Pharmacological management of aggression in patients with traumatic brain injury: a systematic review
H. Smith, E. Collison-Ani, A. Isola, G. Burrah
Abstract
Background
Aggression is a common sequela following TBI. Aggressive behaviour includes agitation, irritability, anger, verbal outbursts, physical aggression against objects, self or other persons, and violent, hostile or assaultive behaviour. Prescribing practice usually derives from management in other patient populations and personal experience of the clinician. This review aims to analyse research in pharmacological intervention following TBI. Given the sparsity of available evidence, an expansive definition of aggression is used to review the evidence for all emotional and behavioural issues subsequent to TBI.
Methods
PubMed, ProQuest, PsychInfo, Ovid Medline, Web of Science and Google Scholar were searched for clinical trials of aggression management in TBI. Inclusion criteria was: participants over 16 with confirmed TBI diagnosis; primary outcome assessing aggression using objective rating scale; trial of single medication. Exclusion criteria was: participants 16 or younger; presence of other neurodegenerative disorder, primary psychiatric diagnosis, or other acquired brain injury e.g. stroke.
Results
10 studies were identified. Of 6 medications used in these studies, only amantadine had level II evidence in support of treatment. Amantadine demonstrated Neuropsychiatric Inventory – Irritability scale reduction of 4.3 compared to 2.6 in placebo (p = 0.0085). Later trials failed to confirm this benefit.
Conclusions
Carbamazepine, methylphenidate, quetiapine, risperidone, and sertraline do not have evidenced benefit in post-TBI aggression management, although are well tolerated. Amantadine demonstrates mixed evidence in support. Larger studies with consistent outcome measures and uniform inclusion criteria should be conducted.
Introduction
Traumatic Brain Injury (TBI) is a major cause of health, social and economic problems. TBI disrupts the normal function of the brain, caused by a bump, blow or jolt to the head or a penetrating head injury. Aggression is the most common sequela post-TBI. The spectrum of aggressive behaviour includes agitation and irritability, anger, verbal outbursts, physical aggression against objects, self or other persons, severe irritability, and violent, hostile, or assaultive behaviour. Aggression affects an individual’s access to rehabilitation, employment and community activities. More effective support and treatment will, therefore, help to improve the quality of life for individuals with aggression post-TBI and their families [1,2].
Post-TBI aggression can be managed with pharmacological and non-pharmacological approaches. There are no Food and Drug Administration (FDA) approved medications for aggression related to TBI. Clinicians rely on experience treating similar conditions and extrapolation of management from non-TBI populations. Resultantly, there is wide variation in prescribing practices; equally efficacious medications, variability of clinical presentation, severity of neurobehavioral symptoms, and lack of clinical practice guidelines contribute to this [3]. Antipsychotics remain the most prescribed medication to manage post-TBI aggression. This contrasts with available evidence and recommendations from systematic reviews [4,5,6].
Methods
Ovid Medline, ProQuest, PubMed, Web of Science, PsychINFO, and Google Scholar were searched between January 1975 and November 2024. The following search terms were used:
(pharmaco* OR drug* OR treatment OR therap* OR management OR beta blocker* OR antipsychotic* OR typical antipsychotic* OR atypical antipsychotic* OR anticonvulsant* OR antiepileptic OR dopamine agonist* OR amantadine OR psychostimulant* OR antidepressant* OR sertraline OR trazodone OR TCAs OR alpha agonist* OR hypnotic* OR anxiolytic* OR benzodiazepine* OR lithium OR methylphenidate OR amphetamine OR propranolol OR pindolol OR buspirone OR carbamazepine) AND (irritability OR hostility OR aggress* OR agitation OR challenging behavior OR violen* OR assault*) AND (traumatic brain injur* OR head injur* OR TBI OR brain trauma)
Clinical trials, both controlled and uncontrolled, were eligible. One author (GB) identified which studies met inclusion criteria, screened each report, and conducted the bias assessment using the Cochrane RobB2 tool. Several papers had some concerning features relating to randomisation, missing data, outcome measures and selection bias. Full assessment is given in the appendix. For each paper, a summary of these features is given in the results section. Two authors (GB, HS), independently collected and reviewed data from each report. Where discrepancies in study screening and data extraction were identified, these authors jointly discussed whether the study met predefined inclusion criteria and verified outcomes. No included paper required further verification steps. Outcomes sought were any change in aggression as identified by a validated, standardised outcome measure (Table 2). A broad definition of aggression, inclusive of all behavioural and emotional changes was included in the search strategy to identify all possible reports given the limited data available.
