R, Rowe B, Cass D, Dreyer J,Holroyd B, Morrison L, Schull M & Laupacis A (2001a) Canadian CT head rule study for patientswith minor head injury: methodology for phase II (validation and economic analysis). Ann EmergMed 38: 317–322.

Stiell IG, Wells GA, Vandemheen K, Clement C, Lesiuk H, Laupacis A, McKnight RD, Verbeek R,Brison R, Cass D, Eisenhauer MA, Greenberg GH & Worthington J (2001b) The Canadian CThead rule for patients with minor head injury. Lancet 357: 1391–1396.

Strich SJ (1956) Diffuse degeneration of the cerebral white matter in severe dementia following headinjury. J Neurol Neurosurg Psychiatry 19: 163–185.

Strich SJ (1961) Shearing of nerve fibres as a cause of brain damage due to head injury. A patholog-ical study of twenty cases. Lancet 2: 443–448.

Stoodley M & Weir B (2000) Contents of chronic subdural hematoma. Neurosurg Clin N Am 11:425–434.

Stuss DT, Ely P, Hugenholtz H, Richard MT, LaRochelle S, Poirier CA & Bell I (1985) Subtle neu-ropsychological deficits in patients with good recovery after closed head injury. Neurosurgery 17:41–47.

Stuss DT, Stethem LL, Hugenholtz H, Picton T, Pivik J & Richard MT (1989) Reaction time afterhead injury: fatigue, divided and focused attention, and consistency of performance. J NeurolNeurosurg Psychiatry 52: 742–748.

Szymanski HV & Linn R (1992) A review of the postconcussion syndrome. Int J Psychiatry Med 22:357–375.

Tandon PN (2001) Acute subdural haematoma: A reappraisal. Neurol India 49: 3–10.Teasdale G & Jennett B (1974) Assessment of coma and impaired consciousness. A practical scale.

Lancet 2: 81–84.Teasdale GM, Nicoll JA, Murray G & Fiddes M (1997) Association of apolipoprotein E polymor-

phism with outcome after head injury. Lancet 350: 1069–1071.Thatcher RW, Cantor DS, McAlaster R, Geisler F & Krause P (1991) Comprehensive predictions of

outcome in closed head-injured patients. Ann NY Acad Sci 620: 82–101.The Brain Trauma Foundation. The American Association of Neurological Surgeons. The Joint Sec-

tion on Neurotrauma and Critical Care (2000a) Trauma Systems J Neurotrauma 17: 457–462.The Brain Trauma Foundation. The American Association of Neurological Surgeons. The Joint Sec-

tion on Neurotrauma and Critical Care (2000b) Role of antiseizure prophylaxis following head in-jury J Neurotrauma 17: 549–553.

Thomas DG, Palfreyman JW & Ratcliffe JG (1978) Serum-myelin-basic-protein assay in diagnosisand prognosis of patients with head injury. Lancet 21: 113–115.

Thompson K, Antony A & Holtzman A (2001) The costs of traumatic brain injury. N C Med J 62:376–379.

Thornhill S, Teasdale GM, Murray GD, McEwen J, Roy CW & Penny KI (2000) Disability in youngpeople and adults one year after head injury: prospective cohort study. Br Med J 320: 1631–1635.

Tiret L, Hausherr E, Thicoipe M, Garros B, Maurette P, Castel JP & Hatton F (1990) The epidemiol-ogy of head trauma in Aquitaine (France), 1986: A community-based study of hospital admissionsand deaths. Int J Epidemiol 19: 133–140.

Townend WJ, Guy MJ, Pani MA, Martin B & Yates DW (2002) Head injury outcome prediction inthe emergency department: a role for protein S-100B? J Neurol Neurosurg Psychiatry 73: 542–546.

Umile EM, Plotkin RC & Sandel ME (1998) Functional assessment of mild traumatic brain injuryusing SPECT and neuropsychological testing. Brain Inj 12: 577–594.

Usui A, Kato K, Abe T, Murase M, Tanaka M & Takeuchi E (1989) S-100a0 protein in blood andurine during open heart surgery. Clin Chem 35: 1942–1944.

80

Viola L, Zotta V, Martino V, Barbato R & Schisano G (2000) Minor head injuries: One year experi-ence according to the New Italian Guideline. Acta Neurochir (Wien) 142: 1281–1285.

Voller B, Benke T, Benedetto K, Schnider P, Auff E & Aichner F (1999) Neuropsychological, MRIand EEG findings after very mild traumatic brain injury. Brain Inj 13: 821–827.

