Journal Neurotrauma

JOURNAL OF NEUROTRAUMA
Volume 23, Number 10, 2006

© The NeuroTrauma Foundation
Pp. 1468–1501
Guidelines for the Pharmacologic Treatment

of Neurobehavioral Sequelae of Traumatic Brain Injury
NEUROBEHAVIORAL GUIDELINES WORKING GROUP MEMBERS
DEBORAH L. WARDEN,1,6,7 BARRY GORDON,2,5,8 THOMAS W. McALLISTER,3,5,9

JONATHAN M. SILVER,1,5,10 JEFFERY T. BARTH,2,11 JOHN BRUNS,1,12
ANGELA DRAKE,3,13 TONY GENTRY,3,14 ANDY JAGODA,1,15 DOUGLAS I. KATZ,2,16
JESS KRAUS,4,17 LAWRENCE A. LABBATE,3,18 LAURIE M. RYAN,2,19
MOLLY B. SPARLING,3,20 BEVERLY WALTERS,4,21 JOHN WHYTE,2,22
ASHLEY ZAPATA,2,23 AND GEORGE ZITNAY1,24
ABSTRACT
There is currently a lack of evidence-based guidelines to guide the pharmacological treatment of neu-
robehavioral problems that commonly occur after traumatic brain injury (TBI). It was our objective
to review the current literature on the pharmacological treatment of neurobehavioral problems after
traumatic brain injury in three key areas: aggression, cognitive disorders, and affective disorders/anx-
iety/psychosis. Three panels of leading researchers in the field of brain injury were formed to review
the current literature on pharmacological treatment for TBI sequelae in the topic areas of affective/anx-
iety/psychotic disorders, cognitive disorders, and aggression. A comprehensive Medline literature search
Subgroups: 1Aggression, 2Cognition, 3Affective, 4Methodology Advisors, 5Workgroup Chairs, 6Guideline Group Coordinator.
7Defense and Veterans Brain Injury Center, Departments of Neurology and Neurosurgery, Walter Reed Army Medical Center,
and Departments of Neurology and Psychiatry, Uniformed Services University of the Health Sciences, Washington, DC. 8De-
partments of Neurology and Cognitive Science, The Johns Hopkins University, Baltimore, Maryland. 9Departments of Psychia-
try and Neuropsychiatry, Dartmouth Medical School, Lebanon, New Hampshire. 10Department of Psychiatry, New York Uni-
versity School of Medicine, New York, New York. 11Departments of Psychiatric Medicine and Neurological Surgery, University
of Virginia School of Medicine, Charlottesville, Virginia. 12Department of Emergency Medicine, Mount Sinai School of Medi-
cine, New York, New York. 13Defense and Veterans Brain Injury Center, San Diego Naval Medical Center, San Diego, Cali-
fornia. 14Partnership for People with Disabilities, Department of Education, Virginia Commonwealth University, Richmond, Vir-
ginia. 15Department of Emergency Medicine, Mount Sinai School of Medicine, New York, New York. 16Department of Neurology,
Boston University School of Medicine, Boston Massachusetts, and Brain Injury Programs, Healthsouth Braintree Rehabilitation
Hospital, Braintree, Massachusetts. 17Department of Epidemiology, UCLA School of Public Health, Los Angeles, California.
18Department of Psychiatry, University of Arkansas for Medical Sciences and VA Medical Center, North Little Rock, Little Rock,
Arkansas. 19Defense and Veterans Brain Injury Center, Department of Neurology, Walter Reed Army Medical Center and De-
partment of Neurology, Uniformed Services University of the Health Sciences, Washington, DC. 20Defense and Veterans Brain
Injury Center, Department of Neurology, Walter Reed Army Medical Center, Washington, DC. 21Department of Neurosurgery,
New York University School of Medicine, New York, New York. 22Moss Rehabilitation Research Institute, Albert Einstein
Healthcare Network and Department of Rehabilitation Medicine, Thomas Jefferson University, Philadelphia, Pennsylvania. 23Med-
ical College of Virginia, Virginia Commonwealth University, Richmond, Virginia. 24Laurel Highlands Neuro-Rehabilitation and
Defense and Veterans Brain Injury Center, Walter Reed Army Medical Center, Washington, DC.
Contributors: Sureyya Dikmen, Hunter Downs, David Thurman, and Peter Quinn.
1468
GUIDELINES FOR PHARMACOLOGIC TREATMENT OF NEUROBEHAVIORAL SEQUELAE
was performed by each group to establish the groups of pertinent articles. Additional articles were ob-
tained from bibliography searches of the primary articles. Group members then independently reviewed
the articles and established a consensus rating. Despite reviewing a significant number of studies on
drug treatment of neurobehavioral sequelae after TBI, the quality of evidence did not support any treat-
ment standards and few guidelines due to a number of recurrent methodological problems. Guidelines
were established for the use of methylphenidate in the treatment of deficits in attention and speed of
information processing, as well as for the use of beta-blockers for the treatment of aggression follow-
ing TBI. Options were recommended in the treatment of depression, bipolar disorder/mania, psychosis,
aggression, general cognitive functions, and deficits in attention, speed of processing, and memory af-
ter TBI. The evidence-based guidelines and options established by this working group may help to guide
the pharmacological treatment of the person experiencing neurobehavioral sequelae following TBI.
There is a clear need for well-designed randomized controlled trials in the treatment of these common
problems after TBI in order to establish definitive treatment standards for this patient population.
Key words: aggression; brain injury; cognition; evidence based reviews; neurobehavior; pharmacology;
psychiatric disorders
INTRODUCTION with TBI are more refractory to treatment, more prone to

the various side effects of psychotropic agents, and may
not respond to traditional psychotropic agents.9 In fact
T RAUMATIC BRAIN INJURY (TBI) is a principal cause of
death and disability in active young adults today. Ap-
proximately 5 million Americans live with disability as
some clinicians feel very strongly that some standard
agents are ineffective, or that they may be associated with
a result of brain injury, with an estimated cost to society excess toxicity in the brain injured population.
of $48.3 billion annually.1,2 Depression after TBI is very common, with best esti-
Neurobehavioral problems commonly seen after TBI in- mates suggesting that 25–60% of individuals with TBI
clude psychiatric disorders, cognitive problems and ag- develop a depressive episode within eight years of their
gression. It is well recognized that neurobehavioral prob- injury.6,10,11 Depression is associated with poorer social
lems are the most debilitating sequelae to individuals with and functional outcome.11 Anxiety occurs frequently with
TBI attempting to reestablish family and work relation- depression,4,6,11,12 and may occur as an independent dis-
ships.3 These sequelae may also impede the rehabilitation order. Van Reekum and colleagues have suggested that
and recovery process, contribute to sub-optimal follow-up, individuals with TBI have increased relative risks of gen-
and ultimately negatively affect overall outcomes. Such eralized anxiety disorder, obsessive compulsive disorder
symptoms may interfere with return to work and quality (OCD), panic disorder, and post-traumatic stress disor-
of life and the cost to the nation in care expenditures, lost der (PTSD) of 2.3, 2.6, 5.8, and 1.8, respectively. PTSD
productivity and family disruption is staggering.4 and depression are common disorders in those who have
While many small clinical trials and case studies have persistent postconcussive symptoms.13,14
been conducted of the effectiveness of pharmacologic Although psychosis is a relatively rare complication of
therapy for these problems after TBI, a comprehensive TBI, psychotic syndromes do occur more frequently in
review to form evidence based practice guidelines has individuals who have had a TBI than in the general pop-
not been conducted. ulation.8,15–17 Psychosis is a good example of a low fre-
quency, high impact complication of TBI, causing enor-
mous distress to individuals and their caregivers.15
I. Affective Disorders, Anxiety, and Psychosis Psychotic syndromes following TBI can occur during the
after TBI period of post-traumatic amnesia, as a complication of
A significant body of evidence suggests that TBI re- post-traumatic epilepsy, in the context of TBI-related
sults in an increased relative risk of developing various mood disorders, or associated with a chronic, schizo-
psychiatric disorders including mood and anxiety disor- phrenia-like syndrome. Bearing these contexts in mind
ders, as well as psychotic syndromes.5–8 Despite this, when assessing the etiology of psychosis helps to guide
there is little consensus about the treatment of these dis- appropriate interventions.
orders. In the absence of good evidence, a body of clin- Thus, considering that depression, anxiety, and psy-
ical lore has developed which suggests that individuals chotic disorders occur at increased rates in individuals
1469
WARDEN ET AL.
with TBI, and that 1–2 million Americans sustain a TBI Attention and speed of processing. The diffuse dam-
each year, it is important to clarify the state of evidence age that accompanies traumatic brain injury tends to di-
with respect to the pharmacologic treatments of these dis- minish attention/concentration, mental processing speed,
orders in the TBI population. and cognitive efficiency. Related to these decreased abil-
ities, patients’ complaints include inability to concentrate,
II. Cognitive Deficits after TBI distractibility, difficulty performing more than one task
at a time, confusion and perplexity in thinking, irritabil-
The cognitive effects of TBI become evident after res- ity, fatigue, and increased time and effort to perform even
olution of posttraumatic amnesia and have been well doc- simple tasks.18,23,24
umented.18,19 Although cognitive deficits vary with
severity of injury, these generally include problems with Memory. Problems with memory are among the most
attention/concentration, memory, and executive func- common cognitive complaints in persons with TBI.
tioning, e.g., problem solving, mental flexibility, and ini- Memory can be fractionated into a variety of processes
tiation. These deficits are found within the overall pic- and subdomains that fit along a number of overlapping
ture of diffuse brain injury and, in particular, damage to dimensions. The studies reviewed for this report that
frontotemporal regions commonly seen in TBI. Cogni- cover memory functions largely consider declarative
tive complaints are common even after mild TBI both memory processes (i.e. recall of facts and events) and as-
acutely and at later follow-up. Prevalence rates for mem- sociated subdomains. Memory processes are highly re-
ory and attention complaints after mild TBI vary but have lated to attention and executive functioning, and prob-
been reported to range from 40 to 60% at 1–3 months lems in these domains may affect the efficiency of
post-injury.20 Deficits can result in significant long-term encoding and retrieval.
morbidity even in those with more mild injuries. The de-
velopment of effective treatment strategies for these cog- Executive functions. Prefrontal brain systems are vul-
nitive sequelae is critical. Medications are frequently nerable to diffuse and focal damage after TBI; conse-
used to treat neurobehavioral sequelae.21,22 However, ev- quently, problems in executive capacities such as reason-
idence-based clinical guidance has been lacking. ing, planning, inhibiting, organizing and sequencing are
In considering the possible effects of pharmacologic common in persons with TBI. Again, a number of subdo-
treatment on the cognitive deficits that can follow TBI, mains within this category of cognition are interdependent
at least three complexities have to be kept in mind. One with other cognitive domains and emotional functioning.
is that cognition is not a single ability; it is now known
to be made up of components that are separable on func- III. Aggression after TBI
tional, neural, and pathological grounds. Another com- Explosive and violent behavior has long been associ-
plexity is that the components of cognition cannot be ated with focal brain lesions, as well as with diffuse dam-
measured in pure form by any available test, so that the age to the central nervous system (CNS).25 Agitation that
translation from actual test results to componential func- occurs during the acute stages of recovery from brain in-
tion is rarely exact. The third complexity is that TBI need jury can endanger the safety of patients and their care-
not respect the componential boundaries and may cause givers. Agitation may be predictive of longer length of
functional deficits that are greater or less than the sum of stay and decreased cognition.26 Subsequently, low frus-
the component deficits. tration tolerance and explosive behavior may develop that
can be set off by minimal provocation or occur without
Cognitive components (domains). The usual division warning. Aggression and irritability are major causes of
of cognitive domains separates them into the following disability to individuals with brain injury and sources of
components: attention, executive function, memory, lan- stress to their families. These episodes range in severity
guage, visuospatial and constructional abilities, and sen- from irritability to outbursts that result in damage to prop-
sory-perceptual-motor skills. Although the traditional de- erty or assaults on others. Aggressive and agitated be-
finitions of each domain is well recognized, rapidly haviors may be treated in a variety of settings, ranging
expanding research has elaborated further complexity from the acute brain injury unit in a general hospital, to
with respect to domain subcomponents and the interrela- a “neurobehavioral” unit in a rehabilitation facility, to
tionship of the neural networks that subserve these do- outpatient environments including the home setting.
mains. For the purpose of our review of outcomes in TBI However, in severe cases, affected individuals cannot re-
pharmacologic treatment research, we have defined these main in the community or with their families, and require
domains using the following framework. care in long-term psychiatric or neurobehavioral facili-
1470
ties. In a survey of all skilled nursing facilities in Con- 9. Arciniegas DB, Topkoff J, Silver JM. Neuropsychiatric as-
necticut, 45% of facilities had individuals with a primary pects of traumatic brain injury. Current Treatment Options
diagnosis of TBI who met the definition of agitation.27 in Neurology. 2000;2:169–186.
It has been reported that during the acute recovery pe- 10. Kreutzer JS, Seel RT, Gourley E. The prevalence and
riod, 35–96% of individuals with brain injury exhibit ag- symptom rates of depression after traumatic brain injury:
itated behavior.28,29 After the acute recovery phase, irri- a comprehensive examination. Brain Injury. 2001;15(7):
tability or bad temper is common, particularly following 563–576.
moderate to severe injury. In the two prospective studies 11. Jorge R, Robinson RG. Mood disorders following trau-
of the occurrence of aggression, agitation, or restlessness matic brain injury. NeuroRehabilitation. 2002;17(4):311–
that have been monitored by an objective rating instru- 324.
ment, the Overt Aggression Scale, 11–34% of TBI patients 12. Hiott DW, Labbate L. Anxiety disorders associated with
were found to be agitated or have aggressive behavior.30,31 traumatic brain injuries. NeuroRehabilitation. 2002;17(4):
In studies that have followed patients from 1 to 15 years 345–355.
after injury, irritability has occurred in up to 71%, and ag- 13. Warden DL, Labbate LA, Salazar AM, et al. Posttraumatic
itation in up to 67%.32–42 In one study, increased irritabil- stress disorder in patients with traumatic brain injury and
ity has also been linked to the number of traumatic brain amnesia for the event? Journal of Neuropsychiatry and
injuries and the presence of loss of consciousness.43 Clinical Neurosciences. 1997;9(1):18–22.
Although there is no medication that is approved by 14. McAllister TW, Arciniegas D. Evaluation and treatment of
the FDA specifically for the treatment of aggression, postconcussive symptoms. NeuroRehabilitation. 2002;
medications are widely used in the management of pa- 17(4):265–283.
tients with acute or chronic aggression. The reported ef-
15. McAllister TW. Traumatic Brain Injury and Psychosis:
fectiveness of these medications is highly variable, as are What Is the Connection? Semin Clin Neuropsychiatry. Jul
the reported rationales for their prescription. 1998;3(3):211–223.
16. Malaspina D, Goetz RR, Friedman JH, et al. Traumatic
brain injury and schizophrenia in members of schizophre-
REFERENCES nia and bipolar disorder pedigrees. Am J Psychiatry. Mar
2001;158(3):440–446.
1. Lewin ICF. The Cost of Disorders of the Brain. The Na-
tional Foundation for the Brain. Washington, D.C.; 1992. 17. Davison K, Bagley CR. Schizophrenia-like psychosis as-
sociated with organic disorders of the central nervous sys-
2. Centers for Disease Control and Prevention. Traumatic
tem. British Journal of Psychiatry. 1969;114:113–184.
Brain Injury in the United States: A Report to Congress.
www] http://www.cdc.gov/ncipc/pub-res/tbicongress.htm. 18. Lezak MD. Neuropsychological assessment. 3rd ed. New
Accessed January 16, 2001, 2001. York: Oxford University Press; 1995.
3. Prigatano GL. Personality disturbances associated with 19. McAllister TW. Neuropsychiatric sequelae of head injuries.
traumatic brain injury. Journal of Consulting and Clinical Psychiatric Clinics of North America. 1992;15(2):395–413.
Psychology. 1992;60:360–368. 20. McAllister TW. Mild traumatic brain injury and the post-
4. Fann JR, Katon WJ, Uomoto JM, et al. Psychiatric disor- concussive syndrome. In: Silver J, Yudofsky S, Hales R,
ders and functional disability in outpatients with traumatic eds. Neuropsychiatry of Traumatic Brain Injury. Wash-
brain injuries. American Journal of Psychiatry. 1995;152(10): ington, DC: American Psychiatruc Press; 1994:357–392.
1493–1499. 21. Evans RW, Evans RI, Sharp MJ. The physician survey of
5. Deb S, Lyons I, Koutzoukis C. Neuropsychiatric sequelae the post-concussion and whiplash syndromes. Headache.
one year after a minor head injury. J Neurol Neurosurg 1994;34:268–274.
Psychiatry. Dec 1998;65(6):899–902. 22. Mittenberg W, Canyock EM, Condit C, et al. Treatment of
6. Hibbard MR, Uysal S, Kepler K, et al. Axis I psy- post-concussion syndrome following mild head injury. J
chopathology in individuals with traumatic brain injury. J Clin Exp Neuropsychol. Dec 2001;23(6):829–836.
Head Trauma Rehabil. Aug 1998;13(4):24–39. 23. Gentilini M, Nichelli P, Schoenhuber R. Assessment of at-
7. van Reekum R, Cohen T, Wong J. Can traumatic brain in- tention in mild head injury. In: Levin H, Eisenberg H, Ben-
jury cause psychiatric disorders? J Neuropsychiatry Clin ton A, eds. Mild Head Injury. New York: Oxford Univer-
Neurosci. Summer 2000;12(3):316–327. sity Press; 1989:163–175.
8. Koponen S, Taiminen T, Portin R, et al. Axis I and II psy- 24. Raskin SA, Mateer CA. Neuropsychological management
chiatric disorders after traumatic brain injury: a 30-year fol- of mild traumatic brain injury. New York: Oxford Uni-
low-up study. Am J Psychiatry. Aug 2002;159(8):1315–1321. versity Press; 2000.
1471
WARDEN ET AL.
25. Elliott FA. Violence. The neurologic contribution: an 41. Rutherford WH. Sequelae of concussion caused by minor
overview. Arch Neurol. Jun 1992;49(6):595–603. head injuries. Lancet. Jan 1 1977;1(8001):1–4.
26. Bogner JA, Corrigan JD, Fugate L, et al. Role of agitation 42. Levin HS, Grossman RG, Rose JE, et al. Long-term neu-
in prediction of outcomes after traumatic brain injury. Am ropsychological outcome of closed head injury. J Neuro-
J Phys Med Rehabil. Sep 2001;80(9):636–644. surg. Apr 1979;50(4):412–422.
27. Wolf AP, Gleckman AD, Cifu DX, et al. The prevalence 43. Carlsson GS, Svardsudd K, Welin L. Long-term effects of
of agitation and brain injury in skilled nursing facilities: a head injuries sustained during life in three male popula-
survey. Brain Injury. 1996;10(4):241–245. tions. J Neurosurg. Aug 1987;67(2):197–205.
28. Levin HS, Grossman RG. Behavioral sequelae of closed
head injury. A quantitative study. Arch Neurol. Nov 1978;
35(11):720–727. METHODS
29. Rao N, Jellinek HM, Woolston DC. Agitation in closed
General Methodology and Search Strategy
head injury: haloperidol effects on rehabilitation outcome.
Archives of Physical Medicine and Rehabilitation. The guidelines group accomplished the evidence based
1985;66:30–34. review of the literature over several meeting sessions.
30. Brooke MM, Patterson DR, Questad KA, et al. The treat- Subgroups based on topic area were identified during the
ment of agitation during initial hospitalization after trau- first meeting held in January 2000. Search parameters for
matic brain injury. Archives of Physical Medicine and Re- each topic were also identified at that time. The primary
habilitation. 1992;73:917–921. comprehensive literature search was conducted at the
31. Tateno A, Jorge RE, Robinson RG. Clinical correlates of meeting held in July 2000. The search engine OVID was
aggressive behavior after traumatic brain injury. J Neu- used to explore the Medline database from 1960 to the
ropsychiatry Clin Neurosci. Spring 2003;15(2):155–160. present. In brief, the overall approach used was to define
32. Rao N, Jellinek HM, Woolston DC. Agitation in closed the proper terms for brain injury, the behavioral syn-
head injury: haloperidol effects on rehabilitation outcome. dromes of interest, and the therapies of interest. The term
Archives of Physical Medicine and Rehabilitation. 1985; Craniocerebral Trauma was the broadest appropriate
66:30–34. term in the OVID program for generating articles rele-
33. McKinlay WW, Brooks DN, Bond MR, et al. The short- vant to traumatic brain injury. This includes such areas
term outcome of severe blunt head injury as reported by as brain injuries, coma, post-head injury, head injuries
relatives of the injured persons. J Neurol Neurosurg Psy- (closed and penetrating), and intracranial hemorrhage
chiatry. Jun 1981;44(6):527–533. (traumatic). Each of these terms in turn has several sub-
34. Brooks N, Campsie L, Symington C, et al. The five year headings. OVID allows these terms to be “exploded”
outcome of severe blunt head injury: a relative’s view. J which means that the search will include all of the sub-
Neurol Neurosurg Psychiatry. Jul 1986;49(7):764–770. headings under the main heading. For example, an “ex-
35. Oddy M, Coughlan T, Tyerman A, et al. Social adjustment ploded” search of Craniocerebral Trauma resulted in
after closed head injury: a further follow-up seven years 59,780 references. The addition of key words “brain in-
after injury. J Neurol Neurosurg Psychiatry. Jun 1985;48(6): jury, head injury, head injuries” resulted in a total of
564–568. 63,702 references. The terms used to generate the brain
36. Thomsen IV. Late outcome of very severe blunt head injury and behavioral syndromes were then combined and
trauma: a 10–15 year second follow-up. J Neurol Neuro- filtered to include humans (as opposed to animals) and
surg Psychiatry. Mar 1984;47(3):260–268. English language articles.
37. van Zomeren AH, van den Burg W. Residual complaints These terms, along with the syndromes of interest,
of patients two years after severe head injury. J Neurol Neu- therapies of interest, and other key words were combined
rosurg Psychiatry. Jan 1985;48(1):21–28. to generate a bibliography of articles to be further eval-
38. McMillan TM, Glucksman EE. The neuropsychology of uated. Subject review articles were included solely for a
moderate head injury. J Neurol Neurosurg Psychiatry. Apr review of bibliographies in order to capture any relevant
1987;50(4):393–397. articles that may have been missed by the computer re-
39. Schoenhuber R, Gentilini M. Anxiety and depression after view. References from research articles were also re-
mild head injury: a case control study. Journal of Neurol- viewed for inclusion. Similarly, all members of the guide-
ogy, Neurosurgery, and Psychiatry. 1988;51:722–724. lines group were asked to review their personal files for
40. Dikmen S, McLean A, Temkin N. Neuropsychological and any pertinent articles that may not have been identified
psychosocial consequences of minor head injury. Journal by the Medline search. Titles of research articles were re-
of Neurology, Neurosurgery & Psychiatry. 1986;49(11): viewed initially for relevance to the topic. When a deci-
1227–1232. sion could not be made, abstracts were obtained. Based
1472
on review of the article titles and abstracts, two group data. Types of studies so classified include: observational
members reduced the list of papers to be read by the studies, cohort studies, prevalence studies, and case con-
working group. If the title or abstract appeared to be rel- trol studies. As mentioned above, class I design studies
evant to the topic, the full text article was reviewed. may be downgraded to class II evidence based on
This review confined itself to research done on individ- methodological flaws.
uals with TBI and a review of other therapies commonly
used for TBI sequelae but with published literature only in Class III evidence. Most studies are based on retro-
other patient populations was beyond the scope of the pro- spectively collected data. Evidence used in this class in-
ject. However, readers should not lose sight of the fact that dicates well-designed and conducted clinical series, data-
evidence of efficacy in other patient populations is also a bases or registries, case reviews, and case reports. Class
useful source of treatment options in TBI in many cases. I and II design studies may be downgraded to class III
evidence due to methodological flaws.
Article Scoring As mentioned above, although a class I study by de-
Papers were read and scored by at least two of the group sign, a RCT could be downgraded by raters to class II or
members independently. Both committee members com- class III evidence or even be judged unusable based upon
pleted data extraction forms. Articles were scored based on the degree of methodological flaws. Articles were also
methodology with consideration to limitations/potential rated unusable if no measurement data were provided,
confounds that would affect the interpretation of results. A patients had acquired rather than traumatic brain injury
consensus review was then completed for each paper, and or if a mixed population was reported for which separate
the level of evidence decided upon by the two primary re- data for patients with TBI were not given. In an effort to
viewers. When the two primary reviewers could not reach minimize confusion, throughout this document classes I,
agreement, the other group members were consulted to II and III refer to classes of evidence (the final decision
achieve a consensus rating. Evidence tables were generated of the group based upon study design and downgrades
using the scoring results on this final group of papers. for methodological flaws) rather than study design, ex-
The scoring method was adapted from the Brain cept where specifically noted in the text.
Trauma Foundation’s Guidelines for the Management of Recommendations for standards, guidelines and op-
Severe Head Injury1: tions were based upon the level of evidence described in
the Table 1 (adapted from Cicerone et al., 20002).
Class I evidence. This is well-designed and conducted
I. Affective Disorders, Anxiety, and Psychosis
prospective randomized controlled trials (RCT)—the
gold standard of clinical trials. However, some may be The Affective Disorders, Anxiety, and Psychosis work
downgraded due to poor design, insufficient patient num- group initially determined their focus to be a review of
bers, or other methodological inadequacies. evidence for the pharmacologic treatment of depression
and anxiety. After further discussion and an initial review
Class II evidence. This is well-designed and conducted of available literature from preliminary searches, it was
clinical studies in which the data were collected prospec- felt that the review should be expanded to include evi-
tively, or retrospective analyses based on clearly reliable dence for the treatment of other mood disorders includ-
TABLE 1. CRITERIA FOR DETERMINING THE LEVEL OF RECOMMENDATION
Standards Guidelines Options
Based on at least 1, well-designed Based on well designed class II Based on class II or class III studies,
class I study with an adequate studies with adequate samples, that with additional grounds to support
sample, or overwhelming class II directly addresses the effectiveness a recommendation as to whether
evidence, that directly addresses of the treatment in question, the treatment be specifically
the effectiveness of the treatment providing fair evidence to support considered for persons with
in question, providing good a recommendation as to whether traumatic brain injury.
evidence to support a the treatment be specifically
recommendation as to whether the considered for persons with
treatment be specifically traumatic brain injury.
considered for persons with
traumatic brain injury.
1473
WARDEN ET AL.
ing mania, as well as the treatment of psychotic syn- mately done using a search string to identify closed head
dromes. Other behavioral syndromes (such as patholog- injury, in conjunction with a variety of cognitive disor-
ical affect or affective lability, PTSD, substance abuse) ders. The search for pharmacologic treatments was done
were not included in this effort. The terms anxiety dis- manually, using the titles of the retrieved articles. The
orders, mood disorders, mania, bipolar disorder, schizo- exact search string that was used for MEDLINE and
phrenia and disorders with psychotic features were used PsychLit through January 2002, was the following:
by the Affective Disorders/Anxiety/Psychosis working
group were used in the Medline search to describe the (Craniocerebral trauma [mh] OR brain injuries [mh]
psychiatric disorders of interest. The terms antidepres- OR head injuries, closed [mh] OR brain damage,
sive agents, antipsychotic agents, anti-anxiety agents, an- chronic [mh]) AND (Delirium, Dementia, Amnes-
ticonvulsants, electroconvulsive therapy, electromagnet- tic, Cognitive Disorders [mh] OR cognition [mh] OR
ics, phototherapy, drug therapy were used to encompass cognitive symptoms [mh] OR memory [mh] OR
the therapies of interest. In an effort to be sure that pa- memory disorders [mh] OR attention [mh] OR
pers exploring the use of newer atypical antipsychotics arousal [mh] OR neuropsychological tests [mh] OR
and other new biological interventions were identified, activities of daily living [mh] OR affective symp-
the key words clozapine, olanzapine, quetiapine, and toms [mh] OR laughter [mh] OR akathisia, drug-in-
transcranial magnetic stimulation were added. The work- duced [mh] OR motor activity [mh] OR dysarthria
ing group reviewed a total of 150 papers including those [mh] OR akinetic mutism [mh] OR delirium [mh]
in the initial medline search and those identified by group OR hallucinations [mh] OR neurologic manifesta-
members and from bibliography searches. The majority tions [mh] OR behavioral symptoms [mh])
of these articles were review papers or not within the
topic area and thus were subsequently excluded after bib- This comprehensive process identified 61 papers that
liography review was conducted. However, 43 remain- were reviewed and entered into the evidence tables for
ing articles were considered for evidence. the Cognitive Disorders working group.
II. Cognitive Disorders III. Aggression