Any trials utilising children under 16 were excluded to remove confounding due to aggression relating to developmental conditions such as ADHD. Any trials where participants had neurodegenerative disorders (e.g. dementia), acquired brain injury (e.g. stroke) or primary psychiatric diagnosis (e.g. schizophrenia) were excluded to remove confounding due to aggression related to these diagnoses. Any trials including participants with a history of behavioural disturbance prior to TBI or those in the acute or posttraumatic amnesic phase were excluded to remove possible confounding with prior data.
Given the limited data available, no data synthesis or meta-analysis was conducted. Results are presented descriptively for each study. The review was not registered in advance with PRISMA.
Results
The PRISMA flow chart (Figure 1) shows the results of the literature search. Ten studies were included in the final review. Seven were RCTs. Three were open trials, one of which was not blinded. A summary of the full-text articles excluded with reasons is given in the appendix. Reasons for exclusion included unspecified time since injury, inconsistent diagnosis, focus on cognitive recovery, and description of case series. Of note, all studies investigating the use of beta blockers were excluded for reasons including participants in the acute-TBI phase, and unspecified outcome measures.
Figure 1: PRISMA summary of identified studies

TBI diagnosis and inclusion criteria
There was significant variability in confirmation of TBI diagnosis (Table 2). For several studies, no thresholds for TBI severity for inclusion were used; therefore, the diagnosis of TBI was equivalent to the inclusion criteria. Some estimation of severity can be made based on the manner in way participants were recruited.
Table 1: Methods of diagnosis and TBI inclusion criteria.
Other inclusion criteria the same as the literature search strategy, such as recruiting only adults, have not been repeated in the table.
|
Study |
Method of TBI diagnosis |
Inclusion criteria relating to TBI severity or chronicity |
|
Azouvi et al, 1999 [7] |
Referral diagnosis of severe closed head injury |
|
|
Deb et al, 2020 [8] |
GCS score < 8 at time of injury |
Minimum 6 months post-injury |
|
Hammond et al, 2015 [9] |
Reported behavioural changes affecting care and/or family reintegration |
Minimum 6 months post-injury |
|
Hammond et al, 2017 [10] |
Mayo Clinic criteria |
Score of > 6 on observer-rated NPI-I Most Problematic questionnaire |
|
Hammond et al, 2014 [11] |
GCS < 13, motor score < 6 |
|
|
Hammond et al, 2021 [12] |
Loss of consciousness |
As above |
|
Kant et al, 1998 [13] |
Post-traumatic amnesia > 24 hours |
Minimum 6 months post-injury |
|
Kim & Bijlani, 2006 [14] |
Positive neuroimaging |
Score of > 2 on NPI-I questionnaire |
|
Mooney & Haas, 1993 [15] |
Focal neurological signs |
As in Hammond et al, 2015 |
Outcome measures
Each paper employed one or more of a battery of available assessments of post-TBI aggression, either specifically or as part of broader mood and behaviour analysis. A summary is given in Table 2. Each yields a quantitative outcome measure.
Table 2: Measures of post-TBI aggression used in the selected papers
|
Outcome measure |
Subtype |
Measure of |
Rated by |
|
ABS (Agitated Behaviour Scale) |
|
Post-traumatic agitation during acute recovery phase |
Observer (objective) |
|
KAS (Katz Adjustment Scale) |
|
Social personality and behavioural characteristics of inpatient psychiatric patients returning to the community |
Observer |
|
NPI (Neuropsychiatry Inventory) |
Most aberrant; Observer NPI-Problematic; NPI-A; NPI-I; NPI-I Distress; Participant problematic |
12 domains, including irritability, agitation/aggression |
Participant or observer |
|
NRS (Neurobehavioral rating scale) |
NRS-R |
Changes in neurobehavioral functioning post-TBI |
Observer |
|
OAS (Overt Aggression Scale) |
MOAS; OAS-M |
Post-TBI aggression |
Observer |
|
POMS (Profile of Mood States) |
|
Mood-assessment |
Participant |
|
STAS (State-Trait Anger Scale) |
|
Fluctuating states of anger, irritability, annoyance |
Participant |
|
STAXI-2 (State-Trait Anger Expression Inventory) |
|
Experience and expression of anger |
Participant or observer |
High variability in outcome measure and trial structure, as well as limited study number and sample sizes necessitate a descriptive summary of individual results rather than data synthesis. Results are reported in positive reduction values, i.e., a larger number indicates greater reduction in test score. A summary of the relevant outcomes is given in Table 3 and Figure 1.
