propranolol has been advocated to control agitation after brain injury. It reportedly lacks some of the deleterious cognitive and emotional effects of other ...
917
The Treatment of Agitation During Initial Hospitalization After Traumatic Brain Injury Marvin M. Brooke, MD, David R. Patterson, PhD, Kent A. Questad, Diana Cardenas, MD, Lisa Farrel-Roberts, RN
PhD,
ABSTRACT. Brooke MM, Patterson DR, Questad KA, Cardenas D, Farrel-Roberts L. The treatment of agitation during initial hospitalization after traumatic brain injury. Arch Phys Med Rehabil 1992;73:917-21. l Agitation after traumatic brain injury is disruptive for patient care, distressing, and difficult to treat. The use of propranolol has been advocated to control agitation after brain injury. It reportedly lacks some of the deleterious cognitive and emotional effects of other medications and physical restraints. This study was designed to test if proprano101is effective in reducing agitated behavior. Subjects had traumatic closed-head injury treated at a combined Level I Trauma Center and Rehabilitation Center. Twenty-one subjects met the criteria of agitation and were treated with propranolol or placebo in a double-blind fashion. The intensity of agitation was significantly lower in the treatment group although the number of episodes were similar. The use of restraints was also significantly lower in the treatment group. The results support the effectiveness of propranolol in reducing the intensity of agitation during the initial hospitalization after closed-head injury. fa 1992 by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation KEY M.ORDS: Agitation; Bruin injury: Propranolol
Agitation after traumatic brain injury (TBI) is disruptive for patient care, distressing to caregivers, and difficult to treat. It is generally believed that agitation is a common problem and that it frequently occurs as patients become responsive. In a previous study,’ investigators found that 11 of 100 consecutive TBI patients had episodic agitation and 35 had restlessness. Attempts to control such behaviors in the TBI population are often through the use of major tranquilizers, which are known to decrease learning abilities in normal individuals.2 It is not known exactly how P-adrenergic blocking medications produce their effects on behavior. There have been several studies on the use of &blockers for situational anxiety in normal subjects, which were recently reviewed by the journal Lancet.3 Subject performance on examination, musical performance, and public speaking improved through the use of these agents. However, there were some reports that they caused impaired memory, complex tracking, and complex reaction time.3 Controlled studies, comparing propranolol and benzodiazepines for the treatment of anxiety, have produced contradictory results.“6 It is difficult to make comparisons with these studies, however, because From Tufts University School of Medicine (Dr. Brooke), Boston, MA, and University of Washington School of Medicine (Drs. Peterson, Questad. Cardenas. and Ms. Farrel-Roberts), Seattle, WA. Submitted Julv 8. 199 I. Accented in revised form December 13. 199 I This research was supported \n part by the National lnstitute~on Disability and Rehabilitation Research grant GOO830076. Department of Education, Washington DC, and Harborview Injury Prevention and Research Center, Center for Disease Control grant CCR49-002570. Reprint requests to Marvin M. Brooke. MD. Chairman, Department ofRehabilitation Medicine. Tufts University School of Medicine, 136 Harrison Avenue, Boston, MA021ll. 6 1992 by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation 000%9993/92/7310-0088$3.00/0
they did not involve subjects with TBI and were concerned with subjective anxiety rather than agitated behavior. Propranolol has been described as an effective medication for agitation after brain injury and does not have some of the serious cognitive and emotional side effects of other medications or physical restraints. Case reports have noted that this drug reduces belligerent behavior in doses of 60 to 520mg/day without producing memory impairment, increasing disorientation, or otherwise altering thought processes.7-9 It has often been reported to be effective with an organically compromised sample of patients even when large doses of major tranquilizers were not.’ There is little literature about the effects of medications such as propran0101in controlled trials with closed-head-injury subjects. Greendyke and associates” recently published a doubleblind crossover, placebo-controlled study on the effects of propranolol in the treatment of assaultive behavior. The study sample consisted of ten subjects with organic brain dysfunction secondary to trauma, tumor, alcoholism, encephalitis, or Huntington’s disease. With a mean dose of propranolol of 520mg a day, a significant reduction in assaultive behavior was reported, as measured by episodes recorded by the nursing staff. Five subjects showed marked improvement and two showed moderate improvement. The episodes for these seven subjects decreased from 88 during the 1 l-week placebo period to 52 during the 1lweek active drug period. It is important to consider the methodological problems that have occurred in studies of propranolol or any medication that is used to control agitation. Rao and coworkers” found a variety of such issues when they treated 11 agitated patients with haloperidol, and compared them to less severely impaired head-injured patients. The problems they noted included finding an acceptable definition of agitaArch Phys Med Rehabil Vol73, October 1992