A preliminary list of search terms was generated by the At the initial meeting of the Aggression working group,
subgroup members through a discussion on cognitive im- the primary focus question for the group was determined
pairments secondary to traumatic brain injury. Group to be “What is the evidence to direct pharmacologic man-
members were able to generate this list from their clini- agement of aggressive disorders following traumatic
cal expertise and knowledge within their respective dis- brain injury?” The group decided to defer an investiga-
ciplines. From this discussion, a list of 188 clinical terms tion of treatment for agitation of individuals in an acute
to describe these impairments was generated, in addition confusional state and to focus the current investigation
to a list of pharmacological agents. These clinical terms on those individuals with TBI seen in the post-acute
were submitted for the initial MESH term search. Two stage, such as in a rehabilitation unit or an outpatient set-
subject heading specialists from the Medical Subject ting. Members discussed topic areas and identified search
Heading Department, Jacqueline Shoalman and Doug terms for the literature search for articles on aggression.
Johnston, were consulted via telephone. Many of the terms Literature from institutionalized populations (e.g., prison
selected did not have exact matches, and the initial search populations) was not reviewed as it was deemed difficult
yielded thousands of additional possible families of search to generalize to the larger population of patients with
terms. This new list of terms to consider was circulated brain injury. The methodology chosen was based on the
to all members of the subgroup, who systematically elim- Institute of Medicine (IOM) Committee to Advise the
inated various terms that were considered irrelevant for Public Health Service on Clinical Practice Guidelines
the purposes of the project. The search process also re- (1990). The search string used in the Medline search en-
vealed that pharmacologic treatments were generally not gine included the following:
indexed hierarchically enough to allow an appropriate
search using index terms. In addition, there was some con- (craniocerebral trauma OR head injury OR brain
cern that “cognitive disorders” of the kind thought to oc- injury) AND (aggression OR irritability OR vio-
cur after traumatic brain injury had also been given vari- lence) AND (drug therapy OR beta blockers OR
ous category labels. After additional consultation with anticonvulsants OR propranolol OR valproate OR
Robert Mormon, Director of the Medical Library at Wal- lithium OR benzodiazepines OR neuroleptics OR
ter Reed Army Medical Center, the final search was ulti- antidepressants)
1474
Later searches were completed for penetrating brain in- be effective for the treatment of depression after TBI. How-
jury, as well as individual anticonvulsants. Titles and ab- ever, side effects may limit their utility in this population.
stracts were reviewed for appropriateness. Articles clearly Two reports indicate that TCAs may be less effective in
not in the topic area, or reporting preclinical work, were patients with TBI than in non-brain injured populations.
excluded. The aggression subgroup identified a total of 127 Serotonin Reuptake Inhibitors:
articles. The majority of these articles were review papers The use of sertraline (25–200 mg/day) is recom-
or not within the topic area and thus were subsequently mended as an option in the treatment of depression after
excluded after bibliography review was conducted. How- TBI based upon an 87% response rate in one class III
ever, 52 remaining articles were considered for evidence. study (n 16).
Bipolar Disorder/Mania: There is insufficient evi-
Search Update dence to support or refute the use of commonly used med-
Due to the lengthy review process undertaken in this ications including lithium, anticonvulsants, and antipsy-
evidence based report, the initial search will be more than chotics in the management of TBI-related bipolar
three years old at the time of publication. Although a disorder or mania.
comprehensive search could not be conducted again for Anxiety: There is insufficient evidence to support or
the intervening time period, a focused literature review refute the use of commonly used medications including
was performed in October 2004, to identify any new ran- TCAs, benzodiazepines, and SRIs in the treatment of anx-
domized controlled trials. The search term “Brain In- iety after TBI.
juries” was limited to include only randomized controlled Psychosis:
trials on Medline and the resulting references and ab- Atypical Antipsychotics:
stracts were reviewed to determine any relevant publica- The use of olanzapine (5–20 mg/day) is recom-
tions. Two recent randomized controlled trials were re- mended at the level of an option for the treatment of psy-
viewed by group members using the same methodology chotic symptoms after TBI based on two case reports.
described above and have been included in this report. Others: There is insufficient evidence to support
or refute the use of other commonly used medications for
psychosis. One of two available reports suggests that the
REFERENCES use of clozapine was associated with significant sedation,
weight gain and seizures, which have also been noted in
1. Bullock R, Chesnut RM, Clifton G, et al. Guidelines for the individuals without TBI.
management of severe head injury. Brain Trauma Founda-
tion. Eur J Emerg Med. Jun 1996;3(2):109–127. There is insufficient evidence in TBI populations to sup-
2. Cicerone KD, Dahlberg C, Kalmar K, et al. Evidence-based port or refute the use of other commonly used medica-
cognitive rehabilitation: recommendations for clinical prac- tions for affective disorders, anxiety and psychosis. How-
tice. Arch Phys Med Rehabil. Dec 2000;81(12):1596–1615. ever, evidence of efficacy in other patient populations is
also a useful source of treatment options in TBI in many
cases.
AFFECTIVE DISORDERS, ANXIETY,
AND PSYCHOSIS II. Search Results
I. Recommendations The review of 43 papers resulted in two class II studies
and 12 class III studies that could be considered for evi-
Standards: There is insufficient evidence to support the
dence in the treatment of psychiatric sequelae after TBI.
development of standards in the treatment of TBI related
Although by research design, other studies seemed to qual-
depression, mania, anxiety and psychosis.
ify for class I, II or III evidence (ex. in three randomized
Guidelines: There is insufficient evidence to support
controlled trials by Adeloye,1 Kitamura,2 and Wroblewski
the development of guidelines in the treatment of TBI re-
et al.,3 significant methodological flaws seriously weak-
lated depression, mania, anxiety and psychosis.
ened the ability to generalize the findings and thus limited
Options:
the conclusions that can be drawn to that of treatment op-
Depression:
tions at best (see Table 2 for summary).
Tricyclic Antidepressants (TCAs)
The use of tricyclic antidepressants is recom-
A. Depression
mended as an option in the treatment of TBI related de-
pression. Specifically, amitriptyline (up to 300 mg/day) Class II. Two class II studies addressed the treatment
and desipramine (150–300 mg/day) have been reported to of depression. In a cohort study of 13 patients with mild
1475
WARDEN ET AL.
TABLE 2. AFFECTIVE DISORDERS, ANXIETY, AND PSYCHOSIS EVIDENCE
Data
Article Description of study class Conclusion
I. Major depression
Dinan, 19924 Cohort study of amitriptyline (up to II Only 4 MTBI patients showed marked improvement on
250 mg): 13 MTBI patients with CGI (vs. 11 of depressed patients without TBI).
depression matched with 13 Depression following MTBI is relatively resistant to
depressed patients without TBI treatment with tricyclic antidepressants—however,
amitriptyline was effective in some TBI patients.
Saran, 19855 Open label cohort study of amitriptyline II After four weeks of treatment, the mean Hamilton score
(200–300 mg per day) without placebo for the MTBI group improved some but was still in
condition. 21 depressed patients: 10 the clinically depressed range, whereas the non-injured
with historyof MTBI. group mean score was in the non-depressed range
(p 0.001). Depression following MTBI is relatively
resistant to treatment with tricyclic antidepressants.
Wroblewski, Randomized, placebo-controlled III Of the 7 patients who completed the study, 6 improved
19963 prospective crossover study of on desipramine treatment. Significant methodological
desipramine (150–300 mg/day): 10 flaws and small sample size limited the strength of the
individuals with severe TBI and findings.
depression
Fann, 20006 Nonrandomized, single-blind, placebo III Following eight weeks of treatment, 87% of patients
run-in trial of sertraline (25–200 mg/ were classified by investigators as responders and 67%
day) for the treatment of depression of patients as in remission based on their improvement
in 16 outpatients with mild TBI in scores on the Hamilton Depression Scale
Baker-Price, Case series: use of weak pulsed III After 5 sessions of therapy conducted over 5 weeks,
19967 magnetic fields in 4 patients with patients showed significant improvement in scores on
frequent or persistent depression the Beck Depression Inventory (p 0.04) and a
after TBI decrease in the magnitude of phobias (p 0.01).
Lack of a control group or condition and small sample
size limit the strength of the study findings.
Newburn, Retrospective case series of III 23 patients were considered responders based on
19998 moclobemide (450–600 mg/day): 26 improvement in scores on the Hamilton Depression
patients with TBI and a diagnosis of Scale. This medication is not FDA approved for use in
depression the United States.
Perino, 20019 Open trial of citalopram (20 mg/day) III Patients showed a significant improvement in BPRS
and carbamazepine (600 mg/day): (p 0.05) and Clinical Global Improvement Scores
20 patients with severe TBI and a (p 0.05). No specific depression rating scale was
diagnosis of depression used. Since combination therapy was used, it is not
possible to determine whether one or the other drug
was primarily responsible for the improvement in
rating scores.
II. Bipolar disorders/mania
Bakchine, Case study of clonidine (150–600 III Treatment with clonidine resulted in 37% symptom
198910 micrograms [mcg]/day) and other reduction, symptoms recurred during a placebo trial.
agents for a patient with severe Levodopa-benserazide resulted in 15% increase in
TBI and mania manicsymptoms. Patient was given clonidine again
(300 mcg/day) and had a 55% drop in symptoms from
baseline and no longer met criteria for diagnosis of
mania. Dosage maintained at 150 mcg/day.
Clark, 198711 Case series of mania in TBI: 1 patient III Case 1 was successfully managed on of amitriptyline
treated with thioridazine (50 mg/ and thioridazine for manic episodes and depression.
day) and amitriptyline (100 mg/ Case II, patient not compliant with lithium therapy
day), 1 patient treated with ECT and was treated with 4 sessions of ECT. When ECT
was discontinued, manic symptoms recurred. 2 more
ECT sessions again reduced symptoms and patient
could then be maintained on lithium.
1476
TABLE 2. AFFECTIVE DISORDERS, ANXIETY, AND PSYCHOSIS EVIDENCE (CONT’D)
Data
Hale, 198212 Open label trial of lithium (900 mg/ III Both patients experienced a marked improvement of
day) in two patients with moderate manic symptoms in the trial. Findings are limited by
to severe TBI and described manic the lack of formal diagnostic criteria, standardized
symptoms outcome measures and placebo condition, as well as
the small sample size. More supporting reports are
needed.
Pope, 198813 Open label trial of valproate (max III Both patients failed to respond to standard treatments
dose 750–100 mg/day): 2 patients including lithium and neuroleptics, but bipolar illness
with bipolar syndromes after remitted after valproate was added. The use of
mild to moderate TBI multiple medications, small sample size and use of
poorly described, non-standardized outcome measures
limit the results.
III. Anxiety disorders
Khouzam, Retrospective study of venlafaxine III Venlafaxine resulted in a decrease in compulsive
199814 (150 mg/day) in the treatment of symptoms (Y-BOCS score decreased from 35 to 3),
a patient with compulsions after reportedly stable over 4 months. Limitations include
TBI lack of a control condition or withdrawal from
medication, the limitations of a single case report and
the lack of formal diagnostic criteria.
IV. Psychotic disorders
Butler, Case report: Olanzapine (5 mg/day) III After 10 days of treatment with olanzapine, the
200015 for Cotard’s syndrome and Capgras individual’s mood stabilized, and there was a complete
delusion following severe TBI resolution of dysphoria, restless apprehension, and
delusional ideation.
Umansky, Case report: Olanzapine (20 mg/day) III After 6 months of treatment with olanzapine, the
200016 for psychosis following second individual no longer heard persecutory voices, had no
severe TBI delusional symptoms or rage outbursts and exhibited
improvements in mood, behavior and follow-up
compliance.
TBI (MTBI) matched with 13 depressed patients without atively resistant to treatment with tricyclic antidepres-
brain injury, Dinan and Mobayed4 studied the effective- sants. Although the authors stated that the groups differed
ness of amitriptyline in the treatment of depression. Par- significantly in the total number of treatment responders
ticipants were responders to a questionnaire sent to indi- (defined as a 50% drop in symptoms), there was no
viduals seen in an emergency department who met entry data presented to support this statement. Other signifi-
criteria following a TBI. Individuals who met cut-off cant limitations included the lack of a placebo condition,
scores for depression were then invited for full assess- and the small sample size. Several MTBI patients did re-
ment. Patients with depression and no history of TBI were spond to treatment and given the small sample size, the
drawn from outpatient referrals. Patients received a start- reviewers could not conclude that MTBI patients were
ing dose of 100 mg of amitriptyline. Dosage was in- resistant to amitriptyline treatment.
creased by 50 mg per week as tolerated over a 6-week Saran5 compared the response of 21 depressed patients
treatment period for a maximum total dosage of 250 mg. to amitriptyline (200–300 mg per day). Ten of the par-
Total dosage did not differ significantly between the two ticipants had experienced a MTBI (defined as loss of con-
groups; the TBI group had a mean daily dose of 158 mg sciousness less than 20 min), 11 had no history of TBI.
per day. In depressed patients with no history of brain in- This was an open label cohort study without a placebo
jury, 11 of 13 showed significant improvement on the condition. Response was measured using the Hamilton
Clinical Global Impression Scale (CGI) and the Hamil- Rating Scale for Depression. At the end of 4 weeks of
ton Rating Scale for Depression (HRSD). Only 4 MTBI treatment although the mean Hamilton score for the
patients showed marked improvement on the CGI. The MTBI group improved some, it was still within the clin-
authors concluded that depression following MTBI is rel- ically depressed range, whereas the non-injured group
1477
WARDEN ET AL.
mean score was in the non-depressed range. Analysis of Newburn and colleagues8 reported a significant im-
covariance using baseline Hamilton score as the covari- provement in scores on the Hamilton Depression Scale
ate showed a significant between group difference (p in 23 of a series of 26 patients with TBI and a DSM-III
0.001). diagnosis of depression treated with moclobemide.
Severity of the patients’ injuries could not be determined
Class III. Wroblewski and colleagues3 studied the ef- from the report and the study was an unblinded retro-
ficacy of desipramine in the treatment of depression in in- spective analysis. Although this agent appears to show
dividuals after TBI. This study was a randomized, promise in the treatment of post-TBI depression, we are
placebo-controlled prospective study, in which 10 indi- unable to recommend medications not FDA approved for
viduals with severe TBI and depression were given use in the United States.
placebo and desipramine in a crossover fashion. All sub- Twenty patients with severe TBI and a DSM-IV diag-
jects were diagnosed using DSM III-R criteria. Of the nosis of depression were given an open trial of citalo-
seven patients who could be evaluated based on the DSM- pram and carbamazepine.9 Patients showed a significant
III-R criteria (two dropped out due to adverse events: ma- improvement in BPRS and Clinical Global Impressions
nia and seizures; one refused to be evaluated), six im- Scores. However, no specific depression rating scale was
proved on desipramine treatment. Again there were a used. As the study used a combination of citalopram and
number of methodological concerns including small sam- carbamazepine, it is not possible to determine whether
ple size, use of other psychotropic treatments, a short one or the other drug was primarily responsible for the
placebo lead in period, use of unspecified exclusion cri- improvement in rating scores or whether it was indeed
teria, the failure of all subjects to meet DSM-III-R MDD the combination therapy that was effective. Thus a rec-
criteria, a relatively mild level of depression severity, and ommendation for either citalopram or carbamazepine
a lack of clarity of who was blind to active treatment af- could not be supported.
ter the first month. These concerns limited the strength of
evidence from this report and those of Dinan and
B. Mania/Bipolar Disorders
Mobayed4 and Saran5 to the level of an option for tricyclic
antidepressants in the treatment of depression after TBI. Class III. Bakchine and colleagues10 treated a patient
Sertraline has also been studied in a series of 16 out- with severe TBI and a DSM-III-R diagnosis of organic
patients with mild TBI and a DSM-III-R diagnosis of de- affective disorder, manic type, with several drugs and
pression.6 Following 8 weeks of treatment, 87% of pa- placebo. The individual experienced more than a 37% re-
tients were classified by investigators as responders and duction in symptoms measured with the Manic Rating
67% of patients as in remission based on their improve- Scale following 8 days of treatment with clonidine (600
ment in scores on the Hamilton Depression Scale. The micrograms/day). Symptoms recurred during a placebo
patients in this study were not randomized to treatment trial. Carbamazepine was also tried (600–1200 mg/day),
condition, there was no control group used, and the in- but had no effect on manic symptoms. A trial of lev-
vestigator was not blind to the treatment given. This well odopa-benserazide (375 mg/day) resulted in a 15% in-
designed class III study supports the recommendation of crease in manic symptoms and this treatment was dis-
sertraline for the treatment of depression after TBI at the continued after 5 days. The patient was then given
option level. clonidine again (300 micrograms/day) and experienced a
A study by Baker-Price and Persinger7 examined the 55% drop in symptoms from baseline and the patient no
use of weak pulsed magnetic fields in 4 patients with fre- longer met criteria for diagnosis of mania. Symptom im-
quent or persistent depression after TBI as diagnosed by provement continued at a slower rate when the dosage
their physician. The patients were within 6 years of in- was maintained at 150 micrograms/day. Twenty-month
jury and refractory to treatment with antidepressants. follow-up found the patient discharged, living alone with
Severity of the patients’ injuries could not be determined little support, and experiencing no manic or depressive
from the report. After 5 sessions of therapy conducted symptoms. While replication is needed to support a treat-
over 5 weeks, patients showed significant improvement ment recommendation as only one individual was stud-
in scores on the Beck Depression Inventory and a de- ied, this report suggests that clonidine can be used suc-
crease in the magnitude of phobias. Lack of a control cessfully in the treatment of mania following severe TBI.
group or condition and small sample size limit the Clark and Davison11 described the treatment of two
strength of the study findings. More supporting reports patients with bipolar disorder after TBI. Case I was a 70-
are needed to support a recommendation for the use of year-old man treated with thioridazine (100 mg/day) for
weak pulsed magnetic fields in the treatment of depres- manic symptoms following TBI. Two months later, he
sion after TBI. developed depressive symptoms and was successfully
1478
treated with 100 mg/day of amitriptyline and reduction lafaxine (150 mg/day) resulted in a decrease in compul-
of thioridazine to 50 mg/day. He was discharged to home sive symptoms (Y-BOCS score decreased from 35 to 3),
and maintained on that regimen without symptom recur- which was reportedly stable over 4 months. The dramatic
rence. Case II was a 60-year-old man with manic be- symptom improvement reported must be tempered by the
havior including physical and verbal aggression after lack of a control condition or withdrawal from medica-
TBI. Tranquilizers were ineffective at controlling his be- tion, the limitations of a single case report and the lack
havior and he was not compliant with lithium therapy. of formal diagnostic criteria.
After two sessions of bilateral ECT, the authors report
that he was less irritable and more cooperative with med- D. Psychotic Disorders
ical treatment. After two more sessions, his mood was
Class III. Only two articles were rated as class III in
only mildly elevated and he was experiencing normal
the area of pharmacological treatment of psychotic dis-
sleeping and eating patterns. When ECT was discontin-
orders after TBI. Butler15 described a 17-year-old man