The predominant population in this review are male. The incidence of TBI in the general population is over twice as high in men as in women [17]. The gender distribution in this review is reflective of the TBI population. The time since injury for 3 studies was at least 6 months, and the remainder had an average of 38 months. This suggests the cohort of participants were likely in the chronic phase and representative of the chronic TBI population. All participants were recruited from community settings. Participants in four studies [7,8,13,15] were recruited from rehabilitation centres and neurology or neuropsychiatry clinics. Another four studies [9,10,11,12] recruited participants from physician referrals, local brain injury support groups, and newsletters. This suggests that our review identifies mild to moderate levels of aggression. Additionally, individuals who are self-motivated to volunteer for studies may be more motivated to adhere to trial conditions than the general population.
Amantadine
Results from amantadine trials are inconsistent, which is in keeping with additional systematic reviews [4] and a Cochrane review [6]. Worsening of agitation has been noted in studies outside the scope of this paper [18]. Where significant effects are found, there is focused benefit on the outcome measure subtypes associated with aggression. Hammond et al conducted three amantadine trials. Participants taking neuroleptic medication or MAOIs were excluded but those on a stable psychoactive stimulant dosage were not.
Hammond et al [11] conducted a 28-day RCT (n = 76) which demonstrated an NPI-I reduction of 4.3 in the treatment group compared to 2.6 in the placebo (p = 0.0085) with 100mg BD amantadine or matched placebo. Exclusion of individuals with minimal to no baseline aggression increased this to 4.56 in treatment versus 2.46 in placebo (p = 0.046). 80% of the treatment group improved by 3 points compared to 44% in the placebo (p = 0.0016). Mean change in NPI-A was not significant compared to placebo for the whole cohort. After exclusion of participants with minimal to no baseline aggression, there was a mean reduction of 4.65 compared to 2.46 in placebo (p = 0.046).
Hammond et al [9] demonstrated no significant effect when p-values were adjusted to account for multiple comparisons. Raw data showed NPI-I Most Aberrant Observer score reduction of 3.74 in the treatment group compared to 3.68 in placebo (p = 0.025), which is a minimal effect. Day 60 reduction in NPI-I Most Problematic Observer score was 4.68 in treatment versus 3.8 in placebo (p = 0.025), and 3.47 in treatment versus 2.29 in placebo for Most Problematic Participant score. Day 60 NPI-I Distress Participant score showed a reduction of 1.87 in treatment versus 1.35 in placebo (p = 0.025). Day 60 reduction in NPI-I Most Aberrant Observer score was 4.39 in treatment versus 3.9 in placebo (p = 0.025) and Most Aberrant Participant score was 3.7 in treatment versus 2.77 in placebo (p = 0.025). Of note, recruitment partially from brain injury support groups, as well as no recording of utilisation of psychosocial support such as rehabilitation therapies and counselling may have increased the apparent placebo response rate.
Hammond et al [10] demonstrated NPI-A Most Problematic Participant score reduction of 5.27 in the treatment group compared to 2.89 in the placebo (p = 0.0118). Significant effects were also reported in NPI-A Distress Participant score with distress > 2. However, all other NPI and STAXI categories demonstrated no significant effect.
Carbamazepine
Our review finds no favourable evidence for the use of carbamazepine in chronic TBI aggression. Some reviews [19] use evidence from case studies to demonstrate improvement; however, they include participants in the acute phases of TBI.
Azouvi et al [7] conducted an open trial (n = 10) demonstrating a NSR-R reduction of 2.3 (p = 0.02) and ABS of 2.2 (p = 0.02) at 2 weeks after treatment initiation. At the end of the trial NSR-R reduced overall by 3.9 (p = 0.01). Participants did not continue to improve and at 8 weeks post-treatment the only significant decreases from pretrial baseline were in NSR-R subgroups irritability (2.4, p = 0.01) and disinhibition (2.04, p < 0.05). Only 5 participants showed a 50% or more decrease in NSR-R scores, 3 of whom were in the subacute post-injury stage. ABS also reduced from 32.7 to 24.4 (p = 0.02); however, any value above 21 is considered as still demonstrating agitated behaviour. Carbamazepine was started at 200mg OD, titrated to a maximum of 600-1200mg OD, although treatment response was independent of dosage. Sedative effects were noted, which limited dose increases. One significant cutaneous allergic reaction occurred.
Hammond et al [12] conducted a 42-day RCT which did not identify significant treatment effect, but also did not note any adverse treatment events. Carbamazepine was titrated to a maximum of 800mg. A history of carbamazepine hypersensitivity, contraindication, or drug interactions were exclusion criteria. Participants taking neuroleptic medication or MAOIs were excluded but those on a stable psychoactive stimulant dosage were not.
