| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ORIGINAL ARTICLES |
From the Department of Behavioural Sciences in Medicine, Faculty of Medicine, University of Oslo, Oslo (A.F.); Department of Physical Medicine and Rehabilitation, University Hospital of Tromsø, Tromsø, (A.W.A.); Sunnaas Rehabilitation Hospital, University of Oslo, Nesoddtangen (E.H., J.K.S.); Centre for Physiotherapy Research and Development, Oslo (K.S.R.); and Department of Surgery, Ullevål Hospital, University of Oslo, Oslo (J.P.-L.), Norway.
Address reprint requests to: A. Finset, PhD, Department of Behavioural Sciences in Medicine, University of Oslo, POB 1111 Blindern, N-0317 Oslo, Norway. Email: arnstein.finset{at}basalmed.uio.no
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
METHODS: Sixty-eight multiple trauma patients were tested with a screening battery consisting of six neuropsychological tasks 3 years after injury. A measure of psychological distress (20-item General Health Questionnaire, or GHQ-20) was also administered.
RESULTS: Patients who neither showed signs of reduced consciousness on admission to the hospital nor reported significant psychological distress at follow-up tended to have normal test performance. In five of the six tasks, cognitive impairment was related to the severity of the traumatic brain injury as measured by the Glasgow Coma Scale (GCS). In both attention span tasks, patients designated as cases by the GHQ had significantly lower scores than noncase patients. These bivariate relationships were upheld in multiple regression analyses, in which age, sex, and GCS and GHQ scores were entered as independent variables. When patients with severe head injuries were excluded from the analyses, GCS scores still contributed to the variance in tests of verbal attention span and delayed recall, but performance on attentional tasks was more strongly related to psychological distress than to GCS scores.
CONCLUSIONS: Cognitive deficits in multiple trauma patients were related both to the severity of the traumatic brain injury and to the degree of psychological distress. The strength of the association between brain injury as indicated by GCS scores and cognitive performance differed between different tasks. Neuropsychological testing may assist in differentiating primary organic from secondary psychogenic impairments.
Key Words: multiple trauma • neuropsychological tests • cognitive impairment • psychological distress • brain injuries
Abbreviations: AIS = Abbreviated Injury Scale; ANOVA = analysis ofvariance; GCS = Glasgow Coma Scale; GHQ = General HealthQuestionnaire; GHQ-20 = 20-item General Health Questionnaire; MT = multiple trauma; PASAT = Paced Auditory Serial AdditionTask; TBI = traumatic brain injury.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Psychological and psychiatric problems are often seen subsequent to MT (3–6). Depressive disorders are the most frequent psychiatric complication (3). Posttraumatic stress symptoms, including intrusive distressing recollections, avoidance behavior, and hyperarousal, are also reported (6, 7). Although the degree of psychological distress is associated with the severity of the impact of injury, the specific emotional responses to injury are reported to reflect the personal meaning of the trauma rather than injury severity per se (8).
Even if psychological sequelae of injury are important to patients, these aspects will often go unrecognized by health personnel, especially during the acute phase (9). Moreover, a number of patients develop their distress response over time, without showing specific adjustment reactions during the first few days or weeks after injury. Consequently, there is a risk that psychiatric disorders will remain untreated.
Changes in cognitive performance among MT patients have been investigated less frequently. Cognitive impairment has been reported mostly in MT sample populations dominated by patients with rather severe brain injury; in the literature, it has often been referred to as "organic brain syndrome" without further qualification (3, 10). Lehmann et al. (11) recently reported deficits in information-processing speed, concentration, recent memory, and learning performance in a study of 58 MT patients examined an average of 5.8 years after injury. However, most of these patients had suffered severe head injury, with a mean duration of coma of 15.4 days. In a more broadly defined MT population, however, there is distinct variability in terms of brain damage, ranging from no cerebral affliction, as can occur with mild head injury, to severe brain injury.
Even if a head injury or cerebral concussion is present, cognitive impairment may be related to different causes. Whereas cognitive changes due to cerebral damage after severe TBI are well documented (12), there are different interpretations in the research literature concerning mild head injury (13). A number of studies conducted in the 1970s and early 1980s reported frequent cognitive problems, such as attentional impairments, after mild head injuries, at least during the first few months after injury, and some investigators tended to attribute these impairments to direct cerebral injury (14, 15). More recent investigations have shown that if and when these impairments occur, they are most often transitory. Levin et al. (16), in their carefully designed multicenter study, concluded that "a single, uncomplicated minor head injury produces no permanent, disabling neurobehavioral impairment in the great majority of patients free of preexisting neuropsychiatric disorder and substance abuse."