ued, the manic symptoms recurred. Two more treatments
with Cotard’s syndrome and Capgras delusion following
again resulted in significant symptom improvement and
severe TBI. After 10 days of treatment with olanzapine
at that point the patient could be maintained on lithium
(5 mg/day), the individual’s mood stabilized, and there
therapy. The findings of this case series are tempered by
was a complete resolution of dysphoria, restless appre-
several weaknesses such as the small sample size and
hension, and delusional ideation. The second class III
lack of formal diagnostic and response criteria and can-
study was also a case report of olanzapine. This report16
not support a treatment recommendation.
details the case of an individual with psychosis after ex-
Hale and Donaldson12 described an open-label trial of
periencing a second severe TBI. Olanzapine (10 mg/day)
lithium (900 mg/day) in two patients with moderate to
was added to his valproic acid treatment (600 mg/day).
severe TBI and described manic symptoms. While both
After two months, some improvement was noted and the
patients experienced a significant improvement of manic
olanzapine dosage was increased to 20 mg/day. After 6
symptoms in the trial, these findings are limited by the
months of treatment at this dose, the individual no longer
lack of formal diagnostic criteria, standardized outcome
heard persecutory voices, had no delusional symptoms or
measures and placebo condition, as well as the small sam-
rage outbursts and exhibited improvements in mood, be-
ple size. More supporting reports are needed to support
havior and follow-up compliance. These reports are lim-
a recommendation for the use of lithium in the treatment
ited by small sample size, lack of standardized diagnos-
of mania after TBI.
tic and outcome measures and lack of a placebo or other
Pope and colleagues13 conducted an open-label trial of
control condition. However, evidence supports the rec-
valproate for bipolar syndromes after TBI. Two patients
ommendation of olanzapine as an option for the treat-
with mild to moderate TBI had sufficient clinical data
ment of psychosis after TBI.
presented to be considered for evidence. Both patients
failed to respond to standard treatments including lithium
and neuroleptics, but bipolar illness remitted after val- E. Other Reports
proate (maximum dose 750–100 mg/day) was added to A large number of other case series and case reports
the regimen. However, the use of multiple medications, address the treatment of depression, anxiety, mania or
small sample size and use of poorly described, non-stan- psychosis after TBI. However, a variety of methodolog-
dardized outcome measures limit the ability to general- ical problems limited the conclusions that could be drawn
ize the results. Again, more evidence is needed to sup- from many of these reports. These methodological issues
port the use of valproate for bipolar syndromes after TBI. took several forms including samples of mixed popula-
tions (e.g., stroke, penetrating brain injury, TBI,
epilepsy),17–19 more than one psychiatric diagnosis,17,20
C. Anxiety Disorders absence of clear diagnostic criteria and inadequate infor-
Class III. There were no class I or II studies found mation given to substantiate the authors diagno-
which addressed the treatment of anxiety disorders. Al- sis,1,17,19,21–26 absence of validated or clear outcome
though some class III studies addressed this patient pop- measures and inadequate information given to substanti-
ulation, the degree of methodological flaws in these stud- ate authors’ claims of improvement,1,19–30 use of more
ies rendered all but one unusable. than one pharmacologic intervention at a time making it
Only one case report could be considered as class III difficult to attribute the improvement to the putative
evidence. The report was a retrospective study of ven- agent,21,27,28,30 absence of placebo condition and/or
lafaxine in the treatment of compulsions in an individual crossover design,19,21–30 and poorly documented link be-
with TBI of unknown severity.14 Treatment with ven- tween remote TBI and current symptoms.24 Several class
1479
WARDEN ET AL.
III studies have addressed the use of selective serotonin Do TBI patients with mood, anxiety, or psychotic disor-
re-uptake inhibitors in this population, but again basic ders respond to standard treatments? (2) Are certain symp-
methodological flaws, and small numbers make it diffi- toms more or less responsive in TBI patients than in id-
cult to draw conclusions from even the better class III iopathic cases? (3) Does treating psychiatric syndromes
studies. improve functional outcome in TBI patients? (4) Are cer-
tain pharmacologic therapies more effective than others?
III. Consideration of Potential Adverse (5) Are side effects more burdensome in TBI patients than
Side Effects in idiopathic cases? (6) Does TBI severity influence treat-
As noted, one of the common tenets of clinical lore ment outcome? (7) Does the presence of multiple psy-
holds that individuals with TBI may be more sensitive to chiatric syndromes affect treatment response?
the side effects of psychotropic medications and other As the etiology of TBI-associated mental disorders re-
pharmacologic interventions. Thus we felt it important to mains unknown, the most easily answered questions can
review reports of potential toxicity associated with phar- be tested first, especially regarding the efficacy of exist-
macologic interventions. While space does not permit a ing standard treatments for TBI patients in randomized-
comprehensive review of these reports, the following pro- controlled clinical trials. Without the ability to discern
vides a summary of agents and their potential side effects etiology, standard syndromal criteria, such as those elab-
in the TBI population as reported in the literature: orated in the DSM-IV, would be a reasonable starting
seizures with tricyclic antidepressants31; weight gain, point for study inclusion. Other inclusion criteria might
drooling and seizures with clozapine32; mania with de- include patients with similar level of TBI severity and
sipramine33; dysarthria and speech blocking after fluox- cognitive capacity, absence of current substance abuse,
etine treatment34; severe akathisia with both sertraline absence of current epilepsy, absence of prior or comor-
and paroxetine35; and increased cognitive impairment bid psychiatric disorder, and development of symptoms
with lithium (with serum lithium levels at 1.0 meq/L that within a circumscribed time period of the TBI (e.g., 2
resolved when reduced to 0.5 meq/L).36 years). In the absence of validated rating scales for TBI
The overall impression from reading the literature is patients with mood or anxiety disorders, standard symp-
that the full range of side effects associated with the use tom rating instruments (i.e., Hamilton Anxiety and
of psychotropic medications seen in the non-brain injured Hamilton Depression Rating Scales, MADRAS, CGI) in
population can be seen in individuals with TBI. It is not combination with other symptom check-lists (SCL-90)
possible to conclude from this literature that individuals and validated functional rating instruments (i.e., SF-36)
with TBI are more sensitive to these side effects, nor is would be reasonable outcome measures.
it apparent that novel side effects occur in this group. Which treatments to study remain at issue. Effective
standard treatments for idiopathic mood or anxiety dis-
IV. Conclusions orders, especially those with a favorable safety profile
would be reasonable first candidates. The largest number
There is limited evidence in the published literature to of previously studied patients took tricyclic antidepres-
support or refute the use of psychotropic medications sants, though the benefits of the tricyclic antidepressants
commonly used in the general population for individuals are mixed, one study suggesting that mild TBI patients
with traumatic brain injury. There is insufficient evidence are minimally responsive to amitriptyline, but another re-
to support any standards or guidelines for the treatment port suggesting that patients with severe TBI may im-
of affective disorders, mania, or psychosis in this popu- prove with desipramine (Wroblewski, 1996), even though
lation. However, options are recommended for the treat- toxicity may be problematic. Hence, the tricyclics are not
ment of major depression with amitriptyline, desipramine the likeliest first choice for controlled investigation be-
and sertraline, as well as olanzapine for the treatment of cause of their questionable benefit and propensity for sig-
psychosis. Clearly, more well designed and executed ran- nificant toxicity. Similarly, the propensity for toxicity
domized controlled trials are needed to examine the ef- makes the MAOI antidepressants unlikely initial choices
fectiveness of pharmacotherapy for these disorders in infor controlled study. Although electroconvulsive therapy
dividuals with TBI. is often helpful in severe cases of mania or depression,
and there is anecdotal evidence for its benefit in TBI pa-
V. Recommendations for Future Research
tients, this treatment should likely be reserved for re-
Even though mood and anxiety disorders following TBI fractory cases because of its propensity for cognitive tox-
may be etiologically distinct and have phenomenological icity.
differences from idiopathic cases, several critical ques- The serotonin reuptake inhibitor (SRI) antidepressants
tions remain regarding future directions for treatment: (1) have become the consensus first-line treatments for idio-
1480
pathic mood and anxiety disorders because of their broad- tients with severe traumatic brain injury: a controlled,
spectrum effectiveness. These agents would be the nat- prospective study. Journal of Clinical Psychiatry. 1996;
ural first-line choices for study in TBI patients. Any of 57(12):582–587.
the agents would be reasonable candidates, though the 4. Dinan TG, Mobayed M. Treatment resistance of depression
sedation of fluvoxamine would render it less desirable after head injury: a preliminary study of amitriptyline re-
than other SRIs. The preponderance of evidence with flu- sponse. Acta Psychiatrica Scandinavica. 1992;85:292–294.
oxetine and sertraline support their use in depression. In 5. Saran AS. Depression after minor closed head injury: role
anxiety disorders, any of the SRIs would be potentially of dexamethasone suppression test and antidepressants.
useful for study, though fluoxetine may be the least de- Journal of Clinical Psychiatry. 1985;46:335–338.
sirable for panic disorder because of its propensity exac- 6. Fann JR, Uomoto JM, Katon WJ. Sertraline in the treat-
erbate panic at treatment onset. ment of major depression following mild traumtic brain in-
Similarly, for mania following TBI, standard treat- jury. Journal of Neuropsychiatry and Clinical Neuro-
ments, such as valproic acid or lithium, should be exam- sciences. 2000;12(2):226–232.
ined first. Following that, atypical antipsychotics in com- 7. Baker-Price LA, Persinger MA. Weak, but complex pulsed
bination with or separate from standard mood stabilizers magnetic fields may reduce depression following traumatic
would be reasonable choices. Persisting psychotic symp- brain injury. Perceptual and Motor Skills. 1996;83:
toms associated with manic syndromes are relatively un- 491–498.
common following TBI, though they too should be stud- 8. Newburn G, Edwards R, Thomas H, et al. Moclobemide in
ied. Treatment with atypical antipsychotics should be the treatment of major depressive disorder (DSM-3) fol-
examined first because patients with neurological pathol- lowing traumatic brain injury. Brain Injury. 1999;13(8):
ogy are prone to develop extrapyramidal symptoms. 637–642.
Compared with mood and anxiety disorders, psychotic 9. Perino C, Rago R, Cicolin A, et al. Mood and behavioural
disorders uncommonly follow TBI and there is little in- disorders following traumatic brain injury: clinical evalu-
formation to guide treatment of post-TBI psychotic syn- ation and pharmacological management. Brain Injury.
dromes. Future studies could examine the treatment of 2001;15(2):139–148.
psychosis in several categories: (1) with or without as- 10. Bakchine S, Lacomblez L, Benoit N, et al. Manic-like state
sociated significant mood symptoms; (2) with or without after bilateral orbitofrontal and right temporoparietal in-
concomitant seizures; (3) with or without significant cog- jury: efficacy of clonidine. Neurology. 1989;39:777–781.
nitive impairment. Ideally, patients for study would in- 11. Clark AF, Davison K. Mania following head injury: a re-
clude those whose psychosis developed within a reason- port of two cases and a review of the literature. British
able time period from the TBI (i.e., 2 years) and were not Journal of Psychiatry. 1987;150:841–844.
associated with substance abuse. 12. Hale MS, Donaldson JO. Lithium carbonate in the treat-
Few patients would likely experience all the DSM-IV ment of organic brain syndrome. Journal of Nervous and
signs and symptoms of schizophrenia, though patients Mental Disease. 1982;170(6):362–365.
reasonable for study would include those with consistent 13. Pope HG, McElroy SL, Satlin A, et al. Head injury, bipo-
and bothersome psychotic signs or symptoms (i.e., delu- lar disorder, and response to valproate. Comprehensive Psy-
sions or hallucinations). Standard rating scale instruments chiatry. 1988;29(1):34–38.
such as the BPRS or PANSS should be employed. As the 14. Khouzam HR, Donnelly NJ. Remission of traumatic brain
older antipsychotics are more prone to induce extrapyra- injury-induced compulsions during venlafaxine treatment.
midal effects in patients with brain injury, the newer an- General Hospital Psychiatry. 1998;20:62–63.
tipsychotics (i.e., risperidone, olanzapine, quetiapine or 15. Butler PV. Diurnal variation in Cotard’s syndrome (copre-
ziprasidone) would be reasonable choices for study. sent with Capgras delusion) following traumatic brain in-
jury. Australian and New Zealand Journal of Psychiatry.
2000;34:684–687.
REFERENCES 16. Umansky R, Geller V. Olanzapine treatment in an organic
hallucinosis patient. International Journal of Neuropsy-
1. Adeloye A. Clinical trial of fluphenazine in the post-con- chopharmacology. 2000;3:81–82.
cussional syndrome. Practitioner. 1971;206:517–519.
17. Smith RB, Tiberi A, Marshall J. The use of cranial elec-
2. Kitamura K. Therapeutic effect of pyritinol of sequelae of trotherapy stimulation in the treatment of closed-head-in-
head injuries. Journal of International Medical Research. jured patients. Brain Injury. 1994;8(4):357–361.
1981;9:215–221.
18. Ruedrich SL, Chu C, Moore SL. ECT for major depression
3. Wroblewski BA, Joseph AB, Cornblatt RR. Antidepressant in a patient with acute brain trauma. American Journal of
pharmacotherapy and the treatment of depression in pa- Psychiatry. 1983;140:928–929.
1481
WARDEN ET AL.
19. Sloan RL, Brown KW, Pentland B. Fluoxetine as a treat- in traumatic brain injury. Brain Injury. 1997;11(4):
ment for emotional lability after brain injury. Brain Injury. 287–291.
1992;6(4):315–319. 35. Hensley PL, Reeve A. A case of antidepressant-induced
20. Griffith JP, Kamthan M. Obsessive-compulsive disorder akathisia in a patient with traumatic brain injury. Journal
following closed head injury. West Virginia Medical Jour- of Head Trauma Rehabilitation. 2001;16(3):302–305.
nal. 1998;94:198–201. 36. Hornstein A, Seliger G. Cognitive side effects of lithium
21. Wroblewski BA, Guidos A, Leary J, et al. Control of de- in closed head injury. Journal of Neuropsychiatry. 1989;
pression with fluoxetine and antiseizure medication in a 1(4):446–447.
brain-injured patient. American Journal of Psychiatry.
1992;149(2):273.
COGNITIVE DISORDERS
22. Childers MK, Holland D, Ryan MG, et al. Obsessional dis-
orders during recovery from severe head injury: report of
I. Recommendations
four cases. Brain Injury. 1998;12(7):613–616.
23. Mas F, Prichep LS, Alper K. Treatment resistant depres-
Standards. There are insufficient data to support a treat-
sion in a case of minor head injury: an electrophysiologi- ment effect for any of the cognitive functions considered.
cal hypothesis. Clinical Electroencephalography. 1993; GUIDELINES
24(3):118–122. A. General Cognitive Functions. There are insufficient
data to support a treatment effect for this function. How-
24. Greenberg DB. Buspirone for myoclonus, obsessive fears,
and confusion. Psychosomatics. 1993;34(3):270–272.
ever, phenytoin should not be given in severe TBI unless
medically indicated, as there is evidence that it can pro-
25. Moran AJ. Post-traumatic depression treated by exercise duce an impairment in cognitive functions in severe TBI
and monoamine oxidase inhibitors. Medical Journal of at 1 month post-event, although not at 12 months. Where
Australia. 1975;2(23):888.
ongoing anticonvulsant treatment is required, valproate
26. Schreiber S, Klag E, Gross Y, et al. Beneficial effect of or carbamazepine appear to be preferable to phenytoin in
risperidone on sleep disturbance and psychosis following terms of cognitive effects.
traumatic brain injury. International Clinical Psychophar- B. Deficits in Attention and Speed of Processing
macology. 1998;13(6):273–275. Stimulants: Methylphenidate (0.25–0.30 mg/kg bid)
27. Stewart JT, Hemsath RH. Bipolar illness following trau- is recommended to enhance attentional function. The ev-
matic brain injury: treatment with lithium and carba- idence is strongest for an effect on speed of cognitive
mazepine. Journal of Clinical Psychiatry. 1988;49(2): processing and sustained attention/vigilance. Methyl-
74–75. phenidate (0.25–0.30 mg/kg bid) is also recommended to
28. Sinanan K. Mania as a sequel to a road traffic accident. enhance the speed of cognitive processing, although only
British Journal of Psychiatry. 1984;144:330–331. one study provides evidence to support a change in speed
29. Bouvy PF, van de Wetering BJM, Meerwaldt JD, et al. A in a naturalistic task.
case of organic brain syndrome following head injury suc- Cholinesterase Inhibitors: Donepezil (5–10 mg/day)
cessfully treated with carbamazepine. Acta Psychiatrica is recommended to enhance aspects of attention for pa-
Scandinavica. 1988;77:361–363. tients with moderate to severe TBI in subacute and
chronic periods of recovery.
30. Bracken P. Mania following head injury. British Journal
of Psychiatry. 1987;150:690–692.
C. Deficits in Memory
Cholinesterase Inhibitors: Donepezil (5–10 mg/day)
31. Wroblewski BA, McColgan K, Smith K, et al. The inci- is recommended to enhance aspects of memory function
dence of seizures during tricyclic antidepressant drug treat- for patients with moderate to severe TBI in subacute and
ment in a brain-injured population. Journal of Clinical Psy-
chronic periods of recovery.
chopharmacology. 1990;10(2):124–128.
D. Deficits in Executive Functions
32. Michals ML, Crismon ML, Roberts S, et al. Clozapine re- Dopamine Enhancers: Bromocriptine in a dose of
sponse and adverse effects in nine brain-injured patients. 2.5 mg is recommendation for use in enhancing aspects
Journal of Clinical Psychopharmacology. 1993;13(3): of executive functioning (e.g., divided attention/central
198–203.
executive functions) in patients with severe TBI.
33. Santos AB, Ballenger JC. Tricyclic antidepressant triggers OPTIONS
mania in patient with organic affective disorder. Journal of A. General Cognitive Functions
Clinical Psychiatry. 1992;53(10):377–378. Stimulants: Methylphenidate is recommended at the
34. Patterson DE, Braverman SE, Belandres PV. Speech option level for general cognitive functioning in moder-
dysfunction due to trazodone-fluoxetine combination ate to severe TBI.
1482
Dopamine Enhancers: Amantadine may be consid- There is insufficient evidence in TBI populations to sup-
ered for general cognitive functioning in moderate to se- port or refute the use of other commonly used medica-
vere TBI. tions for cognitive disorders. However, evidence of effi-
B. Deficits in Attention and Speed of Processing cacy in other patient populations is also a useful source
Stimulants: Dextroamphetamine may be considered of treatment options in TBI in many cases.
for use in enhancing attentional function in patients with
TBI. In particular, dextroamphetamine may reduce vari- II. Search Results
ability in performance in tasks of attention and working
Of a total of 61 papers meeting the cognitive disorders
memory but this phenomenon was reported in only one
subgroup inclusion criteria, only seven were deemed
study.
class I evidence, six qualified as class II evidence, and
Dopamine Enhancers: Amantadine may be consid-
22 were class III; the remainder (27) were unusable. Two
ered to improve attention and concentration in moderate
additional articles were identified in the updated search
to severe TBI.
of clinical trials conducted in the fall of 2004 and were
Cholinesterase Inhibitors: Physostigmine may be
also rated. Evidence is presented in Table 3 divided by
considered for use in enhancing aspects of attentional
cognitive functional category and by drug class within
function in patients with moderate-to-severe TBI in the
categories.
subacute to chronic phase of recovery.
C. Deficits in Memory
Stimulants: Methylphenidate in a dose of 0.30 mg/
A. General Cognitive Functions
kg bid may be considered as an option to enhance learn- These functions include level of consciousness, atten-
ing and memory. tion, memory, and executive-type functions. This cate-
Other Agents: CDP choline (cytidine diphospheryl gory was used when either (a) studies demonstrated an
choline, 1 gram) may be considered for use in enhanc- effect on two or more of these functions or (b) when the
ing aspects of memory function in patients with mild to tasks reported did not allow a more specific assignment
moderate TBI in the subacute phase of recovery. of effect(s).
TABLE 3. COGNITIVE DISORDERS EVIDENCE
Data
I. General cognitive function