Methylphenidate
Data in support of methylphenidate is poor. Mooney & Haas [15] conducted a 6-week RCT which identified a reduction of 2.37 in STAS trait anger (p < 0.000) and minor improvements of 0.26 in KAS belligerence (p = 0.005) and 0.22 in POMS anger-hostility (p = 0.001). Methylphenidate was titrated up to 30mg OD. As a small (n = 12) single-blinded study with recognised subject randomisation bias, including only male participants, the validity of these results may be questioned.
Speech et al [16] conducted a double-blind, placebo controlled RCT (n = 12) in which participants received 1 week of methylphenidate followed by 1 week of placebo, or vice versa. No information on participant recruitment, concurrent medication use, or randomisation protocol was given. Patients, family members, and assessors were blinded. Dose was determined by weight. No significant difference was identified using the KAS scale. However, all 12 patients correctly determined the order of administration at the end of the trial and 11 reported improvement in personality functioning and cognition, as agreed by 8 of 11 observers.
Quetiapine
Our review finds scope for exploring the use of quetiapine in a further RCT. Kim & Bijlani [14] conducted a 6-week open trial (n = 7), which identified a reduction in OAS-M score of 84.5% (p = 0.002) with quetiapine titrated to a maximum of 800mg OD. Participants using neuroleptics, lithium, or psychoactive drugs were excluded but those taking antidepressants, benzodiazepines, or anticonvulsants were not. However, there was no blinding, no control group, and no compliance or sociodemographic data. Sedation was reported in 43% of participants.
Risperidone
There is minimal data in support of risperidone use. Deb et al [8] conducted a 12-week RCT (n=14) of risperidone vs placebo and identified a MOAS reduction for both placebo (9.5) and risperidone (2.8) to approximately the same level, accounted for by the placebo group having a higher starting MOAS score. No p-values were reported, and the authors stated this was not a significant effect. Aggression was defined as any form of behaviour directed towards the goal of harming or injuring another living being, divided into verbal and physical aggression. Antipsychotic use was an exclusion criterion. Other reviews [6,19] did not identify any further evidence aside from case studies.
Sertraline
Our review finds there is evidence for exploring the efficacy of sertraline in an RCT. Kant et al [13] conducted an 8-week open-label study (n = 13), which demonstrated a reduction in OAS-M aggression score of 226.79 (p < 0.01) and OAS-M irritability of 0.85 (p < 0.01) at 8 weeks with sertraline titrated to 200mg OD. However, this was an open trial with lack of blinding, small size, and lack of a control group. A systematic review [19] did not find that sertraline had significant effect in acute TBI.
Table 3: Summary of results from identified studies with main identified significant treatment effect and relevant study details.
|
Treatment |
Study |
Result |
Conclusion |
Salient study details |
|
Amantadine
|
Hammond et al, 2014 [11] |
NPI-I reduction of 4.3 (treatment) vs 2.6 (placebo) p = 0.0085 |
Positive treatment effect |
RCT n = 76 |
|
Hammond et al, 2015 [9] |
No significant effect when accounting for multiple comparison |
No treatment effect |
RCT n = 168 |
|
|
Hammond et al, 2017 [10] |
NPI-A Most Problematic (participant rating) reduction of 5.27 (treatment) vs 2.89 (placebo) p = 0.0118 |
Positive treatment effect |
RCT n = 118 |
|
|
Carbamazepine
|
Azouvi et al, 1999 [7] |
NSR-R reduction of 3.9 p = 0.01 |
Limited positive treatment effect |
Open trial; no placebo n = 10 |
|
Hammond et al, 2021 [12] |
No significant effect |
No treatment effect |
RCT n = 70 |
|
|
Methylphenidate
|
Mooney & Haas, 1993 [15] |
STAS Anger reduction of 2.37 p < 0.000 |
Positive treatment effect |
RCT n = 12 Randomisation bias |
|
Speech et al, 1993 [16] |
No significant effect |
No treatment effect |
RCT n = 12 |
|
|
Quetiapine |
Kim & Bijlani, 2006 [14] |
OAS-M reduction of 84.5% p = 0.002 |
Positive treatment effect |
Open trial; no placebo n = 7 |
|
Risperidone |
Deb et al, 2020 [8] |
No significant effect |
No treatment effect |
RCT n = 14 |
|
Sertraline |
Kant et al, 1998 [13] |
OAS-M raw score reduction of 226.79 p < 0.01 |
Positive treatment effect |
Open trial; no placebo n = 13 |
Conclusions
Significant results were found for amantadine, carbamazepine, sertraline, and quetiapine. A trial using amantadine conducted by Hammond et al [11] was the only RCT (n = 76) and is therefore the best evidenced; however, follow-up trials failed to replicate these results. The variability in effect size of amantadine may be due to off-target effects, as well as differences in the cognitive profile of included participants relating to severity and time since injury, that may differentially influence the concentration or distribution of drug targets. The studies for carbamazepine, sertraline, and quetiapine that identified significant treatment effect were small open trials. Therefore, amantadine was the only treatment with level 2 evidence of benefit for aggression and irritability in this patient population in our review. Although this is not conclusive evidence, clinicians should consider amantadine for managing aggression in the TBI population, particularly those > 6 months after sustaining a closed head injury. This is supported by other reviews [20,21].