In a minority of patients with mild head injury, however, cognitive impairments persist for a long time after injury. The pathogenesis of these long-term impairments is controversial. Whereas some authors claim that persisting cognitive impairment, even in patients who suffered a slight trauma, is due to direct and primary cerebral injury (17), most researchers today will probably attribute permanently reduced cognitive functioning in patients with mild head injury but no specific risk factors as secondary to psychosomatic or emotional factors (18, 19) or cognitive factors, such as patient expectation (20). Ongoing litigation may possibly play a role in some cases (13). An interaction between cerebral affection caused by the injury and subsequent adjustment problems has also been suggested (21).
A study of long-term sequelae in cognitive performance of MT patients should be done for at least two reasons. First, it is of clinical interest to look into the nature and spectrum of cognitive impairments after MT. Second, because of the variability within the MT population, both in the severity of brain injury and in the emotional responses of patients as time passes, an MT patient sample is suitable for studying the relative influence of the primary organic (in terms of direct cerebral damage) vs. secondary emotional factors in problems of cognitive functioning. The purposes of the present study thus were to assess cognitive test performance 3 years after MT and to explore how it varies as a consequence of the severity of head injury and degree of psychological distress.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
The GCS, a measure of the best motor and verbal responses and degree of eye opening on admission, was applied to measure the state of consciousness as an indicator of the degree of brain injury suffered (25). According to conventional classification criteria, patients were categorized into one of three GCS categories, those with severe (GCS score of 3–8 points), moderate (score of 9–12), and mild or no (score of 13–15) affection of consciousness. In research on TBI, the GCS is a standard procedure for assessment of reduced level of consciousness and is a frequently used marker of injury severity. The scale has good sensitivity and reliability, with an intraclass correlation coefficient of 0.8 to 1.0 for trained users, and well-established cross-sectional construct validity (26).
Psychological distress.
Psychological distress was measured by the GHQ-20. The GHQ is a widely used screening instrument that makes use of a wide range of indicators of psychological distress and psychological disorders in the anxiety/depression spectrum. Each item is scored along a four-point scale ranging from 0 (no distress) to 3 (severe distress) in relation to the particular aspect of distress measured (anxiety, depressed mood, difficulties in coping, etc.). The instrument has been validated in different languages and cultures and has been shown to be a valid and reliable instrument across cultures (27, 28).
GHQ-20 has been validated in a series of accidentally injured adults (29). A caseness cutoff criterion of 24 (GHQ case > 24), based on a 0–1-2–3 scoring procedure, was applied according to the criteria developed in that study.
Neuropsychological tests.
Patients were tested with a screening battery consisting of five neuropsychological tests representing six different tasks. The Digit Span subtest of Wechsler’s Adult Intelligence Scale (30) and the Knox Cubes Imitation Test (31) are frequently used as measures of verbal and nonverbal attention span, respectively.
In the PASAT (32), subjects hear a stream of single digits. Their task is to continuously add the last digit spoken to the one immediately preceding it (32). The speed of presentation can be varied. In the present study, patients were tested using two speeds. A number of patients failed to complete the test at the faster speed. Thus, only data from the test using the slower speed (one digit every 4 seconds) are presented here.
In the Luria-Christensen 10 Words serial learning test, patients are asked to learn 10 common nouns by the end of up to 10 repeated presentations (33). The test includes two tasks, Acquisition and Delayed Recall. The effectiveness of Acquisition is presented in terms of number of words learned. The score of Delayed Recall represents the number of words recalled at the end of the test session.
The Verbal Fluency task, which is particularly sensitive to left prefrontal damage, is a task in which the patient is asked to state as many words he or she can remember, beginning with the letters F, A, and S (in that order) (34). Norms are based on comparisons with responses from noninjured subjects. In the screening battery applied in the present study, patients presented as many words as they could remember during 1 minute using two different first letters (F and A).
Statistical Procedures
Neuropsychological test scores were correlated with GCS and GHQ scores, applying Pearson’s product-moment correlation (Table 3). Differences between subcategories of patients were analyzed by means of simple factorial ANOVA models (Tables 1, 2, and 4). To compare patients in the three GCS categories, post hoc computations of least significant between-group differences, with Bonferroni corrections, were applied. Finally, a series of multiple linear regression analyses were performed, with neuropsychological test scores as dependent variables (Table 5). All independent variables were entered in one block.
|
|
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
The mean GCS score at admission was 12.5 (Table 1). Patients in the 12- to 20-years-old age group tended to have somewhat lower GCS scores than older patients. Ten of the 11 severe head injuries (GCS score of 3–8) were suffered by men. Because of this, there is a trend, although not a statistically significant one, for men to have lower GCS scores than women. Obviously, patients with head injury as the most severe injury had much lower GCS scores than patients in other categories.