Dikmen, 19911 Double-blind, randomized, placebo- I Within 1 month of injury, phenytoin produced
controlled study of phenytoin; 244 adults significant impairment in cognitive functions
with a moderate to severe brain injury (p 0.05). At 12 months post-injury, there
and post-traumatic epilepsy. was no difference between the medication and
placebo groups in general cognitive functioning.
Dikmen, 20002 Randomized, double-blind parallel group I No significant effect of valproate acid (positive or
clinical trial compared the seizure negative) on neuropsychological function at
prevention and neuropsychological 1, 6, or 12 months.
effects of VPA to phenytoin in 279
patients with TBI.
Azouvi, 19993 Open label trial of carbamazepine (400–800 III No global cognitive change found on Mini
mg/day) on agitation and aggressive Mental Status Examination scores.
behavior in 10 patients with severe TBI
Plenger, 19964 Double-blind placebo controlled, II At 1 month, improvement was noted with
randomized trial of methylphenidate methylphenidate on functional outcome on the
(0.6 mg/kg per day) in 23 patients with DRS (p 0.02). At 90 days, there was no
complicated mild to moderately severe significant difference between groups.
brain injury.
Kaelin, 19965 Cohort study of methylphenidate III Trend for improvement on DRS scores (p 0.06).
(30 mg/day) in 10 patients with mixed
severity TBI performing two standard
deviations below the mean on
neuropsychological measures. (continued)
1483
WARDEN ET AL.
TABLE 3. COGNITIVE DISORDERS EVIDENCE (CONT’D)
Data
Kraus, 19976 Case Series of amantadine (up to III 3/5 TBI patients showed improvement (2 STD)
400 mg/day) in 5 patients with severe on at least one neuropsychological test. Patients
TBI and cognitive or behavioral deficits. showed improvement on executive domains
including tests of verbal fluency initiation and
sequencing, as well as overall improved
functioning based on caregiver ratings.
Kraus, 19977 Case report of amantadine for persistent III Improvement in global functioning, divided
frontal lobe dysfunction after severe attention and constructional praxis. Further
TBI. improvement was observed with the addition
of L-dopa/carbidopa.
Schneider, Double blind placebo controlled III Although patients improved over time, they
19998 crossover study of amantadine (100– showed no positive effects from treatment with
300 mg/day) in 10 patients with amantadine over placebo.
moderate to severe TBI and measured
deficits in attention/concentration.
Chapman, Randomized, double-blind, placebo I No effect on measured cognitive functioning.
19999 controlled study of homeopathy for However, there was a significant effect
patients with mild TBI. (p 0.01) on a self-report measure.
McLean, Double-blind placebo controlled III Patients had improvement in verbal memory, but
199110 crossover trial of pramiracetam no effect on attention or speed of information
(1200 mg/day) on memory and other processing
cognitive problems in 2 TBI patients.
II. Deficits in attention
Whyte, 199711 Randomized, placebo-controlled, I Significant drug by performance variable interaction
repeated crossover study of (p 0.01). Specific improvement noted on speed
methylphenidate (0.5 mg/kg per day) for of performance and arousal. Other attentional
attentional problems in 19 patients with components (e.g., distractibility and vigilance)
TBI and. and motor speed were not affected by MP.
Whyte, 200412 Randomized, placebo-controlled, I Patients showed significant improvement on
repeated crossover study of speed of information processing (p 0.01),
methylphenidate (0.6 mg/kg per day) for attentiveness during work tasks (p 0.01) and
attention complaints in 34 patients with caregiver ratings of attention (p 0.01).
moderate to severe TBI.
Plenger, Double-blind placebo controlled, II At 1 month, improvement was noted on tests of
19964 randomized trial of methylphenidate attention and motor performance (p 0.05). At
(0.6 mg/kg per day) in 23 patients with 90 days, there was no significant difference
moderate to moderately severe brain between groups. Possible enhanced rate of
injury. recovery for the drug group in the subacute
recovery period.
Speech, 199313 Double-blind placebo controlled, II No significant effect was found on attentional
randomized crossover trial of measures.
methylphenidate (0.6 mg/kg per day) in
12 patients with moderate to severe TBI.
Gualtieri, Double-blind placebo controlled, II Trend for significance in 10 responders for
198814 randomized crossover study of attention and memory.
methylphenidate (0.15 to 0.3 mg/kg
bid) in 15 patients with severe TBI
and mild to moderate deficits
Kaelin, 19965 Cohort study of methylphenidate III Patients had significant improvement on digit
(30 mg/day) in 10 patients with mixed span, mental control and symbol search
severity TBI performing two standard (p 0.05).
deviations below the mean on
neuropsychological measures.
1484
Data
Hornstein, Retrospective chart review of patients III Ten of 22 TBI patients experienced a positive
199615 with severe TBI treated with effect on attention as indicated by their ability
dextroamphetamine (5–30 mg/day) for to continue in rehabilitation.
severe attentional or initiation problems
interfering with rehabilitation.
Bleiberg, Double-blind, placebo crossover case III Performance on a computerized battery of
199316 study of lorazepam (0.5 mg) and attention and working memory tasks was less
dextroamphetamine sulfate (5 mg) in a variable after administration of dextroampheta-
patient with TBI and complaints of mine. Lorazepam caused more variable
concentration difficulties. performance on one of the subtests.
Generalizability of this finding is limited
because this is a single case and the measures
are experimental.
Zhang, 200417 Double-blind, placebo-controlled I The results demonstrated statistically significant
crossover design of donepezil (5–10 improvement in attention and processes with
mg/day) in 18 patients with moderate donepezil. Moreover, improved test scores were
to severe TBI with attention and/or sustained in Group A in the placebo phase.
memory impairment.
Cardenas, Double-blind, placebo-controlled III Mean digit symbol scores improved with oral
199418 crossover design of scopolamine physostigmine in the responders but not in the
(5 g/h) and physostigmine (2–4 mg non-responders. Possible practice effects and
TID) in 36 patients with severe TBI complex study design are the primary
and significant memory impairment limitations.
on WMS.
Levin, 198619 Double-blind, placebo-controlled I Demonstrated a trend toward significance
crossover study of oral physostigmine (p 0.07) of the effect of oral physostigmine
(3.0 or 4.5 mg/day) and lecithin in combination with lecithin on sustained
(16 g/day) in 16 patients with moderate attention (as measured by the Continuous
to severe head injury and residual Performance Test) when the oral physostigmine
memory deficit measured by the Benton was administered before the placebo.
Visual Retention Test
Kraus, 19976 Case series of amantadine (up to 400 III Statistically significant improvement on Trials A.
mg/day) in 5 patients with severe TBI
and cognitive or behavioral deficits.
Kraus, 19977 Case report of amantadine for persistent III Improvement for global functioning, divided
frontal lobe dysfunction after severe attention and constructional praxis was noted.
TBI. Further improvement was observed with the
addition of L-dopa/carbidopa (75/300 mg/day).
Schneider, Double blind placebo controlled III Although patients improved over time, they
19998 crossover study of amantadine (100– showed no positive effects from treatment with
300 mg/day) in 10 patients with amantadine over placebo.
moderate to severe TBI and measured
deficits in attention/concentration.
III. Deficits in memory
Zhang, Double-blind placebo-controlled I The results demonstrated improvement in memory
200417 crossover design of donepezil (5– processes with donepezil. Moreover, improved
10 mg/day) in 18 patients with test scores were sustained in Group A in the
moderate to severe TBI with attention placebo phase.
and/or memory impairment.
Taverni, Open-label case series (n 2, only 1 with II Patients showed improvement in memory. One
199820 objective data) of donepezil at least patient had measured improvement. In the other
one year post severe TBI with long case the staff noted improvement.
term memory dysfunction. (continued)
1485
WARDEN ET AL.
Data
Masanic, Open-label case series of donepezil (5–10 III Trend for improvement in memory and behavior
200121 mg/day) for 4 patients at least 2 years (Rey Auditory Verbal Learning Test, Complex
post TBI Figure Test, Rivermead Behavioural Memory
Test, Neuropsychiatric Inventory). Treatment
seemed to specifically impact new learning
and recall.
Cardenas, Double-blind placebo-controlled cross- II 16 subjects showed improvement in memory test
199418 over design of scopolamine (5 g/h) performance (storage and recall of information).
and oral physostigmine (2–4 mg TID) Possible practice effects and complex study
in 36 patients with severe TBI and design are the primary limitations.
significant memory impairment as
measured on WMS.
Levin, 198619 Double-blind placebo-controlled II Improvement was only seen on sustained
crossover study of oral physostigmine attention (see table above), not memory
(3.0 or 4.5 mg/day) and lecithin measures.
(16 g/day) in 16 patients with moderate
to severe head injury and residual
memory deficit measured by the Benton
Visual Retention Test
Gualtieri, Double-blind placebo controlled II Improved memory and attention test scores in
198814 randomized crossover study of 10 MPH responders.
methylphenidate (0.15–0.3 mg/kg bid)
in 15 patients with severe TBI and
mild to moderate deficits
Speech, 199313 Double blind placebo controlled II Study failed to show a significant effect of the
randomized crossover trial of study drug on memory measures. However,
methylphenidate (0.6 mg/kg per day) in statistical power was low for this study.
12 patients with moderate to severe TBI.
Mooney, Randomized placebo controlled single III No significant effects were noted on memory
199322 blind study of methylphenidate measures in patients receiving methylphenidate.
(30 mg/day) in 38 patients with
moderate to severe TBI.
Tiberti, Randomized double-blind, placebo- III No significant effects were noted on memory
199823 controlled study of methylphenidate measures.
(10–40 mg) in 10 patients with TBI
and amnesia.
Levin, 199124 Double-blind, placebo-controlled study II Results showed improvement in verbal and visual-
of CDP choline (1 g) in 14 patients spatial memory, although the only significant
with mild-to-moderate TBI in the finding was for spatial recognition memory in
subacute level of recovery after post which the CDP-choline group improved by over
traumatic amnesia (PTA) cleared. 100% versus 29% in the placebo group (p 0.02).
Fewtrell, Double-blind crossover study of III Patients had no significant improvement on verbal
198225 vasopressin (16 IU/day) conducted in or visual memory.
6 patients with severe TBI.
Eames, 1999 Cohort study of vasopressin (8 IU/day) III Patients had improved verbal and visual memory.
in 26 patients with severe TBI.
IV. Deficits in executive functions
McDowell, Double-blind, placebo-controlled I Treatment resulted in significant improvement on
199826 crossover trial of bromocriptine for executive function tasks thought to tap
patients with severe TBI. prefrontal functioning (p 0.05). No
improvement was seen for measures of working
memory that did not involve executive
control processes.
1486
Anticonvulsants. Phenytoin: A class I study by Dik- the option level for the treatment of attentional function
men and colleagues1 showed that phenytoin produced after TBI.
significant impairment in cognitive functions acutely (1
month post-injury) in patients with severe TBI. This Other agents. Homeopathy: A single class II study by
study focused on patients with post-traumatic epilepsy Chapman et al.9 compared a menu of 18 possible home-
and examined cognitive side effects of phenytoin in this opathic medications that were individually matched to the
population. At 12 months post-injury, there was no dif- clients’ pattern of symptoms vs. placebo treatment. Three
ference between the medication and placebo groups in to seven total doses were given. No significant differ-
general cognitive functioning. Valproate: One class I ences were noted in subsections of the Woodcock-John-
study of patients with post-traumatic epilepsy by Dikmen son tests of Cognitive Abilities–revised. However, self-
et al.2 showed no positive or negative cognitive side ef- reports on a variety of functional tasks tapping a range
fects of valproate. Carbamazepine: A class III open-la- of cognitive capacities, and self-reports of symptoms,
bel study examined the effect of carbamazepine (400–800 many of which concerned cognitive difficulties (e.g.,
mg/day) on agitation and aggressive behavior in 10 pa- short-term memory problems), though measured on un-
tients with severe TBI.3 There was no global cognitive validated scales, showed significantly greater improve-
change noted on Mini Mental Status Examination scores. ment on the active treatment than on placebo. While pa-
Therefore, the use of valproate or carbamazepine may be tients’ subjective improvement is noteworthy, as no
preferable to phenytoin in patients experiencing cogni- measurable improvement was found on formal cognitive
tive impairment after TBI. This recommendation is sup- testing, these results do not support a treatment recom-
ported at the guideline level. mendation. Pramiracetam: A class III study examined
the effects of pramiracetam (1200 mg/day) on memory
Stimulants. Methylphenidate (MP): One small ran- and other cognitive problems in 2 patients with TBI.10
domized controlled study (class II)4 and one small class The study found improvement in verbal memory, but no
III study5 reported greater early recovery with effect on attention or speed of information processing.
methylphenidate treatment, as measured by the Disabil- More well-designed and conducted studies are needed to
ity Rating Scale. The Plenger study was a parallel group support a recommendation for pramiracetam for cogni-
study, but the treatment group had significantly higher tive deficits after TBI.
GCS scores at treatment onset than the placebo group.
The Kaelin study was an AABA design, where the pace B. Deficits in Attention and Speed of Processing
of recovery was slightly greater during the treatment in-
terval that during the interval between the initial base- The assessment of drug benefits on attention is com-
lines. The results of these studies support recommenda- plicated by a lack of agreement about the relevant sub-
tion of methylphenidate as an option for the treatment of divisions within the broad domain of arousal/attention.
general cognitive deficits after TBI. Even within a given subdomain (e.g., distraction) there
is disagreement about the optimal measures to use for de-
Dopamine Enhancers. Amantadine: Three class III tection of drug benefits. Thus, some of the discrepancies
reports examine the use of amantadine for general neu- in results may be due to differences in the choice of mea-
rocognitive function in moderate to severe TBI with two sures to assess drug effects. In addition, most of the mea-
of the three reports demonstrating improvement on test- sures of drug benefit with respect to attention focus at
ing. One six-case and one single-case study by Kraus and the impairment level, making it difficult to assess the real-
Maki6,7 (class III) revealed evidence for improvement in world impact of any improvements noted.
cognitive function using amantadine (200–400 mg/day).
In the single case study further improvement was ob- Stimulants. Methylphenidate (MP): One class I
served with the addition of L-dopa/carbidopa (75/300 study11 and one class II study4 found beneficial effects
mg/day). In the case series, 3/5 TBI patients showed im- of MP on sustained attention/vigilance while a different
provement (2 STD) on at least one neuropsychological class II study,13 using a similar dose and methodology,
test. Patients showed improvement on executive do- found no significant benefit. Another class II study14 did
mains including tests of verbal fluency initiation and se- find improvement in a post hoc analysis in 10 of 15 pa-
quencing, as well as overall improved functioning based tients, i.e., medication responders. The available evidence
on caregiver ratings. One class III study by Schneider regarding the impact of methylphenidate on distractibil-
et al.8 reported no positive effects of amantadine ity is mixed. The class I study by Whyte and colleagues11
(100–300 mg/day) in seven TBI subjects. Overall, these failed to find a significant reduction in distractibility in
findings support the recommendation of amantadine at a laboratory task, but the effect size was medium and the
1487
WARDEN ET AL.
study was small. This study also assessed distractibility involves working memory and attentional processes. Pa-
in a naturalistic environment and found an effect size of tients in Group A received donepezil first while patients
zero. In their more recent class I study,12 however, MP in Group B received placebo first. The results demon-
was associated with less off-task behavior during indi- strated improvement in both attention and memory
vidual work time. The class II study13 using the Gordon processes with donepezil. Moreover, improved test scores
Diagnostic System, similarly found no improvement in were sustained in Group A in the placebo phase.
distraction. There is evidence for an effect on speed of Donepezil is recommended as a guideline for the treat-
processing. The two class I studies11,12 found improvement of attentional function after TBI. Physostigmine:
ments in processing speed during MP treatment. Only Two studies suppor the recommendation for physostig-
one class II study13 that looked specifically at process- mine at the option level. Cardenas et al.18 reported im-
ing speed, failed to find a positive effect. Another class proved scores on one measure tapping attentional
III study5 found a positive effect on speed in only one of processes in patients who were felt to be medication re-
two speed-related subtests. The collective evidence sup- sponders in terms of memory functioning. However, this
ports guidelines for MP in the treatment of both atten- was a complex study design, results were difficult to tease
tional function and speed of cognitive processing. Dex- out, and data and statistical analyses were not reported.
troamphetamine: Two class III studies support the use The other study19 was a double-blind placebo-controlled
of dextroamphetamine for attention in patients with TBI. study of oral physostigmine in combination with lecithin
Hornstein and colleagues performed a retrospective chart in a small group of patients. While multiple measures of
review of patients with severe TBI treated with dex- attention and memory were used, there was a statistical
troamphetamine (5–30 mg/day) for severe attentional or trend for patients on physostigmine to show improvement