These studies indicate amantadine, sertraline, quetiapine, methylphenidate and risperidone are tolerated well in this patient population. Only one patient had a severe allergic reaction to carbamazepine. Methylphenidate was not noted to have increased or worsened aggressive behaviours.
There was significant variability in patient inclusion criteria in these studies. Study protocols tended to include subjects stabilised on concurrent medication; however, delayed or additive effects from these treatments cannot be discounted. Further studies should focus on the application of uniform inclusion criteria to increase the likelihood of producing reliable and reproducible results. They could also consider withdrawing medications with a washout period to remove the additive effect of concurrent medication.
There was significant variability in aggression scales. Each scale attempts to characterise a broad variety of behaviours. To allow data synthesis, utilisation of a variety of aggression scales, each focused on a sub-group of behaviour (e.g. verbal vs physical) verified in the post-TBI community would allow flexibility of selection for the specific study population as well as inter-publication comparison. However, the authors accept this would require further consensus in the research community that may not yet be present.
References
- Gould KR, Hicks AJ, Hopwood M, Kenardy J, Krivonos I, Warren N, et al. The lived experience of behaviours of concern: A qualitative study of men with traumatic brain injury. Neuropsychol Rehabil. 2019 Mar 16;29(3):376–94.
- Tam S, McKay A, Sloan S, Ponsford J. The experience of challenging behaviours following severe TBI: A family perspective. Brain Inj. 2015 Jul 1;29(7–8):813–21.
- Francisco GE, Walker WC, Zasler ND, Bouffard MH. Pharmacological management of neurobehavioural sequelae of traumatic brain injury: A survey of current physiatric practice. Brain Inj. 2007;21(10):1007–14.
- Rahmani E, Lemelle TM, Samarbafzadeh E, Kablinger AS. Pharmacological Treatment of Agitation and/or Aggression in Patients with Traumatic Brain Injury: A Systematic Review of Reviews. Journal of Head Trauma Rehabilitation. 2021 Jul 1;36(4):E262–83.
- Williamson D, Frenette AJ, Burry LD, Perreault M, Charbonney E, Lamontagne F, et al. Pharmacological interventions for agitated behaviours in patients with traumatic brain injury: A systematic review. BMJ Open. 2019 Jul 1;9(7).
- Fleminger S, Greenwood RR, Oliver DL. Pharmacological management for agitation and aggression in people with acquired brain injury. Cochrane Database of Systematic Reviews. 2006 Oct 18;
- Azouvi P, Jokic C, Attal N, Denys P, Markabi S, Bussel B. Carbamazepine in agitation and aggressive behaviour following severe closed-head injury: Results of an open trial. Brain Inj. 1999;13(10):797–804.
- Deb S, Aimola L, Leeson V, Bodani M, Li L, Weaver T, et al. Risperidone versus placebo for aggression following traumatic brain injury: A feasibility randomised controlled trial. BMJ Open. 2020 Sep 10;10(9).
- Hammond FM, Sherer M, Malec JF, Zafonte RD, Whitney M, Bell K, et al. Amantadine Effect on Perceptions of Irritability after Traumatic Brain Injury: Results of the Amantadine Irritability Multisite Study. J Neurotrauma. 2015 Aug 15;32(16):1230–8.
- Hammond FM, Malec JF, Zafonte RD, Sherer M, Bogner J, Dikmen S, et al. Potential Impact of Amantadine on Aggression in Chronic Traumatic Brain Injury. Journal of Head Trauma Rehabilitation. 2017;32(5):308–18.
- Hammond FM, Bickett AK, Norton JH, Pershad R. Effectiveness of amantadine hydrochloride in the reduction of chronic traumatic brain injury irritability and aggression. Journal of Head Trauma Rehabilitation. 2014 Sep 1;29(5):391–9.