918
AGITATION TREATMENT AFTER BRAIN INJURY, Brooke
tion, the validity of dependent variables, and the lack of randomized and controlled group trials with brain-injured patients. As a solution to one of these problems, the Overt Aggression Scale has recently been used in an attempt to develop a valid and reliable rating scale for these behaviors.12 Most of the literature supporting the efficacy of propran0101in reducing agitation in brain-injured patients is anecdotal in nature. The recent study by Greendyke” is a notable exception in that a sophisticated study design was used. It would be useful to replicate these results with a more homogeneous sample, by limiting subjects to those with exclusive diagnoses of TBI, and to use more sophisticated measures of agitation. This study was designed to see if propranolol is more effective than a placebo in reducing agitated behavior.
METHODS Subjects All severe, traumatic closed-head-injury patients at a combined Level I Trauma Center and Rehabilitation Center over an 1S-month period were assessed to determine if they met the admission criteria of the study throughout their hospital stay. The study included only patients with severe, traumatic, closed-head injuries with more than 1 hour of unconsciousness, a Glasgow Coma Scale less than 8 on admission, and demonstrated agitation. Agitation was defined as episodic motor or verbal behavior that interfered with patient care, therapy, or safety. The criteria used for this study included any of the following: (1) any agitation severe enough to be scored on the Overt Aggression Scale; (2) the presence of clear documentation of this behavior by direct observation, staff interview, or chart review; or (3) the appropriate use of restraints or medications for agitation. We excluded patients who died, did not stay in the hospital at least one week, or who had a history of drug, alcohol, or psychiatric problems severe enough to require hospitalization for these diagnoses independent of TBI. Subjects were also excluded from the study if they had had metabolic or drug-induced coma, cardiac arrhythmias, congestive heart failure, baseline blood pressure less than 90/50mmHg, heart rate below 50bpm, the presence or history of bronchial asthma, diabetes mellitus, or did not give informed consent.
Measures Several times weekly, a research nurse reviewed hospital chart notes and interviewed direct care staff to determine if any subjects had had possible episodes of agitation. Each possible episode of agitation was rated on the Overt Aggression Scale (OAS; Yudofsky7), which categorizes types and severity of such behavior. The OAS forms were scored on a weekly basis and the maximum OAS score was used to represent the level of agitation of each subject for the week. The maximum level was used instead of the average or median level because it is theoretically independent of the Arch Phys Med Rehabil Vol73, October 1992
number of episodes. The number of OAS forms scored greater than 0 for each subject was totaled on a weekly basis to track the number of episodes per week. Use of restraints and supplemental medications for agitation and sedation for reasons other than agitation were also recorded. The number of subjects who were placed in restraints or received supplemental medications were added together to give totals of subjects in each category for each week.
Procedure Subsequent to determining eligibility and obtaining informed consent, subjects were randomly assigned to the double-blind, placebo-controlled treatment phase. After one week of baseline measurement, each subject received scheduled increasing and then tapering doses of proprano101or placebo. The propranolol (or placebo) was sent by the pharmacy as propranolol LA,” beginning with 60mg a day, increased by 60mg a day every third day, to a maximum of 420mg unless agitation ceased or side effects occurred. After three weeks, active or inert drugs were tapered over two weeks. Subjects received analgesic and anxiolytic medications as were judged necessary by their primary physician for the treatment of pain, agitation, sleeplessness, or anxiety. No rehabilitation therapy or treatment was altered by the enrollment of any subject into the study.