initiation problems interfering with rehabilitation.15 Ten on just one test of sustained attention. These studies sup-
of 22 TBI patients experienced a positive effect on at- port the recommendation of physostigmine at the option
tention as indicated by their ability to continue in reha- level for the treatment of attentional function after TBI.
bilitation. However, no formal cognitive testing was con- Physostigmine is rarely used in clinical practice for en-
ducted. A single case, double-blind, crossover study by hancing cognitive functioning but donepezil and other re-
Bleiberg et al.16 used a computerized battery of tasks of cently developed cholinesterase inhibiting medications
attention and working memory. The effect of dextroam- are commonly used for memory, attention, and behav-
phetamine (5 mg) was determined by comparing perfor- ioral dysfunction for patients with dementia. No other
mance on subtests of the battery to performance follow- studies were available using these medications in TBI.
ing administration of lorazepam (0.5 mg) or placebo
given randomly on consecutive days. Performance was Dopamine enhancers. Amantadine: Three class III re-
less variable on three of the five subtests after adminis- ports examine the use of amantadine for attention/con-
tration of dextroamphetamine. Lorazepam caused more centration in moderate to severe TBI with two of the three
variable performance on one of the subtests. There were reports demonstrating improvement on testing. One six-
no reported side effects of dextroamphetamine. Although case and one single-case study6,7 (class III) revealed ev-
the effect in this study was significant, the generalizabil- idence for improvement in sustained attention, initiation,
ity of this finding is limited because this is a single case and mental flexibility using amantadine (200–400
and the measures are experimental. Dextroamphetamine mg/day). In the single case study further improvement in
is recommended as an option for the treatment of atten- these cognitive domains as well as constructional praxis
tional function after TBI. was observed with the addition of L-dopa/carbidopa
(30/300 mg/day). One class III study by Schneider et al.8
Cholinesterase inhibitors. Donepezil: One small class reported no positive effects for the use of amantadine
I study,17 a double-blind, placebo-controlled two-group (100–300 mg/day) in seven TBI subjects. Overall, these
cross-over design (n 18) used moderate to severe TBI findings studies support the recommendation of amanta-
patients 2–24 months post-injury with attention and/or dine at the option level for the treatment of attentional
memory impairment. Immediate verbal memory was as- function after TBI.
sessed with the Wechsler Memory Scale Third Edition
(WMS-III) Auditory Immediate Index, including imme-
diate recall of stories and pairs of associated words. Im- C. Deficits in Memory
mediate visual memory was assessed with the WMS-III Cholinesterase inhibitors. Donepezil: One small class
Visual Immediate Index including recall of faces and de- I study,17 described in detail above (Deficits in Attention
tails of a complex picture. The Paced Auditory Serial Ad- and Speed of Processing) examined the use of donepezil
dition Test (PASAT) was also administered and this test for memory deficits. The results demonstrated improve-
1488
ment in both memory and attention scores with donepezil. memory, although the only significant finding was for
Moreover, improved test scores were sustained in Group spatial recognition memory. Although this study was de-
A in the placebo phase. Two class III studies20,21 con- signed as a randomized controlled trial, the small num-
sisted of open-label donepezil case series at least one year ber of subjects and the multiple statistical comparisons
post severe TBI. Masanic et al.21 had four subjects, and limit the strength of this recommendation to the option
Taverni et al.20 had one out of two subjects with objec- level.
tive data. Both studies suggested improvement in verbal
and visual-spatial memory with donepezil. These studies Vasopressin. Two class III studies examined the use
support the recommendation of donepezil for the treat- of vasopressin to improve learning and memory after
ment of memory deficits after TBI at the guideline level. TBI. No effect on visual or verbal memory in a double
Physostigmine: Two class-II studies support this rec- blind crossover study of vasopressin (16 IU/day) con-
ommendation at the option level. One study18 was a dou- ducted in 6 patients with severe TBI.25 Another class III
ble blind, placebo-controlled cross-over design of 36 pa- study of vasopressin (8 IU/day) in 26 patients with se-
tients (ages 19–51) with severe TBI, at least three months vere TBI found improved verbal and non-verbal mem-
post-injury with significant, on-going memory impair- ory.27 However, this study had no control condition and
ment. Patients received oral physostigmine, transdermal patients were actively undergoing rehabilitation concur-
scopolamine, or placebo in different sequences in four rent with the study. There is not enough evidence at this
treatment conditions. Scopolamine was evaluated be- time to support a recommendation for vasopressin in the
cause of its potential detrimental affects on memory, as treatment of memory deficits after TBI.
a possible predictor of responsiveness to ACh inhibitors.
Findings demonstrated improvement (50% or greater in- D. Deficits in Executive Functions.
crease) in verbal long-term storage and retrieval for 44%
of patients. Limitations to the interpretation of this study Dopamine enhancers. Bromocriptine: One class I
include potential practice effects for the neuropsycho- study26 found improvement on measures of executive
logical measures and the lack of data and statistical functioning (e.g., tasks involving initiation, mental flex-
analyses available for some measures. The other study19 ibility) in 24 severe TBI patients. Specifically, this study
was a double-blind placebo-controlled study of oral used a double blind, placebo-controlled crossover trial,
physostigmine in combination with lecithin in a small counterbalanced for order. Improvement on executive
group of patients. While multiple measures of attention function tasks thought to tap prefrontal functioning (Dual
and memory were used, there was a statistical trend for Task: counting, Dual Task: digit span, Trail Making Test,
patients on physostigmine to show improvement on just was observed with the with administration of bromocrip-
one test of sustained attention. As mentioned above in at- tine; significant drug by test interaction; positive effect
tentional processes, physostigmine is rarely used in clin- for Dual Task-counting p.028; Dual Task-Digit span,
ical practice for enhancing cognitive functioning, but p.016; Trail Making Test, Stroop Test, Wisconsin Card
donepezil and other recently developed medications in Sorting Test, Controlled Oral Word Association Test).
this class are commonly used for memory, attention, and No improvement was seen, however, for measures of
behavioral dysfunction for patients with dementia. No working memory that did not involve executive control
other studies were available using other cholinesterase in- processes.
hibitors in patients with TBI.
III. Consideration of Potential Adverse
Stimulants. Methylphenidate (MP): One class II
Side Effects
study14 found improved verbal learning and memory on The following section provides a summary of the po-
MP. One class II study13 failed to show a significant ef- tential side effects from the above listed medications in
fect of the study drug, however, statistical power was low TBI patients as reported in the literature:
for this study. Two class III studies22,23 also showed no
significant effect of the study drug. Amantadine: In one study,8 one subject experienced a
side effect of light-headedness that resolved with de-
Other agents. CDP Choline: One class II study24 sup- creased dosage.
ported the use of CDP choline at the option level. The Homeopathy: Side effects, including nausea, vomiting,
study was a double blind, placebo-controlled study of 14 dizziness, fever, depression, and temporary increases
patients with mild-to-moderate TBI in the subacute level in cognitive complaints, were seen in 10% of the treat-
of recovery after post traumatic amnesia (PTA) cleared. ment group but none of those on placebo in the Chap-
Results showed improvement in verbal and visual-spatial man et al.9 study.
1489
WARDEN ET AL.
Stimulants: In terms of side effects with methyl- ious cognitive domains, investigators should typically use
phenidate (MP), studies that reported side effects found several measures of a cognitive domain in a given study, at
no significant differences between drug and placebo. least in early phases of research. To enhance comparability
However, the Kaelin et al study5 did drop one of 11 across studies, efforts to develop consensus measures for
study subjects due to drug-related tachycardia, sug- treatment studies in specific domains should be encouraged
gesting, as expected, that this is a potential side effect by professional organizations and by federal research fund-
of treatment with MP. In the class I study by Whyte ing agencies. In some cases, research support to develop op-
and colleagues,28 reduced appetite was the only sub- timal outcome measures in a given domain will be required.
jective side effect of methylphenidate that approached With respect to specific drugs to study, the current
statistical significance. There was a small but signifi- guideline and option level drugs (e.g., methylphenidate,
cant increase in pulse and blood pressure on active drug, amantadine) need to be examined with more methodolog-
which, in a few subjects, was clinically significant. ically rigorous studies so that they may be able to reach
Cholinesterase inhibitors: Side effects for this general the standard level. Methylphenidate has the most evidence
class of medication included nausea, vomiting, diar- with respect to the treatment of attentional disturbances af-
rhea, and insomnia in a small number of patients. ter TBI warranting a guideline recommendation. Clearly,
Dopamine agonists: Common side effects for this gen- however, large scale randomized control trials are needed
eral class of medication include nausea, dizziness, and to see if it can reach the standard level. Other psychos-
insomnia. timulants such as dextroamphetamine and the new atom-
Anticonvulsants: As noted earlier, a class I study by Dik- oxetine also need further study. Amantadine has limited
men and colleagues1 reported that phenytoin produces evidence regarding the treatment of attentional problems
significant impairment in cognitive functions in pa- and general cognitive functioning and further study is
tients acutely (one month post injury) with severe TBI. needed to see if it remains a viable option in the future.
CDP choline: Gastrointestinal distress was reported as a Cholinesterase inhibitors first found to be useful in the
side effect for CDP choline.24 treatment of dementia such as donepezil and physostig-
mine have some evidence to suggest usefulness in the
IV. Conclusions treatment of memory and attention deficits after TBI but
again larger scale randomized control trials are needed.
While many studies have been conducted to examine
Other cholinesterase inhibitors (e.g., rivastigmine, galan-
the effects of pharmacotherapy on cognitive disorders
tamine), as well as other medication classes found use-
following TBI, the majority of the available evidence was
ful in the treatment of cognitive dysfunction in the de-
limited by the design of the study or its execution. There
mentias (e.g., memantine) should also be examined to see
is insufficient evidence to support the development of any
if there is utility in the TBI population. CDP choline and
standards for the treatment of cognitive disorders after
methylphenidate also have limited evidence for memory
TBI. Evidence does support guidelines for the use of
disorders and need further study.
methylphenidate and donepezil for deficits of attention
Bromocriptine is the only medication found to have any
and speed of processing, the use of donepezil for mem-
evidence in the treatment of executive dysfunction (initi-
ory deficits, and bromocriptine for deficits in executive
ation, mental flexibility). Executive dysfunction can have
functioning after TBI. At the option level, methyl-
significant morbidity with respect to social and vocational
phenidate and amantadine are recommended for deficits
functioning and well-controlled studies examining possi-
of general cognitive function; dextroamphetamine, aman-
ble pharmacologic intervention are much needed.
tadine and physostigmine are recommended for deficits
of attention and speed of processing; methylphenidate
and CDP choline are recommended for memory deficits.
More well-designed and executed randomized controlled REFERENCES
trials are needed to develop treatment standards for cog-
nitive disorders in individuals with TBI. 1. Dikmen SS, Temkin NR, Miller B, et al. Neurobehavioral
effects of phenytoin prophylaxis of posttraumatic seizures.
Journal of the American Medical Association. 1991;
V. Recommendations for Future Research 265(10):1271–1277.
In order to achieve the level of rigor required to develop 2. Dikmen SS, Machamer JE, Winn HR, et al. Neuropsycho-
treatment standards for cognitive disorders after TBI, large logical effects of valproate in traumatic brain injury: a ran-
randomized controlled studies (parallel and cross-over de- domized trial. Neurology. 2000;54:895–902.
signs) are needed. Moreover, given the remaining contro- 3. Azouvi P, Jokic C, Attal N, et al. Carbamazepine in agita-
versy about the definitions and appropriate measures of var- tion and aggressive behaviour following severe closed-head
1490
injury: results of an open trial. Brain Injury. 1999;13(10): memory and sustained attention after traumatic brain in-
797–804. jury. Arch Phys Med Rehabil. Jul 2004;85(7):1050–1055.
4. Plenger PM, Dixon CE, Castillo RM, et al. Subacute 18. Cardenas DD, McLean A, Farrell-Roberts L, et al. Oral
methylphenidate treatment for moderate to moderately se- physostigmine and impaired memory in adults with brain
vere traumatic brain injury: a preliminary double-blind injury. Brain Injury. 1994;8(7):579–587.
placebo-controlled study. Archives of Physical Medicine 19. Levin HS, Peters BH, Kalisky Z, et al. Effects of oral
and Rehabiliation. 1996;77(536–540). physostigmine and lecithin on memory and attention in
5. Kaelin DL, Cifu DX, Matthies B. Methylphenidate effect on closed head-injured patients. Central Nervous System
attention deficit in the acutely brain-injured adult. Archives Trauma. 1986;3(4):333–342.
of Physical Medicine and Rehabiliation. 1996;77:6–9. 20. Taverni JP, Seliger G, Lichtman SW. Donepezil mediated
6. Kraus MF, Maki P. Effect of amantadine hydrochloride on memory improvement in traumatic brain injury during post
symptoms of frontal lobe dysfunction in brain injury: case acute rehabilitation. Brain Injury. 1998;12(1):77–80.
studies and review. Journal of Neuropsychiatry and Clin- 21. Masanic CA, Bayley MT, van Reekum R, et al. Open-la-
ical Neurosciences. 1997;9:222–230. bel study of donepezil in traumatic brain injury. Archives
7. Kraus MF, Maki P. The combined use of amantadine and of Physical Medicine and Rehabiliation. 2001;82:896–901.
l-dopa/carbidopa in the treatment of chronic brain injury. 22. Mooney GF, Haas LJ. Effect of methylphenidate on brain-
Brain Injury. 1997;11(6):445–460. injury related anger. Archives of Physical Medicine and Re-
8. Schneider WN, Drew-Cates J, Wong TM, et al. Cognitive habiliation. 1993;74:153–160.
and behavioural efficacy of amantadine in acute traumatic 23. Tiberti C, Sabe L, Jason L, et al. A randomized, double-
brain injury: an initial double-blind placebo-controlled blind, placebo-controlled study of methylphenidate in pa-
study. Brain Injury. 1999;13(11):863–872. tients with organic amnesia. European Journal of Neurol-
9. Chapman EH, Weintraub RJ, Milburn MA, et al. Homeo- ogy. 1998;5(3):297–299.
pathic treatment of mild traumatic brain injury: a random- 24. Levin HS. Treatment of postconcussional symptoms with
ized, double-blind, placebo-controlled clinical trial. Jour- CDP-choline. Journal of Neurological Services. 1991;
nal of Head Trauma Rehabilitation. 1999;14(6):521–542. 103:S39–S42.
10. McLean A, Cardenas DD, Burgess D, et al. Placebo-con- 25. Fewtrell WD, House AO, Jamie PF, et al. Effects of vaso-
trolled study of pramiracetam in young males with mem- pressin on memory and new learning in a brain-injured pop-
ory and cognitive problems resulting from head injury and ulation. Psychological Medicine. 1982;12:423–425.
anoxia. Brain Injury. 1991;5(4):375–380. 26. McDowell S, Whyte J, D’Esposito M. Differential effect
11. Whyte J, Hart T, Schuster K, et al. Effects of methyl- of a dopaminergic agonist on prefrontal function in trau-
phenidate on attentional function after traumatic brain in- matic brain injury patients. Brain. 1998;121:1155–1164.
jury: a randomized, placebo-controlled trial. American 27. Eames P, Wood RL. Lysine vasopressin in post-traumatic
Journal of Physical Medicine and Rehabilitation. 1997; memory disorders: an uncontrolled pilot study. Brain In-
76:440–450. jury. 1999;13(4):255–260.
12. Whyte J, Hart T, Vaccaro M, et al. Effects of methyl- 28. Alban JP, Hopson MM, Ly V, Whyte J. Effect of
phenidate on attention deficits after traumatic brain injury: methylphenidate on vital signs and adverse effects in adults
a multidimensional, randomized, controlled trial. Am J with traumatic brain injury. American Journal of Physical
Phys Med Rehabil. Jun 2004;83(6):401–420. Medicine and Rehabilitation 2004;83:131–137.
13. Speech TJ, Rao SM, Osmon DC, et al. A double-blind con-
trolled study of methylphenidate treatment in closed head
injury. Brain Injury. 1993;7(4):333–338. AGGRESSION
14. Gualtieri CT, Evans RW. Stimulant treatment for the neu-
robehavioural sequelae of traumatic brain injury. Brain In-
I. Recommendations
jury. 1988;2(4):273–290. STANDARDS. There is insufficient evidence to sup-
15. Hornstein A, Lennihan L, Seliger G, et al. Amphetamine in port the development of standards in the treatment of ag-
recovery from brain injury. Brain Injury. 1996;10(2):145–148. gression after TBI.
16. Bleiberg J, Garmoe W, Cederquist J, et al. Effects of GUIDELINES
dexedrine on performance consistency following brain in- Beta blockers: Beta blockers are recommended as a
jury: a double-blind placebo crossover case study. Neu- guideline for the treatment of aggression after TBI. Stud-
ropsychiatry, Neuropsychology, and Behavioral Neurol- ies reported the efficacy of both propranolol (maximum
ogy. 1993;6(4):245–248. dose 420–520 mg/day) and pindolol (maximum dose
17. Zhang L, Plotkin RC, Wang G, et al. Cholinergic aug- 40–100 mg/day) in the treatment of aggression in this
mentation with donepezil enhances recovery in short-term population.
1491
WARDEN ET AL.
OPTIONS as a component of a multi-drug regimen. The majority of