- Hammond FM, Zafonte RD, Tang Q, Jang JH. Carbamazepine for Irritability and Aggression after Traumatic Brain Injury: A Randomized, Placebo-Controlled Study. J Neurotrauma. 2021 Aug 15;38(16):2238–46.
- Kant R, Smith-Seemiller L, Zeiler D. Treatment of aggression and irritability after head injury. Brain Inj. 1998;12(8):661–6.
- Kim E, Bijlani MBAM. A Pilot Study of Quetiapine Treatment of Aggression Due to Traumatic Brain Injury. J Neuropsychiatry Clin Neurosci [Internet]. 2006;18(4):547–9. Available from: http://neuro.psychiatryonline.org547
- Mooney GF, Haas LJ. Effect of Methylphenidate on Brain Injury-Related Anger. Arch Phys Med Rehabil. 1993;74(2):153–60.
- Speech TJ, Rao SM, Osmon DC, Sperry LT. A double-blind controlled study of methylphenidate treatment in closed head injury. Brain Inj. 1993;7(4):333–8.
- Frost RB, Farrer TJ, Primosch M, Hedges DW. Prevalence of traumatic brain injury in the general adult population: A meta-analysis. Vol. 40, Neuroepidemiology. 2013. p. 154–9.
- Gramish JA, Kopp BJ, Patanwala AE. Effect of Amantadine on Agitation in Critically Ill Patients with Traumatic Brain Injury. Clin Neuropharmacol. 2017 Sep 1;40(5):212–6.
- Levy M, Berson A, Cook T, Bollegala N, Seta E, Tursanski S, et al. Treatment of agitation following traumatic brain injury: A review of the literature. NeuroRehabilitation. 2005;20:279–306.
- Mohamed MS, El Sayed I, Zaki A, Abdelmonem S. Assessment of the effect of amantadine in patients with traumatic brain injury: A meta-analysis. Vol. 92, Journal of Trauma and Acute Care Surgery. Lippincott Williams and Wilkins; 2022. p. 605–14.
- Theofanopoulos A, Tsitsopoulos P. The Role of Amantadine in Traumatic Brain Injury. A Systematic Review. Brain and Spine. 2024;4(3):127.
About the authors
Dr Hugo Smith is a Foundation Year 2 doctor at Royal Surrey County Hospital.
Drs Emma Collison-Ani & Adetunji Isola are Psychiatry Higher Trainees at St. Andrew‘s Hospital, Northampton.
Dr Gayathri Burra is a Consultant Psychiatrist at St Andrew’s Hospital, Northampton.
APPENDIX
Appendix 1: Cochrane risk of bias assessment for RCTs
Table A1: Cochrane risk of bias assessment for RCTs
|
Bias in selection of the reported result |
Some concerns: no pre-planned analysis as it was a feasibility study |
Low risk |
Low risk |
Low risk |
Low risk: intention-to-treat analyses |
Some concerns: data not analysed as per pre-specified plan |
Low risk |
|
Measurement of outcome |
Low risk |
Low risk |
Low risk |
Low risk |
Low risk |
Some concerns: single blinded |
Some concerns: no discussion on whether researchers were aware of grouping |
|
Missing data |
Some concerns: 3 dropped out, and data available for 1 |
Low risk: 6% did not complete study |
Low risk: 6% did not complete the study |
Low risk: 4 dropped out |
Low risk: 3 withdrew, but entire sample analysed |
Low risk: all subjects analysed |
Low risk |
|
Deviations |
Some concern: descriptive analysis |
Low risk |
Low risk |
Low risk: intention-to-treat analyses |
Low risk |
Some concerns: altered analysis |
High risk: participants were aware of their allocation status |
|
Randomisation |
Low risk: web-based randomisation |
Low risk: computer generated |
Low risk: computer generate |
Low risk: computer generated |
Low risk: computer generated |
Some concerns: both groups not similar |
Some concerns: no details of randomisation, baseline characteristics, etc. |
|
Author, date |
Deb et al, 2020 |
Hammond et al, 2015 |
Hammond et al, 2017 |
Hammond et al, 2014 |
Hammond et al, 2021 |
Mooney et al, 1993 |
Speech et al, 1993 |
Appendix 2: list of full-text articles excluded, with reasons
Table A2: list of full-text articles excluded, with reasons. NOSIE: the nurses observation scale for inpatient evaluation
|
Study |
Features |
Intervention |
Reason for exclusion |
|
Beresford et al, 2022 |
Design: Double-blind placebo controlled Participants: n = 50. Working age adults with self-reported TBI at least 1 year prior to study. Outcome measure: ABS, NRS-R, CGI-S Follow-up: 8 weeks |
Divalproex sodium |
Included patients with concurrent alcohol use disorder |
|
Brooke, 1992 |
Design: Double-blind placebo controlled Participants: n=21. CHI inclusion criteria including loss of consciousness, GCS, agitation Outcome measure: OAS Follow-up: 8 weeks
|
Propranolol |
Time since injury not specified Unclear if patients were in post-traumatic amnesia phase |
|
Brossart, 2008 |
Design: Randomised double-blind crossover Participants: n=10. From outpatient clinic – TBI or hypoxic brain injury Outcome measure: ABS Time since injury: >1 year prior to study Follow-up: 6-12 weeks |