Thirty-four patients had a GCS score of 15, indicating no reduction in consciousness on admission to the hospital. This does not, however, rule out a possible concussion. In fact, 16 of these 34 patients had AIS scores indicating a minor concussion component in their injury. However, a minor concussion without any effect on consciousness hardly produces a brain injury at all, and GCS score was chosen as the best indicator of a head injury component.
Psychological Distress
The mean GHQ score in the sample at follow-up 3 years after injury was 22.1 (Table 1). There was a trend, although not a statistically significant one, of increased GHQ score with increasing age. No significant GHQ score differences between the sexes were found.
Distress was not related to head injury severity; in fact, there was a zero correlation between GHQ scores and TBI severity as measured by the GCS (r = 0.01). As indicated in Table 1, the GHQ score tended to be slightly higher among patients with injuries to the thorax or abdomen than in other injury categories.
Of the 67 patients who completed the GHQ, 18 (26.9%) met the case criterion (GHQ score > 24). Almost half (47.1%) of the patients with injuries to the thorax or abdomen were GHQ cases, compared with only 19.2% of head injury patients, but the difference did not reach statistical significance.
Neuropsychological Test Data
The means and standard deviations of each of the six tasks are presented in Table 2. For the sake of comparison, test scores of the subsample of 25 patients who had both a GCS score of 15 (indicating no affection of consciousness level) and a GHQ score below the case criterion are presented in Table 2. There are statistically significant differences between the two groups on four of the six tasks.
|
Female patients had significantly higher scores on the Verbal Fluency and Acquisition tasks. When GCS scores were included as covariates, differences in scores failed to reach statistical significance.
Test Results and TBI Severity
Pearson’s product-moment correlations were computed between all six neuropsychological tasks and GCS total scores. Each individual task, except the Knox Cubes test, was positively correlated with GCS scores (Table 3).
Test scores on all six tasks for patients in the three GCS categories (those with severe (GCS scores of 3–8 points), moderate (GCS score of 9–12), and mild or no (GCS score of 13–15) affection of consciousness) were compared in a pairwise manner, applying post hoc least significant differences analyses with Bonferroni corrections, for a total of 18 (6 x 3) comparisons. Patients with severe disturbance of consciousness had significantly lower scores than those with mild or no affection of consciousness on five of the six tasks: Digit Span (df = (2,64), p = .001), PASAT (df = (2,55), p = .028), Verbal Fluency (df = (2,65), p = .001), Words Learned (df = (2,56), p = .006), and Words Recalled (df = (2,65), p = .038). Patients in the moderate category had significantly lower scores than patients in the mild or no affection of consciousness group on one test only, Digit Span (df = (2,64), p = .032). Moreover, on the Verbal Fluency task, patients with severe disturbance of consciousness had significantly lower scores than patients with a moderate disturbance (df = (2,65), p = .010).
Test Results and Level of Psychological Distress: GHQ Scores
Each of the neuropsychological tasks was correlated with total GHQ scores. Both the attention span tasks (Digit Span and Knox Cubes test) were positively correlated with GHQ scores (Table 3).
Moreover, test scores for all six tasks were broken down against whether patients were rated as GHQ cases on follow-up. Again, on both attention span tasks (Digit Span and Knox Cubes test), patients designated as GHQ cases had significantly lower scores than noncase patients (Table 4). Moreover, although there were no differences on the Acquisition task in terms of the number of words learned, GHQ case patients needed a significantly higher number of presentations than required by GHQ noncase patients to reach the criterion (10 words).
|
Both the GCS and GHQ scores contributed significantly to the variance in the Digit Span task (GCS: ß = 0.42, p < .001; GHQ: ß = -0.36, p < .01). GHQ score alone contributed significantly to scores on the Knox Cubes test (ß = -0.37; p < .01). Moreover, GCS scores significantly predicted performance of the PASAT (ß = .38, p < .01), the Verbal Fluency task (ß = 0.31, p < .02), and the Acquisition (ß =.31, p < .01) and Delayed Recall (ß = 0.33, p < .01) tasks of the 10 Words test. Neither age nor sex had significant independent contributions to any of the test scores.
To test the specific influence of a moderate to mild injury on test scores, a parallel series of regression analyses were performed, excluding the 11 patients with a GCS score of 3 to 8, which indicates severe brain injury. GCS scores still represented a significant contribution in two separate tests (Digit Span: ß = 0.28, p < .05; Delayed Recall: ß = 0.29, p < .05) and nearly reached significance in one other (Acquisition: ß = 0.25, p = .051).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
The fact that neuropsychological test performance is related to the degree of head injury severity is certainly no news. But the results of the study go beyond that rather obvious observation. The main findings are that the strength and nature of the association between the brain injury indicator (GCS score) and cognitive performance differ between different tasks and, perhaps more interestingly, that performance on some tasks is more strongly related to psychological distress at the time of testing than to brain injury severity.