Methylphenidate: Methylphenidate (dose) is recom- patients reported showed good response to treatment.
mended as an option for the treatment of aggression. Al- However, it should be noted that several patients had to
though evidence was mixed, the study with the greatest be discontinued secondary to side effects.
number of participants showed a positive effect. There is
clear evidence that methylphenidate may be safely used There is insufficient evidence in TBI populations to sup-
without concern of adverse effects on cognition. How- port or refute the use of other commonly used medica-
ever, it should be noted that one case report reported in- tions for aggression. However, evidence of efficacy in
creased agitation with methylphenidate. other patient populations is also a useful source of treat-
Cranial Electrical Stimulation (CES): CES is rec- ment options in TBI in many cases.
ommended at the option level for the treatment of ag-
gression following TBI. Although the supporting class II II. Search Results
study was well constructed, there were no additional sup-
The review of 52 papers resulted in 33 studies that
porting studies to support the recommendation of CES at
could be considered for evidence in the treatment of ag-
the guideline level.
gression after TBI. Nine articles were determined to have
Homeopathy: Homeopathic therapy is recom-
the potential for class I or II evidence (i.e., a randomized
mended at the option level for the treatment of self-re-
controlled trial or a study that utilized a comparison
ported irritability and anger following mild TBI. This rec-
group). No studies were ultimately classified as class I
ommendation is based on a single randomized controlled
evidence due to study design flaws. Six studies were clas-
trial. Although were no other studies to offer supporting
sified as class II evidence. Three of the initial nine were
evidence, the strength of the design of this single study
classified as class III because though unusable as ran-
merits its consideration at the option level.
domized trials, they could be considered as case series.
Serotonin Reuptake Inhibitors: SSRIs are recom-
Other potential class I or II articles were often consid-
mended at the option level for the treatment of aggres-
ered unusable because the treatment group was not lim-
sion following TBI. Specifically, sertraline (25–200
ited to individuals with traumatic brain injuries, and those
mg/day) and paroxetine (20 mg/day) have been reported
TBI patients could not be analyzed separately from avail-
to be effective for the treatment of aggression in this pop-
able data. Evidence is provided in Table 4 by drug cate-
ulation.
gory for the treatment of aggression after TBI.
Valproate: The use of valproate (750–2250 mg/day
to reach therapeutic serum level) is recommended at the
Beta Blockers
option level based on two case reports and one case se-
ries describing marked improvement in aggressive or as- Studies have been conducted over the past 20 years re-
saultive behavior. garding the use of beta blockers in the treatment of ag-
Lithium: The use of lithium is recommended at the gression after a number of neurological disorders, in-
option level for the treatment of aggression after TBI. Be- cluding traumatic brain injury. While several of the
havioral response was achieved at therapeutic levels rang- articles reviewed were class I by design, significant flaws
ing from 0.4 to 1.4 mEq/L. Although the majority of pa- limited their contributions to evidence to class II at best.
tients reported showed a positive response to treatment
with lithium, it should be noted that one patient showed Class II. Brooke1 performed a randomized placebo-
no response and two patients experienced increased irri- controlled 7-week clinical trial of propranolol in 21 in-
tability/agitation. Neurotoxicity and increased EEG spik- dividuals with severe TBI and agitation (episodic motor
ing have also been reported. Thus, lithium should be used or verbal behavior that interfered with patient care, ther-
only with careful monitoring of cognitive status. apy, or safety) rated on the Overt Aggression Scale
Tricyclic Antidepressants: The use of the tricyclic (OAS). Propranolol LA (n 11) or placebo (n 10) was
antidepressants is recommended as an option for the treat- given after one-week baseline measurement, beginning
ment of aggression after TBI. Specifically, amitriptyline with 60 mg/day, and increased by 60 mg/day every third
and desipramine (both up to 150 mg/day) have been re- day to a maximum of 420 mg unless agitation ceased or
ported to be effective for the treatment of aggression in side effects occurred. After three weeks the study drug
this population. was tapered over the course of two weeks. Patients ex-
Buspirone: The use of buspirone (10–60 mg/day) is perienced a significant reduction in intensity of the most
recommended as an option for the treatment of aggres- severe episode per week, but no significant change in fre-
sion after TBI. There are several case series and case re- quency of episodes. Limitations of the study include loss
ports of the use of buspirone as single agent therapy and of subjects to follow-up and small sample size.
1492
TABLE 4. AGGRESSION EVIDENCE
Data
Beta blockers
Brooke, Randomized placebo controlled trial of II Patients experienced a significant reduction in
19921 propranolol (up to 420 mg/day) in 21 intensity of the most severe episode per week
individuals with severe TBI and (p 0.05), but no significant change in
agitation rated on the OAS frequency of episodes.
Greendyke, Double blind, placebo controlled III There were significantly fewer assaults and
19862 crossover study of propranolol attempted assaults during propranolol treatment
(520 mg/day) for violent behavior in a (p 0.05). Although the specific efficacy in
mixed population (4/9 patients had TBI). TBI patients alone in this sample is difficult to
determine, the significant overall group
response warrants consideration.
Greendyke, Double blind, placebo controlled III Statistically significant improvement in number
19863 crossover study of pindolol (60– of assaultive episodes, and other aggression
100 mg/day) for violent behavior in a ratings (p 0.05). The need for supplemental
mixed population (5/11 patients had medication was reduced significantly.
TBI). Optimal response was at 40–60 mg/day.
Greendyke, Double blind, placebo controlled III There was a trend toward decreased aggressive
19894 crossover study of pindolol (20 mg) for behavior for the group as a whole, but this
violent behavior in a mixed population did not reach statistical significance. For the
(3/13 patients had TBI). three individuals with TBI, clinical improvement
was rated as “marked,” “moderate,” and
“none.”
Elliot, 19775 Case series of propranolol (40–120 III Propranolol controlled belligerent behavior without
mg/day) in 5 patients for belligerent inducing general sedation. 2 patients
behavior (4 with CHI and 1 with PHI). experienced bradycardia, hypotension and
lightheadedness.
Mansheim, Case study of propranolol (20 mg TID), III Number of outbursts decreased from 2.5 to 1 per
19816 A-B-A design, for treatment of violent month and no suicidality on 3-month trial of
outbursts and suicide attempts 5 years propranolol 20 mg TID; behavior recurred
post TBI. when withdrawn for 2 weeks, and improved
when reinstated
Yudofsky, Case report of propranolol (320 mg/day) III Propranolol 320 mg/day (and chlorpromazine 400
19817 for violent episodes 4–10 times daily mg/day) improved behavior, and patient could
after severe TBI. be managed at home.
Ratey, 19838 Case report of propranolol (300 mg/day) III Propranolol given for tremor caused dramatic
in patient with severe TBI, psychotic improvement in behavior. When dose reduced
depression, suicide attempts and due to bradycardia, episodes recurred.
periodic assaultive episodes and
no response to ntidepressants or
aantipsychotics
Mattes, 19859 Case report of metoprolol (100 mg BID) III Substantial decrease of outbursts when metoprolol
for aggressive outbursts after penetrating was added to carbamazepine 800 mg/d.
TBI.
Methylphenidate
Mooney, Randomized single-blind placebo- II Scores on measures of anger—KAS Belligerance,
199310 controlled 6-week trial of State-Trait Anger Scale (State), and POMS
methylphenidate (30 mg/day) of 38 anger/hostility factor significantly improved
men with moderate to severe TBI. (p 0.01). However, because study entry was
not based on relevant anger symptoms, one
cannot determine whether anger is a
significant clinical problem for this group of
individuals.
(continued)
1493
WARDEN ET AL.
TABLE 4. AGGRESSION EVIDENCE (CONT’D)
Data
Speech, Randomized double-blind, placebo- II While this study was not designed to treat
199311 controlled crossover study of aggression, the belligerence variable on Katz
methylphenidate (0.3 mg/kg BID) in scale showed no significant adverse reaction
12 patients with moderate to severe of methylphenidate.
TBI and cognitive impairment.
Cranial electrical stimulation
Smith, 199412 CES (1.5 mA output, alternating current, II Patients had statistically significant decreases in
pulsing 100 times/sec) was administered Tension/Anxiety and Anger/Hostility, as well as
to 10 chronic severe TBI patients all other subscores on the POMS.
Homeopathy
Chapman Randomized controlled trial of homeopathy II Treated group had a decrease in 2 of 3 items
199913 in 61 outpatients with mild TBI. relating to irritability and aggression on a 34-
item scale. Analysis was conducted on the
symptom scale as a whole, though scores for
individual items were also presented. However,
because study entry was not based on relevant
anger symptoms, patients had a questionable
level of irritable and aggressive symptoms at
study entry.
Serotonin reuptake inhibitors
Fann, 200014 Single-blind study of sertraline (25–150 III During the 8-week study, scores on the Brief
mg) in 16 mild TBI patients with Anger and Aggression Questionnaire dropped
major depression. significantly (p 0.05). Scores of irritability
and loss of temper on the Head Injury
Symptom Checklist also improved during
treatment.
Kant, 199815 Open-label 8-week trial of sertraline III Significant changes in aggression and irritability
(50–200 mg) in 13 mixed severity scores on the OAS-M were seen at both week
TBI patients with complaints of 4 and 8, compared with baseline (p 0.01).
irritability and/or aggression.
Kim, 200116 Case series of 3 older adults with TBI III All 3 patients were said to have experienced
and aggressive behaviors treated with improved mood and a prompt decrease in
sertraline (100–150 mg/day) or aggression
paroxetine (20 mg/day).
Valproate
Wroblewski Case series of 4 patients with TBI and III Patients were reported to show improvement in
199717 aggressive behaviors treated with all behaviors in a dose dependent manner.
valproate (750–2250/day).
Horne, 199518 Case study of divalproex (serum level III Irritability improved and patient could be
50 micrograms/ml) for severe discharged to a residential home.
agitation after TBI
Geracioti, Case study of valproate (750–1000 mg/ III Dramatic report of reduction of symptoms
199419 day) for episodic explosiveness
after TBI.
Lithium
Glenn, 198920 Case series of lithium for aggression III Five patients were reported to have had a
following TBI. Mixed population, 3/10 “dramatic response” to treatment.
patients had TBI (therapeutic levels
ranged from 0.7 to 1.4 mEq/L).
Bellus, 199621 Case report of lithium (900 mg/day) for III Patient experienced decreased numbers of
severe behavior problems after TBI seclusions, and aggressive and self destructive
(therapeutic levels ranged from 0.48 to behaviors
0.78 mEq/L).
1494
TABLE 4. AGGRESSION EVIDENCE (CONT’D)
Data
Haas, 198522 Case report of lithium for irritability and III Lithium was calming, but patient remained calm
aggression after TBI (therapeutic levels after placebo was given.
ranged from 0.4 to 0.8 mEq/L).
Schiff, 198223 Case report of lithium added to III Report described a paradoxical increase of
phenytoin and carbamazepine aggressivity and EEG spiking when lithium was
aggression after TBI. added to regimen.
Tricyclic antidepressants
Jackson, 198924 Open randomized trial of amitriptyline III 67% responded with 50% decrease in number
(n 5) and desipramine (n 10) of agitated episodes over 7 days.
(both up to 150 mg/day) on agitated
behavior in patients with severe
agitation 2–10 months post-TBI.
Jackson, 198525 Case report: amitriptyline (50 mg/day) III Improvement was seen in aggressive behaviors
for aggressive behaviors. after 2 weeks.
Buspirone
Gualtieri Retrospective case series of thirteen III Buspirone therapy resulted in an average decline
199126 TBI patients treated with buspirone of the Neurobehavioral Rating Scale from 42 to
(10–45 mg/day) for agitated behaviors. 22. Patients with a positive response to
buspirone therapy had mild TBI with no severe
motor or cognitive deficits. However, 6 patients
discontinued due to side effects.
Stanlislav, Retrospective chart review series of 8 III Six patients improved, one had equivocal results,
199427 patients with TBI in a rehabilitation and one was clearly worse. Four of eight
facility treated with buspirone patients had at least a 50% reduction in
(10–20 mg) for more than 3 months. behavioral target symptoms. Improved subjects
had higher doses than non-responders.
Holzer, 199828 Case report of buspirone (60 mg/day) III All behaviors were reported to improve with
for assaultive behavior after severe TBI treatment.
Ratey, 199229 Case report of buspirone (10 mg/day) for III Adding buspirone to existing regiment of lithium,
violent outbursts after severe TBI Carbamazepine, and nadolol significantly
decreased masturbatory behavior. The patient no
longer required restraints, and was discharged
to the community
Carbamazepine
Azouvi, 199930 Open trial of carbamazepine (400–800 III Group had significant improvement on measures
mg/day) in 10 patients with aggressive of agitation and disinhibited behavior after
behavior following severe TBI. treatment (p 0.05). Analysis of individual
response showed that five patients had marked
improvement, three had moderate improvement,
and two had no improvement at all.
Estrogen
Arnold, 199331 Case report of estrogen (1.25 mg/day) for III Patient experienced a “dramatic response” with
refractory aggression after TBI in a sustained improvement after 4 months.
male patient.
Amantadine
Chandler, Case series of amantadine (up to 400 III Patients experienced a decrease in the frequency
198832 mg/day) for refractory aggressive of aggressive behaviors.
behavior after TBI (n 2).
Pyritinol
Kitamura, Placebo-controlled cohort study of II Patients treated with pyritinol reported improvements
198133 pyritinol (600 mg/day) in 270 patients in several symptoms including irritabiity on a
with head injury or sequelae following global self-rating scale. Limitations include
neurosurgery. mixed patient population, lack of detail in
reported study design and lack of statistical rigor.
1495
WARDEN ET AL.
Class III. Three studies conducted by Greendyke and dolol 20 mg bid or placebo. At the end of 10 weeks, there
colleagues were class I in design, but were downgraded was a six-day crossover period. There was a trend toward
by reviewers to class III evidence due to significant flaws. decreased aggressive behavior for the group as a whole,
Greendyke2 conducted a double-blind, placebo-con- but this did not reach statistical significance. For the three
trolled crossover study of propranolol for violent behav- individuals with TBI, clinical improvement was rated as
ior in a mixed VA hospital inpatient psychiatric popula- “marked,” for a 38-year-old man (also on thioridazine
tion. Of the 9 individuals in the study, four had traumatic 200 mg and imipramine 75 mg), “moderate,” for a 53-
brain injury. Subjects had “lack of impulse control and a year-old man (also on thioridazine 150 mg and
high frequency of violent behavior.” Patients were ran- imipramine 100 mg), and “none” for a 72-year-old man
domly assigned to receive propranolol LA (starting at 80 who had an old CVA (no other medications). Overlap
mg/day with increases of 80 mg every 3–4 days to a to- with patients enrolled in Greendyke’s other two studies
tal of 520 mg) or placebo for 11 weeks. After this time, is possible, but not reported.
drug was tapered over 3 weeks and crossover to the other Supporting case reports and case series for propranolol
study agent. There were significantly fewer assaults and included Elliott5 (seven patients), Mansheim6 (one pa-
attempted assaults during propranolol treatment. Seven tient), Yudofsky et al.7 (one patient), and Ratey8 (one pa-
patients had episodes of hypotension or bradycardia, tient). There was one case report of successful treatment
which was not clinically evident. Although the specific with metoprolol9 (one patient). Three case reports pri-
efficacy in TBI patients alone in this sample is difficult marily focused on other drug treatments noted a lack of
to determine, the significant overall group response war- response to beta blockers: Haas and Cope22 (one patient),
rants consideration as class III evidence. Arnold31 (one patient), and Holzer28 (one patient). How-
Later, Greendyke and Kanter3 conducted a double- ever, the preponderance of evidence supports a guideline
blind, placebo controlled crossover study of pindolol for for the use of beta blockers in the treatment of aggres-
violent behavior in a mixed VA hospital inpatient psy- sion after TBI.
chiatric population. Five of the 11 individuals in the study
had traumatic brain injury. Subjects were “pathologically
impulsive, assaultive, openly hostile, and generally un- Methylphenidate
cooperative. All were severely demented.” Patients were Class II. Mooney10 conducted a randomized single-
randomly assigned to receive pindolol (starting at 10 blind placebo-controlled 6-week trial of methylphenidate
mg/day with increases of 10 mg/day every 3–4 days up in 38 men who had sustained TBI. All patients experi-
to a total daily dose of 60 mg) or placebo for 10 days. enced a TBI with LOC of 6 h or greater, or PTA of 24
Further dose increases were permitted up to 100 mg/day. h or longer. Subjects were randomly assigned to receive
After this time, drug was tapered and patient crossover either methylphenidate or placebo (single blind). Med-
to the other study drug. It was not clear how long treat- ication was gradually increased over the first 4 weeks of
ment was for each phase of the study. All regular treat- the study, and remained at the final dosage (methyl-
ment with psychotropic medications was discontinued 2 phenidate 30 mg/day) for the final two weeks. Scores on
weeks before initiation of the study. measures of anger-KAS, State-Trait Anger Scale, and
There was a statistically significant improvement in POMS anger/hostility factor significantly improved. Sev-
number of assaultive episodes, and ratings of hostility, eral weaknesses temper the results of this trial. The au-
uncommunicativeness, uncooperativeness, and repetitive thors did not provide data on each treatment group with
behaviors. The need for supplemental medication was re- reference to confounders and or any details on random-
duced significantly. Optimal response was 40–60 ization methods. There was no a priori definition of a re-
mg/day. On higher dosages, some patients appeared to sponder. Thus, it is not clear what the clinical meaning
be over stimulated. While individual response of the five is for these changes in rating scale scores. Because study
patients with TBI was not given, the significant overall entry was not based on relevant anger symptoms, one
group response warrants consideration as class III evi- cannot determine whether anger is a significant clinical
dence. problem for this group of individuals. Individuals who
Also, Greendyke4 conducted a double-blind, placebo- had greater pretreatment anger appeared to have greater
controlled crossover study of pindolol for violent behav- improvement. Therefore, one explanation of the results
ior in a mixed VA hospital inpatient psychiatric popula- is regression of these anger scores toward the mean. Due
tion. Three of the 13 individuals in the study had to these significant flaws, this class I design study was
traumatic brain injury. Subjects had behavioral or man- downgraded to class II evidence. It is important to note
agement problems. The OAS was used to monitor phys- that patients did not exhibit increased aggression on
ical and aggressive acts. Individuals received either pin- methylphenidate.
1496
Speech11 conducted a randomized double-blind, measure of post concussive symptoms. Fifty subjects of
placebo-controlled crossover design study in 12 patients the 61 originally randomized were available for analysis
with closed TBI with cognitive impairment. Patients were (27 in active treatment and 23 in placebo). The persis-
14–108 months post-injury. They were required to have tence of post concussive symptoms is implied in the
LOC or PTA of at least 1 day (average 14.4 days). Pa- methodology. The treated group reported a decrease in
tients received either 0.3 mg/kg methylphenidate BID for two of three items relating to irritability and aggression
1 week, followed by 1 week of placebo, or the opposite on a 34-item scale. Statistical analysis was conducted on
regimen. Behavior was assessed with the Katz Adjust- the symptom scale as a whole, though scores for indi-
ment Scale, as well as several cognitive tasks. While this vidual items were also presented. Weaknesses of the
study was not designed to treat aggression, the belliger- study include a loss to follow-up of 18% of patients and
ence variable on Katz scale showed no significant ad- questionable level of irritable and aggressive symptoms
verse reaction of methylphenidate. at study entry. This study supports the recommendation
of homeopathy as an option for the treatment of aggres-
Class III. One case report of lithium for irritability and sion after TBI.
aggression after TBI noted increased agitation when the
patient was treated with methylphenidate.22 However, the
Serotonin Reuptake Inhibitors
majority of the evidence supports an option for the use
of methylphenidate in the treatment of aggression after Class III. Three open case series examined sertraline
TBI. in the treatment of post-TBI aggression. Fann et al.14 con-
ducted a trial of sertraline in 16 patients with mild TBI
Cranial Electrical Stimulation who met criteria for major depression (DSM-III-R) and
had at least 18 on the HAM-D. During this 8-week sin-
Class II. Smith and colleagues12 administered cranial
gle blind study of sertraline, scores on the Brief Anger
electrotherapy stimulation (CES) to 10 of 21 chronic se-
and Aggression Questionnaire dropped significantly
vere TBI patients in a sheltered living facility. CES was
(from 9.3 at baseline to 6.5 post-treatment), as did the
delivered with a maximum of 1.5 mA output, alternating
HAM-D scores. Scores of irritability and loss of temper
current, pulsing 100 times per second. CES or sham treat-
on the Head Injury Symptom Checklist also improved
ment was delivered for 45 min, 4 consecutive days/week
during treatment. The final dose of sertraline ranged from
for 3 weeks. The POMS was administered at baseline and
25 to 150 mg. Limitations of this study include the lack
after the 12 sessions to CES (n 10), sham controls (n
of ability to control for natural improvement, and re-
5) and ordinary controls (n 6) with no access to the de-
gression to the mean. Although the clinical significance
vice administering sham or active intervention. Forty-
of the statistical improvement of the aggression scale im-
three percent of patients had received chemical depen-
provement is not clear, the PCS scale clearly showed pa-
dency treatment prior to their injury and 86% of patients
tients reporting improvement in irritability and anger out-
had a seizure disorder; all patients were on 3–5 medica-
bursts compared with baseline levels.
tions. Although details of the sham treatment were pro-
Kant et al.15 reported an 8-week open label trial of ser-
vided, the reviewers were not clear how patients could
traline in 13 patients with mixed severity TBI presenting
be unaware of receiving above threshold current to their
to a neuropsychiatry clinic with complaints of irritability
scalp. Still, the report was suggestive for the statistically
and/or aggression. Dosage ranged from 50 to 200 mg.
significant decreases in Tension/Anxiety and Anger/
Ten patients completed the trial. Significant changes in
Hostility, as well as all other subscores on the POMS.
aggression and irritability scores on the OAS-M were
Although this study was well constructed, there were no
seen at both week 4 and 8, compared with baseline. This
additional supporting studies to support a higher recom-
study was limited by lack of a control group or blind
mendation. Therefore, CES is recommended as an option
ratings.
for the treatment of aggression after TBI.
Kim et al.16 reported three older adults (45 years)
who had sustained TBI in their 20’s. Each had a remote
Homeopathy
or recent history of alcohol abuse and decreased cogni-
Class II. Homeopathy uses minute amounts of sub- tive abilities. Each received sertraline for aggressive be-
stances derived from plant, animal, or mineral sources, haviors and was described to experience improved mood
which are produced by serially agitated dilution (SAD) and a prompt decrease in aggression.
with water or pharmaceutical alcohol. Chapman et al.13 These studies support the recommendation of serotonin
studied outpatient volunteers with mild TBI who were reuptake inhibitors at the option level for the treatment
followed for 4 months in a RCT study and rated on a new of aggression after TBI.
1497
WARDEN ET AL.
Valproate Buspirone
Class III. Two case reports and one case series sup- Class III. Two case reports and two case series exam-
port the recommendation of valproate as an option for ined the use of buspirone for post-TBI aggression.
this indication. Wroblewski34 reported a series of four Gualtieri26 reported a retrospective case series of 13 TBI
patients who showed improvement in all behaviors in a patients treated with buspirone for agitated behaviors. Six
dose-dependent manner. Horne18 described one patient patients had therapy discontinued for side effects, in-
whose irritability improved and could be discharged to a cluding headache (2), lightheadedness (2), rash (1), or
residential home. Geracioti19 also reported dramatic non-response (1). Patients with a positive response to
symptom reduction in one patient. buspirone therapy had mild TBI with no severe motor or
cognitive deficits. Three of the responders had dyspho-
Lithium ria and restlessness consistent with post traumatic
Class III. There have been several case series/case re- akathisia and four had agitation, irritability or angry out-
ports on the use of lithium. Supporting reports include bursts, all with temporal lobe lesions. Buspirone (10–45
Glenn et al.20 (mixed population, 10 patients, five pa- mg/day) therapy resulted in an average decline of the
tients had a “dramatic response,” three of these individ- Neurobehavioral Rating Scale from 42 to 22.
uals had a traumatic brain injury); Bellus et al.21 (one Stanislav27 described a series of nine patients with TBI
case); and Haas22 (one case). The dramatic improvement in a rehabilitation facility. Retrospective pharmacy and
in aggressive episodes described by all of these reports chart review over a 3-year period was undertaken to iden-
supports a recommendation at the option level. It should tify individuals receiving buspirone therapy. Outcome
be noted, however, that one case report described a para- was documented by quantified and qualified aggressive
doxical increase of aggressivity and EEG spiking on events by behavioral therapists. Of the patients with
lithium when added to phenytoin and carbamazepine.23 greater than three months of buspirone therapy (10–20
Thus, while lithium is recommended at the option level, mg tid), eight had TBI. Of these, six improved, one had
cognitive status should be carefully monitored. equivocal results and one was clearly worse. Four of eight
patients had at least a 50% reduction in behavioral tar-
Tricyclic Antidepressants (Amitriptyline get symptoms. Improved subjects had higher doses than
and Desipramine) non-responders. Of the four patients receiving buspirone
on admission, one had TBI. This patient failed a 6-week
Class III. Jackson et al.24 investigated the effect of trial of buspirone in combination with carbamazepine and
TCAs on agitated behavior and cortisol secretion and sup- haloperidol resulting in its discontinuation.
pression in a study of 35 patients with severe TBI un- Supporting case reports of buspirone in the treatment
dergoing rehabilitation. Patients exhibiting severe per- of aggression after TBI include Holzer28 and Ratey et
sistent agitation at 2–10 months post-TBI (n 15) were al.29 These class III reports support the recommendation
openly randomized to amitriptyline (n 5) or de- of buspirone as an option in the treatment of aggression
sipramine (n 10). Doses were less than 150 mg total after TBI.
per day. Sixty seven percent responded with 50% de-
crease in number of agitated episodes over 7 days by
Carbamazepine
prospective nursing records. TCA non-responders had a
shorter duration of coma (4.8 vs. 9.6 days), a longer du- Class III. Azouvi et al.30 reported on a therapeutic trial
ration from injury to rehabilitation admission (104 vs. 50 of carbamazepine in 10 patients with aggressive behav-
days) and a lower percentage of PTA clearing at 81 days ior following severe TBI. The group showed significant
(40% vs. 80%) compared to TCA responders. Of the TCA improvement on measures of agitation and disinhibited
responders, seven were receiving phenytoin and three behavior after treatment. Analysis of individual response
were receiving phenobarbitol. Of the TCA non-respon- showed that five patients had marked improvement, three
ders, five were receiving phenytoin and none were re- had moderate improvement, and two had no improve-
ceiving phenobarbitol. There was essentially no associa- ment at all. While quantitative measures of therapeutic
tion identified between cortisol dynamics and response response were utilized, the open-label design, absence of
to TCAs. Jackson et al25 also reported a single case controls, and variation in concomitant pharmacologic
treated with amitriptyline (50 mg QHS) with improve- therapies limit the conclusions that can be drawn from
ment seen in aggressive behaviors after 2 weeks. These this trial. Indeed, the authors concluded that it “might
studies support the recommendation of tricyclic antide- help,” but that significant inter-individual variability ex-
pressants at the option level for the treatment of aggres- isted. Although carbamazepine seems to be effective for
sion following TBI. some patients experiencing aggression after TBI, the re-
1498
ported literature did not support a recommendation at this spite reports of depression with the use of beta-blockers,
time. controlled trials indicate that it is a rare occurrence.35,36
Estrogen Methylphenidate. Of most importance was that there