Propranolol |
Single case research design No information about the methodology of the clinical trial |
|
Chatham-Showalter, 1996 |
Design: Case series Participants: n=7. Trauma critical care setting Outcome measure: Rancho scale Follow-up: 14-40 days
|
Carbamazepine |
Case series Patients in the acute phase of TBI |
|
Fann et al, 2000 |
Design: nonrandomised single blinded placebo controlled study Participants: n=15. Mild TBI in the community Outcome measure: Various rating scales Follow-up: 8 weeks |
Sertraline |
Patients had primary psychiatric diagnoses prior to sustaining TBI |
|
Gramish, 2017 |
Design: Retrospective cohort study Participants: n=139. TBI patients in surgical ICU Outcome measure: RASS Follow-up: 7 days |
Amantadine vs non-amantadine |
Patients in the acute phase of TBI |
|
Greendyke, 1986 |
Design: Randomised double-blind placebo controlled crossover study Participants: n=10. Organic brain impairment referred to neurobehavioral ICU Outcome measures: NOSIE clinical record Time since injury: 1-30 years Follow-up: 25 weeks
|
Propranolol |
Patients had varied diagnoses, 8 had acquired brain injury No standardised rating scaled Used many of the same partipants as OTHER GREENDYKE PAPER |
|
Greendyke, 1986 |
Design: Randomised double-blind placebo controlled crossover study Participants: n=11. Various brain diseases referred for behavioral disturbances Outcome measure: record of assaultive behaviors Time since injury: 1-38 years |
Pindolol |
Patients had various diagnoses, 11 had acquired brain injury No standardised rating scales Used many of the same participants as OTHER 1986 PAPER |
|
Greendyke, 1989 |
Design: Randomised double-blind placebo controlled crossover study. RCT for 21 weeks, open trial for 12 weeks Participants: n=13. Agitation with varied diagnoses Outcome measure: OAS Follow-up: 33 weeks |
Pindolol |
Patients had varied diagnoses, 10 had acquired diagnosis Time since injury not specified |
|
Gualtieri et al, 1988 |
Double blind placebo-controlled crossover study Participants: n=15. Outcome measure: various rating scales and neuropsychological tests Time since injury: 47 months Follow-up: 6 weeks |
Methylphenidate |
Aggression not a primary outcome |
|
Hammond, 2018 |
Design: Randomised placebo-controlled trial Participants: Patients with TBI and behavioral disturbance Time since injury: > 6 months Follow-up: 60 days |
Amantadine |
NPI-I for entry into the study Cognition was the main outcome |
|
Hart, 2018 |
Design: Randomised placebo-controlled trial Participants: n=32. TBI specialist inpatient unit Outcome measure: ABS Time since injury: 2 months Follow-up: 3 weeks |
Dextroamphetamine |
Study aimed at cognitive recovery from TBI |
|
Maryniak, 2001 |
Design: Retrospective chart review Participants: n=120. Acquired brain injury inpatient unit Time since injury: within 3 months |
Methotrimeprazine |
Retrospective chart review No control |
|
McKay, 2021 |
Design: observational prospective study Participants: n=125. Inpatient TBI rehabilitation unit Outcome measure: ABS Time since injury: patients in post-traumatic amnesia Follow-up: length of stay / 67 days |
Antipsychotics |
Patients in acute phase of TBI |
|
Michals, 1993 |
Design: case series Participants: n=9. Brain injury with cognitive impairment |
Clozapine |
Case series |
|
Schneider et al, 1999 |
Design: randomised double-blind placebo-controlled crossover design Participants: n=10. TBI in acute brain injury rehabilitation unit. Outcome measure: Neurobehavioral rating scale Follow-up: 6 weeks |
Amantadine |
Patients may have been in acute phase of recovery Aggression not defined |
|
Stanislav, 1994 |
Design: retrospective medical records review Participants: n=20 Outcome measure: Behavior therapy reports (no rating scale) Follow-up: 1 week-7 months |
Buspirone |
Retrospective chart review Only 70% had TBI No controls |
|
Wroblewski, 1997 |
Design: case series Participants: n=5 |
Valproic acid |
Case series No controls No rating scale |
Appendix 3: reference list of full-text articles excluded
- Beresford T, Ronan P, Hipp D, Schmidt B, Thumm E, Temple B, Wortzel H, Weitzenkamp D, Emrick C, Kelly J, Arciniegas D. A double-blind placebo-controlled, randomized trial of divalproex sodium for posttraumatic irritability greater than 1 year after mild to moderate traumatic brain injury. The Journal of Neuropsychiatry and Clinical Neurosciences. 2022;34(3):224-232.
