- Email: [email protected]
Occult head injury: Its incidence in spinal cord injury
227
Occult Head Injury: Its Incidence in Spinal Cord Injury Conal B. Wilmot, MD, D. Nathan Cope, MD, Karyl M. Hall, EdD, Mary Acker, PhD Santa Clara Valley Medical Center, San Jose. CA 95128 ABSTRACT. Wilmot CB, Cope DN, Hall KM, Acker M: Occult head injury: its incidence in spinal cord injury. Arch Phys Med Rehabil66:227-231, 1985. l This study investigated the suspicion that a significant proportion of individuals having spinal cord injury (SCI) also sustain a concomitant undiagnosed occult head injury during the trauma accident. The criteria for high risk of head injury included the following: (1) quadriplegia with high energy deceleration accident, (2) loss of consciousness at time of injury, (3) brainstem or cortical neurologic indicators, or (4) respiratory support required at time of injury. In this study, 67 patients admitted to the rehabilitation unit were given a neuropsychologic evaluation a median of 48 days after injury. Motor free scales used were the Galveston Orientation and Amnesia Test (GOAT), Quick Test, Raven Progressive matrices, serial 7s, Shipley Hartford, Stroop Color/Word Interference, and the Wechsler Memory Scale Associate Learning Tests. Forty-three of the 67 patients. (64%) scored mildly to profoundly impaired on the test battery. Evidence of poor premorbid academic history was present in 19 (44%) of those with impaired performance on the neurologic evaluation and in only three (13%) of those scoring unimpaired. Consequently, 56% (24/43) of the impaired had no previous record of scholastic difftculties, presumably acquiring cognitive impairment at the time of injury. The implications of this high incidence of impaired cognitive functioning for treatment of individuals with SC1 are significant. KEY WORDS: Brain injuries: Cognition disorders: Head injuries; Spinal cord injuries
At the Northern California regional spinal cord injury (SCI) center, the combination of acute SC1 with head injury leading to subacute neurologic complications has been sufficiently frequent to warrant a formal investigation. Consequent to the cognitive deficits of head injury, significant management and rehabilitation training problems can develop which might better be managed with an appreciation of the occult head injury. Rime1 and associates’4 studied a large sample of individuals with minor head injury having unconsciousness 20 minutes or less. They found that three months after injury, 59% of patients complained of memory loss and 34% with previous employment were now unemployed. Results of neurologic evaluation were normal in nearly all patients. In contrast, neuropsychologic examination was sensitive to problems of attention, concentration, memory, and judgment in these patients. There have been several published reports referring to the high incidence of coexisting multiple injuries in primary diagnoses of head injury and SCI.10.‘7 Weiss” reasoned that anyone receiving a blow severe enough to render him/her unconscious is a candidate for a possible cervical spine injury. PorterI found that in 62 (29%) of 216 nonacute SC1 mortalities there was evidence of significant associated injury to the brain or skull on pathologic examination. These findings concur with those of Guttman7 and Harris.’ Davis and associates4 reported that 46% of autopsied spinal cord injuries had. demonstrable pathology of the brain. Most cases were upper as compared to lower cervical lesions. Cases with lesions below TI were not examined. However, Porter” points out that almost half of his brain injured sample had lower spinal cord injuries (T5-L4). Young2’ recorded major head injuries coincident with paraplegia at 16% and with quadriplegia, 7%. Wagner and associates I7 found that of those with head injuries, complete quadriplegic individuals had severe head injuries as clinically assessed at 24 hours, whereas complete paraplegic individuals had mild injuries. Patients with incomplete lesions also sustained head injuries but these more often were mild in both paraplegic and quadriplegic individuals.
Still to be determined is the actual incidence of head injury in a typical acute SC1 population and whether screening procedures such as neuropsychologic testing would be a sensitive and cost-effective means of diagnosing and assessing the problem. It is our clinical impression that patient management and incidence of neurosurgical complications could both be improved with a greater awareness of the coexistence of these two conditions. The purposes of this study were to: (1) establish an incidence of cognitive dysfunction in the SC1 population; (2) document the degree of cognitive impairment and relate findings to level and completeness of SC1 lesion; (3) investigate the impact of head injury on length of hospital stay; (4) investigate which clinical factors and/or tests best indicate head injury; (5) establish the prevalence of premorbid problems as determined by scholastic performance for comparison with those diagnosed head injured after injury; and (6) assess adequacy of standard admitting history, physical examination, and mental status examination in identifying patients with cognitive dysfunction.