RESULTS Twenty-one subjects met the inclusion/exclusion criteria and entered the study. The number of subjects in the study week by week decreased as subjects were discharged. There were 11 subjects randomly assigned to the treatment group and ten to the placebo group. A Wilcoxin matched pairs test, however, showed that there was no significant difference in the number of subjects in the active drug and placebo groups (z = -.7338) and a Spearman rank order correlation coefficient showed the pattern of attrition was quite similar in the two groups (r = 889, p < .OS). It was not necessary for any subject to withdraw from the study because of adverse side effects. Subjects who dropped out of the study were either transferred to another hospital or refused all medications and most care. The average maximum intensities of agitation by week for the placebo and the active drug groups are shown in figure 1. A Wilcoxin matched pairs test showed they were significantly higher in the placebo group (z = -2.028, p < .05) and a Spearman rank order correlation coefficient showed the patterns ofincrease and decrease were not significantly similar (r = .49 1). The average number of agitation episodes by week for the placebo and the active drug groups are shown in figure 2. A Wilcoxin matched pairs test showed they were not significantly greater in the placebo group (z = - 1.5213) and a Spearman rank order correlation coefficient showed the patterns of increase and decrease were significantly similar (r = .892, p < -05). The numbers of subjects who received other interventions that might affect agitated behavior are shown in table 1. These interventions were all initiated by either the sub-
919
AGITATION TREATMENT AFTER BRAIN INJURY, Brooke
Numbers of Subjects in Propranolol Drug and Placebo Groups Who Were Placed in Restraints, Receiving Propranolol Medications Other Than Propranolol for Pain, Sleep, etc, Which Might Affect Agitation
Placed in Restraints Group Study Week
0
Prcpranotol:
60
160
300
420
420
420
120
Drug
Placebo
Receiving Propranolol Medication for Agitation Group Drug
Placebo ____
Receiving Propranolol Medication for Pain, Sleep, Etc Group ---__ Drug
Placebo
0
Fig I-Maximum intensity of agitation episodes for the placebo and active drug groups. The maximum OAS scores representing the level of agitation were averaged for the placebo group (broken line) and the active drug, propranolol, group (solid line) each week of the study. Week 0 was a baseline week with no medication, Week 1 began with 60mg of propranolol (or placebo) a day and ended with 180mg a day, Week 2 ended with 300mg a day, and Week 7 ended with the medication tapered with nothing.
ject’s primary physician or nurse who were unaware of the subject’s study status-drug or placebo. A Wilcoxin matched pairs test showed significantly more subjects in the placebo group were placed in restraints during the study (z = -2.022, p < .05), but a Spearman rank order correlation coefficient showed there was no significant similarity in the pattern of restraint use of the two groups (r = -.OSO). A Spearman rank order correlation coefficient showed there was also a significant similarity in the pattern of subjects receiving drugs for sedation to treat pain, sleeplessness, or other indications of discomfort (r = -.886, p < .05) but a Wilcoxin matched pairs test showed no significant difference between the groups in the numbers of subjects receiving sedating drugs (z = -. 104). Finally, there were no significant differences or similarities between the two groups in the numbers of subjects who received drugs for agitation. DISCUSSION The predictions in this study were that TBI patients who received propranolol would show less agitation on a variety
(1.i’
Weeks:
Pmpranolol:
0
60
180
300
420
420
420
120
0
Fig 2-Average number of agitation episodes for the placebo group (broken line) and the active drug, propranolol, group (solid line) each week of the study.