- Brooke MM, Patterson DR, Questad KA, Cardenas D, Farrel-Roberts L. The treatment of agitation during initial hospitalization after traumatic brain injury. Archives of Physical Medicine and Rehabilitation. 1992;73(10):917-21.
- Brossart DF, Meythaler JM, Parker RI, McNamara J, Elliott TR. Advanced regression methods for single-case designs: Studying propranolol in the treatment for agitation associated with traumatic brain injury. Rehabilitation Psychology. 2008;53(3):357-69.
- Chatham-Showalter PE. Carbamazepine for combativeness in acute traumatic brain injury. The Journal of Neuropsychiatry and Clinical Neurosciences. 1996;8(1):96-9.
- Fann J, Uomoto J, Katon W. Sertraline in the treatment of major depression following mild traumatic brain injury. The Journal of Neuropsychiatry and Clinical Neurosciences. 2000;12(2):226-32.
- Gramish JA, Kopp BJ, Patanwala AE. Effect of amantadine on agitation in critically ill patients with traumatic brain injury. Clinical Neuropharmacology. 2017;40(5):212-6.
- Greendyke RM, Kanter DR, Schuster DB, Verstreate S, Wootton J. Propranolol treatment of assaultive patients with organic brain disease: a double-blind crossover, placebo-controlled study. Journal of Nervous and Mental Disease. 1986;174(5):290-4.
- Greendyke RM, Kanter DR. Therapeutic effects of pindolol on behavioral disturbances associated with organic brain disease: a double-blind study. The Journal of Clinical Psychiatry. 1986;47(8):423-6.
- Greendyke RM, Berkner JP, Webster JC, Gulya A. Treatment of Behavioral Problems with Pindolol. Psychosomatics. 1989;30(2):161-5.
- Gualtieri CT, Evans RW. Stimulant treatment for the neurobehavioural sequelae of traumatic brain injury. Brain injury : BI.2(4):273-90.
- Hammond FM, Sherer M, Malec JF, Zafonte RD, Dikmen S, Bogner J, et al. Amantadine Did Not Positively Impact Cognition in Chronic Traumatic Brain Injury: A Multi-Site, Randomized, Controlled Trial. J Neurotrauma. 2018;35(19):2298-305.
- Hart T, Whyte J, Watanabe T, Chervoneva I. Effects of dextroamphetamine in subacute traumatic brain injury: A randomized, placebo-controlled pilot study. Journal of Neuroscience Research. 2018;96(4):702-10.
- Maryniak O, Manchanda R, Velani A. Methotrimeprazine in the treatment of agitation in acquired brain injury patients. Brain Injury. 2001;15(2):167-74.
- McKay A, Trevena-Peters J, Ponsford J. The Use of Atypical Antipsychotics for Managing Agitation After Traumatic Brain Injury. Journal of Head Trauma Rehabilitation. 2021;36(3):149-55.
- Michals ML, Crismon M, Roberts S, Childs A. Clozapine response and adverse effects in nine brain-injured patients. Journal of Clinical Psychopharmacology. 1993;13(3):198-203.
- Schneider WN, Drew-Cates J, Wong TM, Dombovy ML. Cognitive and behavioural efficacy of amantadine in acute traumatic brain injury: An initial double-blind placebo-controlled study. Brain Injury. 1999;13(11):863-72.
- Stanislav SW, Fabre T, Crismon M, Childs A. Buspirone’s efficacy in organic-induced aggression. Journal of Clinical Psychopharmacology. 1994;14(2):126-30.
- Wroblewski B, Joseph A, Kupfer J, Kalliel K. Effectiveness of valproic acid on destructive and aggressive behaviours in patients with acquired brain injury. Brain Injury. 1997;11(1):37-47.