was no increase in aggression or irritability.10,11,22
Class III. Arnold31 published a case report on a male
patient with refractory aggression post-TBI who signifi-
Homeopathy. As noted in the cognitive section, side
cantly improved with estrogen therapy. More supporting
effects occurred, including nausea, vomiting, dizziness,
cases are needed to support a recommendation for estro-
fever, depression, and temporary increases in cognitive
gen therapy in the treatment of aggression after TBI.
complaints, were seen in 10% of the treatment group but
none of those on placebo in the Chapman et al.13 study.
Amantadine
Class III. Chandler and colleagues32 reported on two Lithium. Neurotoxicity20 and increased EEG spiking23
patients with refractory aggressive behavior tried on have been reported. Thus, lithium should be used with
amantadine (up to 400 mg/day) who experienced a de- careful monitoring of cognitive status.
crease in the frequency of aggressive behaviors. More
supporting cases are needed to support a recommenda- Buspirone. Reported side effects in one study included
tion for the use of amantadine in the treatment of ag- headache, lightheadedness, or rash.26 Three of the re-
gression after TBI. sponders had dysphoria and restlessness consistent with
post-traumatic akathisia, and four had agitation, irritabil-
Pyritinol ity, or angry outbursts
Class II. One placebo-controlled cohort study exam-
IV. Conclusions
ined the use of pyritinol in 270 patients with head injury
or sequelae following neurosurgery.33 Patients treated Although many publications examine the use of phar-
with pyritinol reported significant improvements in sev- macotherapy for the treatment of aggression after TBI,
eral symptoms including irritability on a global self-rat- the majority of studies suffer significant methodological
ing scale. Limitations of this study included the mixed flaws. To date, there is insufficient evidence to support
patient population, lack of detail in reported study design the development of any standards for the treatment of ag-
and lack of statistical rigor. Although this agent may gression following TBI. Evidence does support guide-
show promise in the treatment of post-TBI aggression, lines for the use of the beta blockers propranolol and pin-
we are unable to recommend medications not FDA ap- dolol for treating aggression in TBI patients. At the option
proved for use in the United States. level, methylphenidate, cranial electrical stimulation,
homeopathy, serotonin reuptake inhibitors, valproate,
III. Consideration of Potential Adverse lithium, tricyclic antidepressants, and buspirone are rec-
Side Effects ommended for the treatment of aggression following TBI.
More well-designed and executed randomized controlled
Not all of the studies and case reports reported adverse trials are needed to develop treatment standards for ag-
events. For many of these medications, which have been gression in individuals with TBI.
studied for other neurobehavioral problems after TBI,
side effects have been reported. The reader is referred to
V. Recommendations for Future Research
those studies noted above. Specifically in the studies of
aggression, valproate, carbamazepine, amantadine, and Primary recommendations for future research on phar-
SSRI’s did not list specific adverse reactions. We assume macologic treatment of aggressive behavior include the use
that the side effect profile for these medications will be and development of defined measures of aggression, con-
similar to those experienced by other patient populations. sistent assessment of aggression severity, reports focused
Specific data was available for several medications: specifically on TBI rather than mixed patient populations,
inclusion of data on comorbid disorders (especially de-
Beta blockers. The major side effects of beta-blockers pression), and the use of placebo-controlled designs.
when used to treat aggression are a lowering of blood Due to the paucity of current studies, there are a num-
pressure and pulse rate. Because peripheral beta recep- ber of medications that can be studied. It will be impor-
tors are fully blocked in doses of 300–400 mg/day, fur- tant to pursue research on medications that appear promis-
ther decreases in these vital signs usually do not occur ing regardless of the level of industry support for individual
even when doses are increased to much higher levels. De- agents.
1499
WARDEN ET AL.
Standardized measures of aggression should be utilized. 5. Elliot FA. Propranolol for the control of belligerent be-
This report was limited by the non-standardized descrip- havior following acute brain damage. Annals of Neurology.
tions of aggression provided in the current literature. Al- 1977;1:489–491.
though theoretical discussions of aggression and irritabil- 6. Mansheim P. Treatment with propranolol of the behavioral
ity define concepts in discrete terms, we found that these sequelae of brain damage. Journal of Clinical Psychiatry.
treatment reports commonly did not define the behaviors 1981;42(3):132.
in these terms. There are spontaneous day-to-day and 7. Yudofsky SC, Williams DT, Gorman J. Propranolol in the
week-to-week fluctuations in aggression that cannot be treatment of rage and violent behavior in patients with
validly interpreted without prospective documentation. In chronic brain syndromes. American Journal of Psychiatry.
addition, aggression like certain mood disorders may have 1981;138(2):218–220.
cyclic exacerbations. The use of an objective measure- 8. Ratey JJ, Morrill R, Oxenkrug G. Use of propranolol for
ment scale such as the Overt Aggression Scale provoked and unprovoked episodes of rage. American
(OAS),37–39 the Overt Agitation and Severity Scale,40 or Journal of Psychiatry. 1983;140(10):1356–1357.
the Agitated Behavior Scale41 would greatly strengthen 9. Mattes JA. Metoprolol for intermittent explosive disorder.
the findings of future studies of aggression after TBI. Fu- American Journal of Psychiatry. 1985;142(9):1108–1109.
ture studies also need to focus specifically on patients with 10. Mooney GF, Haas LJ. Effect of methylphenidate on brain-
TBI. Several studies we reviewed included patients with injury related anger. Archives of Physical Medicine and Re-
varied neuropsychiatric disorders. In order to eliminate habiliation. 1993;74:153–160.
the possibility of differential response by diagnosis, stud- 11. Speech TJ, Rao SM, Osmon DC, et al. A double-blind con-
ies should be conducted exclusively with patients with trolled study of methylphenidate treatment in closed head
TBI. Studies should include measures of co-existing psy- injury. Brain Injury. 1993;7(4):333–338.
chiatric disorders, such as depression, psychosis, anxiety, 12. Smith RB, Tiberi A, Marshall J. The use of cranial elec-
to explore whether these may predict efficacy. There is trotherapy stimulation in the treatment of closed-head-in-
some suggestion that the presence of depression correlates jured patients. Brain Injury. 1994;8(4):357–361.
with aggressive behavior after TBI.42 13. Chapman EH, Weintraub RJ, Milburn MA, et al. Homeo-
Study design must consider the expected change with pathic treatment of mild traumatic brain injury: a random-
medication so that sufficient numbers of patients are in- ized, double-blind, placebo-controlled clinical trial. Jour-
cluded. We believe that adequate statistical power can only nal of Head Trauma Rehabilitation. 1999;14(6):521–542.
be obtained through the use of multi-center studies. Due 14. Fann JR, Uomoto JM, Katon WJ. Sertraline in the treat-
to the spontaneous fluctuations in the number of aggres- ment of major depression following mild traumtic brain in-
sive episodes, and the natural tendency for regression to jury. Journal of Neuropsychiatry and Clinical Neuro-
the mean (where patients are entered at the most severe sciences. 2000;12(2):226–232.
levels of aggression and naturally exhibit fewer episodes 15. Kant R, Smith-Seemiller L, Zeiler D. Treatment of ag-
with time), a placebo control arm must be included. gression and irritability after head injury. Brain Injury.
1998;12(8):661–666.
16. Kim KY, Moles JK, Hawley JM. Selective serotonin re-
REFERENCES uptake inhibitors for aggressive behavior in patients with
dementia after head injury. Pharmacotherapy. 2001;21(4):
1. Brooke MM, Patterson DR, Questad KA, et al. The treat- 498–501.
ment of agitation during initial hospitalization after trau- 17. Wroblewski B, Joseph AB. Tricyclic antidepressant in TBI.
matic brain injury. Archives of Physical Medicine and Re- Archives of Physical Medicine and Rehabilitation.
habilitation. 1992;73:917–921. 1997;78:109.
2. Greendyke RM, Kanter DR, Schuster DB, et al. Propranolol 18. Horne M, Lindley SE. Divalproex sodium in the treatment
treatment of assaultive patients with organic brain disease: a of aggressive behavior and dysphoria in patients with or-
double-blind crossover, placebo-controlled study. Journal of ganic brain syndromes. Journal of Clinical Psychiatry.
Nervous and Mental Disease. 1986;174(5):290–294. 1995;56(9):430–431.
3. Greendyke RM, Kanter DR. Therapeutic effects of pindolol 19. Geracioti TD. Valproic acid treatment of episodic explo-
on behavioral disturbances associated with organic brain siveness related to brain injury. Journal of Clinical Psy-
disease: a double-blind study. Journal of Clinical Psychi- chiatry. 1994;55(9):416–417.
atry. 1986;47(8):423–426. 20. Glenn MB, Wroblewski B, Parziale J, et al. Lithium car-
4. Greendyke RM, Berkner JP, Webster JC, et al. Treatment bonate for aggressive behavior or affective instability in ten
of behavioral problems with pindolol. Psychosomatics. brain-injured patients. American Journal of Physical Med-
1989;30(2):161–165. icine and Rehabilitation. 1989;68(5):221–226.
1500
21. Bellus SB, Stewart D, Vergo JG, et al. The use of lithium 34. Wroblewski BA, Joseph AB, Kupfer J, et al. Effectiveness
in the treatment of aggressive behaviours with two brain- of valproic acid on destructive and aggressive behaviours
injured individuals in a state psychiatric hospital. Brain In- in patients with acquired brain injury. Brain Injury.
jury. 1996;10(11):849–860. 1997;11(1):37–47.
22. Haas JF, Cope DN. Neuropharmacologic management of 35. Ko DT, Hebert PR, Coffey CS, et al. Beta-blocker therapy
behavior sequelae in head injury: a case report. Archives and symptoms of depression, fatigue, and sexual dysfunc-
of Physical Medicine and Rehabilitation. 1985;66: tion. JAMA 17 2002;288(3):351–357.
472–474. 36. Yudofsky SC. Beta-blockers and depression. The clini-
23. Schiff HB, Sabin TD, Geller A, et al. Lithium in aggres- cian’s dilemma. JAMA 1992;267(13):1826–1827.
sive behavior. Am J Psychiatry. Oct 1982;139(10):1346– 37. Silver JM, Yudofsky SC. Documentation of Aggression in
1348. the Assessment of the Violent Patient. Psychiatric Annals.
24. Jackson RD, Mysiw WJ. Abnormal cortisol dynamics af- 1987;17:375–384.
ter traumtic brain injury. American Journal of Physical 38. Silver JM, Yudofsky SC. The Overt Aggression Scale:
Medicine and Rehabilitation. 1989;68(1):18–23. overview and guiding principles. J Neuropsychiatry Clin
25. Jackson RD, Corrigan JD, Arnett JA. Amitriptyline for ag- Neurosci. 1991;3(2):S22–29.
itation in head injury. Archives of Physical Medicine and 39. Yudofsky SC, Silver JM, Jackson W, et al. The Overt Ag-
Rehabilitation. 1985;66:180–181. gression Scale for the objective rating of verbal and phys-
26. Gualtieri CT. Buspirone for the behavior problems of pa- ical aggression. Am J Psychiatry. 1986;143(1):35–39.
tients with organic brain disorders. Journal of Clinical Psy- 40. Yudofsky SC, Kopecky HJ, Kunik M, et al. The Overt Ag-
chopharmacology. 1991;11(4):280–281. itation Severity Scale for the objective rating of agitation.
27. Stanislav SW, Fabre T, Crismon ML, et al. Buspirone’s ef- J Neuropsychiatry Clin Neurosci. 1997;9(4):541–548.
ficacy in organic-induced aggression. Journal of Clinical 41. Bogner JA, Corrigan JD, Stange M, et al. Reliability of the
Psychopharmacology. 1994;14(2):126–130. Agitated Behavior Scale. J Head Trauma Rehabil. 1999;
28. Holzer JC. Buspirone and brain injury. Journal of Neu- 14(1):91–96.
ropsychiatry and Clinical Neurosciences. 1998;10(1): 42. Tateno A, Jorge RE, Robinson RG. Clinical correlates of
113. aggressive behavior after traumatic brain injury. J Neu-
29. Ratey JJ, Leveroni CL, Miller AC, et al. Low-dose bus- ropsychiatry Clin Neurosci. 2003;15(2):155–160.
pirone to treat agitation and maladaptive behavior in brain-
injured patients: two case reports. Journal of Clinical Psy-
chopharmacology. 1992;12(5):362–364. ACKNOWLEDGMENTS
30. Azouvi P, Jokic C, Attal N, et al. Carbamazepine in agita-
tion and aggressive behaviour following severe closed-head Funding for meetings and administrative support by
injury: results of an open trial. Brain Injury. 1999;13(10): National Brain Injury Research Treatment and Training
797–804. Foundation, Centers for Disease Control, and Defense
31. Arnold SE. Estrogen for refractory aggression after trau- and Veterans Brain Injury Center.
matic brain injury. American Journal of Psychiatry. 1993
1993;150(10):1564–1565. Address reprint requests to:
32. Chandler MC, Barnhill JL, Gualtieri CT. Amantadine for Peter Quinn
the agitated head-injury patient. Brain Injury. 1988;2(4): The NeuroTrauma Foundation
309–311. 555 Madison Ave.
33. Kitamura K. Therapeutic effect of pyritinol of sequelae of New York, NY 10022
head injuries. Journal of International Medical Research.
1981;9:215–221. E-mail: [email protected]
1501