Patients Without Signs of Brain Injury or Distress
When the 25 patients who neither had signs of reduced consciousness on admission to the hospital nor qualified as GHQ cases were considered separately, their mean test scores were well in accordance with mean values in the published norms for each test (Table 2). Thus, we may infer that the MT patients in our study who neither suffered TBI nor experienced significant psychological distress displayed normal cognitive functioning 3 years after injury.
Brain Injury and Test Performance
As expected, the severity of head injury as indicated by the GCS score was significantly related to cognitive test performance on most tasks, and these bivariate relationships were upheld in multiple regression analyses, in which age, sex, and GHQ scores were entered as independent variables together with GCS scores.
There was one interesting exception to this rule. Nonverbal attention span, as measured by the Knox Cubes Imitation Test, did not differentiate between patients with and without brain injury. We interpret this to be the result of a floor effect in this particular task when it comes to direct sequelae of brain injury. Performance on the Knox Cubes test is reported to be significantly impaired in patients with severe, focal brain lesions (31), but few patients in our sample had a level of brain damage severe enough to produce reduced Knox Cubes scores as a result of primary cerebral injury.
Obviously, patients with severe head injuries tended to have the most significantly lowered test performance compared with other categories. However, in keeping with other studies, many patients also in the mild to moderate TBI categories displayed clinically significantly reduced cognitive functioning in the areas of verbal attention span, speed of processing (35), and memory (36), even in the absence of psychological distress.
Psychological Distress and Test Performance
More than one-fourth of patients met the caseness criterion on the GHQ-20, indicating a higher level of distress than is present in the normal population (6, 29). On the basis of results of earlier studies of MT patients, we may assume that this increased level of distress is related to the injuries suffered by the subjects in this study (3–6).
On average, patients with psychological distress, operationalized as GHQ cases, had consistent problems in terms of attention span, both verbal and nonverbal, and learning efficiency. These findings are in accordance with those published in the literature, indicating specific problems of depressed patients in the areas of attention (37, 38) and effort-demanding tasks (39, 40). It is especially interesting that GHQ cases tended to exhibit impaired performance in the task of nonverbal attention, the Knox Cubes test. This task appears to be specifically sensitive to the type of attentional deficits suffered by patients with a high degree of psychological distress.
There was a zero correlation between distress (GHQ scores) and GSC scores. Patient response to injury in terms of distress is thus quite independent of head injury severity. Even a seemingly mild or moderate injury may have devastating personal affects and produce high scores on tests of psychological distress for the individual patient (8).
Cognitive Dysfunction in MT Patients—Organic or Psychogenic?
The study confirmed that different psychological tests are sensitive to different aspects of cognitive sequelae subsequent to traumatic injury. For instance, the Verbal Fluency task is sensitive to a rather severe head injury but not to mild TBI or psychological distress. However, we also found that verbal attention span and memory tasks may be moderately affected in patients with mild to moderate injury, even when psychological distress is controlled for. The latter findings indicate that one should be careful not to rule out that a cognitive impairment may be related to brain injury, even when the head injury itself is mild to moderate.
On the other hand, a pattern of reduced attentional capacity, perhaps especially in terms of nonverbal processing and slow and inefficient performance on effort-demanding tasks, may be related to psychological distress regardless of whether a head injury is present. Similar findings are reported in other patient populations characterized by psychological distress, fatigue, and/or pain but without cerebral injury, such as patients with chronic fatigue syndrome (41, 42), fibromyalgia (43), and common whiplash (44). It is certainly important not to interpret all cognitive impairment as related to brain injury, even in patients who have suffered head trauma.
Conclusion
In conclusion, we found that a substantial number of patients in our sample of MT patients, particularly those who had neither suffered a significant head injury nor reported psychological distress, did not display signs of cognitive impairment 3 years after injury. However, several patients had test results indicative of cognitive impairments that were significantly related to their GCS and GHQ scores. Although the patterns of test performance of patients with moderate head injuries and those with psychological distress shared similar features, different neuropsychological tests seemed to be sensitive to different aspects of cognitive functioning. In interpreting findings on cognitive impairment, due attention should thus be given to the sensitivity of different tests to diverse areas of cognitive functioning that may reflect primary cerebral mechanisms of TBI, secondary psychological reactions to the injury, or both.
Received for publication August 13, 1998.
Revision received May 4, 1999.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