of measures than when they received placebo. The results suggested that (1) the average maximum intensity of rated agitation was higher in the placebo group: (2) more subjects were placed in restraints in the placebo group: and (3) there were no differences in numbers of agitation episodes, use of drugs for agitation, or sedation for other reasons, such as pain or sleeplessness. The results also suggest that the groups were quite similar in the number of agitation episodes that occurred, and the number of subjects that received sedating medication for reasons other than agitation. Taken together, these findings indicate that propranolol may be an effective agent for reducing the intensity, but not the frequency, of agitation in patients after severe head injuries. The similarities in numbers of episodes and other sedating indications between the groups further supports the inference that propranolol is effective. These similarities suggest that the differences in intensity of episodes were not “accidental” or artifactual differences between the two groups, which might have caused the difference in intensity of agitation, such as chance differences in the number of provocations of agitated behavior or in the propranolol use of other sedating medications. Several precautions should be mentioned in the application of these results. A previous descriptive study’ resulted in an estimate of the typical duration of agitation that is lower than the duration observed in the subjects of this study. It is possible that the use of a study drug caused care givers to decrease the use of standard measures for agitation such as decreased stimulation, increased staffing, restraints or other medications even though they were blinded to whether active or placebo medications were being given in the study. In addition, direct care staff may have been unusually vigilant in observing and recording agitated behavior. Thus, the overall magnitude and frequency or agitation may have been greater or over estimated because the patients were in a study. Without this hypothetical exaggeration in overall agitation, it is possible that the differences and similarities between study groups would not have been observed. Arch Phys Med Rehabil Vol73, October 1992
920
AGITATION
TREATMENT
AFTER BRAIN INJURY, Brooke
In this study, agitation was defined very carefully and separated from less severe or more continuous behaviors, such as restlessness. These results should not be compared to other studies, other populations, or individual patients whose behavior and injuries do not meet the same criteria. The reported results of the treatment are from a small number of subjects, and even though there is a difference in the two groups, replication of the results in a larger group is needed. The population from which this study sample was selected was specifically from all the consecutive admitted severe TBI patients admitted to a Regional Trauma Center. Application of these results to other patient groups with different diagnostic categories, severity of injuries, or referral patterns should be carried out with caution. The mechanism of propranolol’s effect on behavior is poorly understood. It is postulated that propranolol affects the central nervous system, because it has been shown to go across the blood-brain barrieri3,i4 and lowers the seizure threshold in mice. is Peripheral effects of P-blocker drugs may reduce the autonomic correlates of emotion and reduce agitation by that mechanism as we11.4,L6 The findings of this study, that propranolol may have reduced the intensity but not the frequency of agitation, suggest that it may influence behavior in a subtle and specific fashion. For example, it may be that the primary mechanism of the drug is to reduce only the emotional intensity of agitated responses by dampening autonomic correlates. Some of the newer P-adrenergic blocking medications have predominantly central effects and thereby avoid the peripheral side effects. Further studies on the behavioral effects of these medications may shed more light on the site of action that appears to effect agitated behavior and, in turn, lead to their more effective use clinically. Presently, agitation may be controlled through the use of major tranquilizers, which are known to decrease learning abilities in normal individuals.2 Patients may also be restrained or restricted in other ways, eg, to their room or in the Craig Bed, and isolated from staff and friends. It is our opinion that there has been a significant, although poorly documented increase in the frequency of use of physical and chemical restraints in favor of eliminating causes of disruptive behavior, one-to-one attendent care, and other less intrusive strategies. In our experience, care givers are often distressed by TBI agitation and cite this as a major problem when working with these patients. If propranolol continues to prove effective in decreasing agitation after brain injury, there may be several significant benefits. The ability of patients to cooperate and learn new skills in rehabilitation may increase, and the distress of patients, family, and staff may be decreased. It may decrease the time and intensity of institutionalization and amount of support needed outside the hospital for patients. The use of propranolol to promote these benefits would likely be cost effective because the medication is inexpensive and familiar to most care providers. An informal examination of the data reported here points to questions regarding the clinical process or managing agitation. The most intriguing questions come from examining the comparison of the number of subjects restrained in each group (table 1). The lack of correlation over Arch Phyr Med Rehabil Vol73, October 1992
the weeks of the study appears to be due to a lack of change in restraint use during the first weeks in the active drug group and a corresponding increase in restraint use in the placebo group. This occurred despite the relative consistency in the average maximum intensity of agitation and a consistent decrease in the number of episodes. One possible implication is that standard treatment is to apply restraints at the first incidence of agitation and to leave the patient in restraints regardless of changes in the patient’s behavior. Reducing the intensity of initial agitation with propranolol may also reduce the probability of patients being placed in restraints. Questions like these can be examined formally within the design of this study but larger numbers of subjects are needed. Despite the potential advantages, the potential detrimental effects of the use of propranolol might also be considered. First, even though there is currently no evidence for this, it cannot be ruled out that propranolol may retard neural recovery and cognitive performance among head injured patients in ways that were not detectable in previous research with normal subjects. Second, there may be a slight possibility that agitated and restless patients may exhibit a higher total level of verbal and motor behavior and hence, recover more and faster simply because of more practice of these behaviors. Thus, it may be that any effort to suppress agitated behavior may result in slower or lower levels of recovery after serious traumatic brain injury despite the apparent disruptive and counter-productive nature of such behaviors. It is also likely, however, that these hypothetical negative effects are diverse and small, so definitive answers to these questions can only come from large, longitudinal, multivariate studies ofthe early recovery from traumatic head injury.