Journal Neurotrauma

Uploaded by

Copyright:

Available Formats

Journal Neurotrauma

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Journal Neurotrauma

Uploaded by

Copyright:

Available Formats

JOURNAL OF NEUROTRAUMA

Volume 23, Number 10, 2006

Guidelines for the Pharmacologic Treatment

NEUROBEHAVIORAL GUIDELINES WORKING GROUP MEMBERS

DEBORAH L. WARDEN,1,6,7 BARRY GORDON,2,5,8 THOMAS W. McALLISTER,3,5,9

INTRODUCTION with TBI are more refractory to treatment, more prone to

TABLE 1. CRITERIA FOR DETERMINING THE LEVEL OF RECOMMENDATION

Standards Guidelines Options

II. Cognitive Disorders III. Aggression

TABLE 2. AFFECTIVE DISORDERS, ANXIETY, AND PSYCHOSIS EVIDENCE

TABLE 2. AFFECTIVE DISORDERS, ANXIETY, AND PSYCHOSIS EVIDENCE (CONT’D)

TABLE 3. COGNITIVE DISORDERS EVIDENCE

I. General cognitive function

TABLE 3. COGNITIVE DISORDERS EVIDENCE (CONT’D)

TABLE 3. COGNITIVE DISORDERS EVIDENCE (CONT’D)

TABLE 3. COGNITIVE DISORDERS EVIDENCE (CONT’D)

OPTIONS as a component of a multi-drug regimen. The majority of

TABLE 4. AGGRESSION EVIDENCE

TABLE 4. AGGRESSION EVIDENCE (CONT’D)

TABLE 4. AGGRESSION EVIDENCE (CONT’D)

Estrogen Methylphenidate. Of most importance was that there

You might also like