Vos PE, Battistin L, Birbamer G, Gerstenbrand F, Potapov A, Prevec T, Stepan CA, Traubner P, Tw-ijnstra A, Vecsei L & von Wild K (2002) EFNS guideline on mild traumatic brain injury: reportof an EFNS task force. Eur J Neurol 9: 207–219.

Wade DT, Crawford S, Wenden FJ, King NS & Moss NEG (1997) Does routine follow up after headinjury help? A randomised controlled trial. J Neurol Neurosurg Psychiatry 62: 478–484.

Wagner AK, Hammond FM, Sasser HC, Wiercisiewski D & Norton HJ (2000) Use of injury severityvariables in determining disability and community integration after traumatic brain injury. J Trau-ma 49: 411–419.

Waller PF, Stewart JR, Hansen AR, Stutts JC, Lederhaus Popkin C & Rodgman EA (1986) The po-tentiating effects of alcohol on driver injury. JAMA 256: 1461–1466.

Walsh DC, Hingson RW, Merrigan DM, Levenson SM, Cupples LA, Heeren T, Coffman GA, BeckerCA, Barker TA, Hamilton SK, McGuire TG & Kelly CA (1991) A randomized trial of treatmentoptions for alcohol-abusing workers. N Engl J Med 325: 775–782.

Ward AA Jr (1966) The physiology of concussion. Clin Neurosurg 12: 95–111.Ward RE, Flynn TC, Miller PW & Blaisdell WF (1982) Effects of ethanol ingestion on the severity

and outcome of trauma. Am J Surg 144: 153–157.Wardlaw JM, Easton VJ & Statham P (2002) Which CT features help predict outcome after head in-

jury? J Neurol Neurosurg Psychiatry 72: 188–192.Wilberger JE Jr, Harris M & Diamond DL (1991) Acute subdural hematoma: morbidity, mortality,

and operative timing. J Neurosurg 74: 212–218.Williams DH, Levin HS & Eisenberg HM (1990) Mild head injury classification. Neurosurgery 27:

422–428.Wilson JTL, Pettigrew LEL & Teasdale GM (2000) Emotional and cognitive consequences of head

injury in relation to the Glasgow Outcome Scale. J Neurol Neurosurg Psychiatry 69: 204–209.Woertgen C, Rothoerl RD, Metz C & Brawanski A (1999) Comparison of clinical, radiologic, and

serum marker as prognostic factors after severe head injury. J Trauma 47: 1126–1130.Woertgen C, Rothoerl RD, Holzschuh M, Metz C & Brawanski A (1997) Comparison of serial S-100

and NSE serum measurements after severe head injury. Acta Neurochir (Wien) 139: 1161–1165.Wong PP, Dornan J, Schentag CT, Rosa IP & Keating AM (1993) Statistical profile of traumatic brain

injury: A Canadian rehabilitation population. 7: 283–294.Wrightson P & Gronwall D (1981) Time off work and symptoms after minor head injury. Injury 12:

445–454.Xiao-Sheng H, Sheng-Yu Y, Xiang Z, Zhou F & Jian-ning Z (2000) Diffuse axonal injury due to lat-

eral head rotation in a rat model. J Neurosurg 93: 626–33.Yanagawa T, Bunn F, Roberts I, Wentz R & Pierro A (2002) Nutritional support for head-injured pa-

tients. Cochrane Database Syst Rev 3: CD001530.Young A & Willatts S (1998) Controversies in management of acute brain trauma. Lancet 352: 164–

165.Ytrebø LM, Nedredal GI, Korvald C, Holm Nielsen O-J, Ingebrigtsen T, Romner B, Aarbakke J &

Revhaug A (2001) Renal elimination of protein S100ß in pigs with acute encephalopathy. ScandJ Clin Lab Invest 61: 217–226.

Zimmer DB, Cornwall EH, Landar A & Song W (1995) The S100 protein familiy: History, function,and expression. Brain Res Bull 37: 417–429.

Zimmer DB & van Eldik LJ (1987) Tissue distribution of rat S100α and S100β and S100-bindingproteins. Am J Physiol 252: 285–289.

81

Zimmerman RA, Bilaniuk LT, Gennarelli T, Bruce D, Dolinskas C & Uzzell B (1978) Cranial com-puted tomography in diagnosis and management of acute head trauma. Am J Roentgenol 131: 27–43.