CONCLUSIONS This study provides preliminary support for the effectiveness of propranolol in reducing the intensity of agitation and the probability of restraint use during the initial hospitalization after closed-head injury. Several questions remained unanswered, however, and larger longitudinal and multivariate studies are needed to develop optimal strategies for managing agitation in patients after serious traumatic head injury. References 1. Brooke MM, Questad KA, Patterson DR, Bashak KJ. Agitation and restlessness after closed head injury: a prospective study of 100 consecutive admissions. Arch Phys Med Rehabil 1992;73:320-30. 2. Swonger AK, Constantine LL. Drugs and therapy: a handbook of psychotropic drugs. Boston: Little, Brown, 1983: pp 305-l 1. 3. Editor. Beta-blockers in situational anxiety. Lancet 1985; 8448(2): 193. 4. Tyer PJ, Lader MJ. Response to propranolol and diazepam in somatic and psychic anxiety. Br Med J 1975;2: 14-6. 5. Johnson G, Singh B, Leeman M. Controlled evaluation of the beta adrenoceptor blocking drug oxprenolol in anxiety. Med J Aust 1976;1:909-12.
AGITATION
TREATMENT
AFTER BRAIN INJURY, Brooke
Petrie WM. Ban TA. Propranolol in organic agitation. Lancet 1981;1:324. Whitlock FA. Price J. Use of beta-adrenergic receptor blocking drugs in psychiatry. Drugs 1974;8: 109-24. Elliot FA. Propranolol for the control of belligerent behavior following acute brain damage. Ann Neurol 1977; 1:489-9 1. Y udofsky S, Williams N, Groman J. Propranolol in the treatment of rage and violent behavior in patients with chronic brain syndrome. Am J Psychiat 198 1; 1382 18-20. 0. Greendyke RM. Kanter DR, Schuster DB, Verstreate S, Wootton J. Propranolol treatment of assaultive patients with organic brain disease. J Nerv Ment Dis 1986;174:290-4. I Rao N. Jelhnek H. Wollston D: Agitation in closed head injury: haloperidol effects on rehabilitation outcome. Arch Phys Med Rehabil 1985:66:30-4.
921
12. Yudofsky S, Silver JM, Jackson W. Endicott J, Williams D. The overt aggression scale for the objective rating of verbal and physicaf aggression. Am J Psychiat 1986;143:35-9. 13. Yorkston MK, Zaki SA, Malik MK. et al. Propranolol in the control of schizophrenic symptoms. Br Med J 1974:4:633-5. 14. Young RY, Growden JH, Shahani BT: Beta-adrenergic mechanisms in action tremor. N Engl J Med 1975;293:950-3. 15. Yeoh PN, Wolf HH: The effects of some adrenergic agents in low frequency electroshock seizures. J Pharm Sci 1968: 57:340-2. 16 Cole JO, Altesman RI, Weingarten CH: Beta-blocking drugs ’ in psychiatry. McLean Hosp J 1979;4:40-68. Supplier
a. Wyeth-Ayerst 19101-1245.
Research,
PO Box 8299. Philadelphia,
PA
Arch Phys Med Rehabil Voi 73, October 1992