SUBJECTS AND METHODS Subjects All SC1 patients admitted to the regional spinal injury center within one year after injury who met any of the following additional criteria for high risk were candidates for inclusion in the study: (1) quadriplegic with high energy deceleration accident (any significant impact); (2) history of or documented loss of consciousness at time of injury; (3) significant brain-
This research was supported in part by grant GO08202589 to the Northern California Regional Spinal Injury System (Project 128EH20015) from the Rehabilitation Services Administration. U.S. Department of Education. Washington, DC. Submitted for publication February 8. 1984. Accepted in revised form July I I, 1984.
Arch Phys Med Rehabil Vol66,
April 1965
INCIDENCE OF OCCULT HEAD INJURY IN SCI, Wilmot
228
Table 1: Scoring for Neuropsychologic
Test Battery Imoairment
Test l.GOAT
levels
1
2
3
4
5
6
None
Mild
Moderate
Severe
Extreme
Profound
90+
x0-w
70-79
ho-641
95+
K-Y-!
75-M
K-71
50+
-IO-JY
x1-39
X-ZY
4-Y
7
(100minus error points 2. Quick I Q 3-5. Stroop Color Word (T-scores) (Do not use interference score for these impairment
categories) O- I
7
i-3
5-h
7. Raven IQ
95+
x5-94
75-M
65-74
8. Wechsler Word Pairs IQ
95 +
US-94
75-x-1
65-74
0+
l-8
5-h
3-4
6. Serial 7s (errors)
9. Wechsler Delayed Word Pairs (number correct) 10. Shipley Hartford Vocabulary (number correct)
xt
77.75 _.. _.
18-21
I-1-17
11. Shipley Hartford Abstraction (numbercorrect)
77+
‘2-26
16-21
IO- I5
stem or cortical neurologic indicators; and (4) respiratory support required at time of accident. During the study (February to December 1983). 7 1 of 77 trauma patients admitted to the rehabilitation unit met the criteria. Sixty-seven (55 males, 12 females) actually participated. Four were not tested; none refused to participate. Initially. we planned to compare “high risk” patients with those who did not meet criteria for high risk of head injury, but it quickly became evident that almost all SC1 patients qualified as high risk, given the criteria used for this study. Study patients had an average age of 29 years (range: 12 to 69 years). Average length of hospital stay was 116 days (median 115 days). Sample distribution by level of lesion was: 18 high quadriplegics (C2-4), 33 low quadriplegics (C5-8). eight upper thoracic (Tl-7), and eight lower thoracic (T8 and below). Of the 67 patients, 41 (61%) had complete lesions. Patients were given the neuropsychologic evaluation as soon as possible after they were medically stable and could com-
”I
Number of patients in each category of impairment. Arch Phys
Med Rehabil
Vol 66, April 1985
plete the 90-minute testing battery with the psychologist. Patients were tested an average of 63 days after injury with a range of ten to 246 days (median 48 days). Some patients had been hospitalized for a considerable period of time when the study began. thus lengthening average days to testing for the total group.
Methods Tests used for neuropsychologic evaluation excluded procedures dependent upon motor functions. Tests used were: Galveston Orientation and Amnesia Test (GOAT).” This test consists of ten items addressing orientation to person. place. and time. and memory for events. Duration of post-traumatic amnesia, as defined by persistence of defective GOAT scores. was longer in patients with CT-scan evidence of diffuse or bilateral brain injury as compared to cases with focal unilateral lesions. Initial standardization was based on 50 patients who had sustained mild closed head injury. Quick Test. ’ This procedure. which measures vocabulary and simple verbal reasoning, requires the test subject to identify which of four pictures on a card best represents words given by the examiner. Six equivalent forms are available, with 50 words on each form. Mental age norms for levels 1.5 through 19.0 years are provided. along with adult percentiles and Wechsler-type IQs based on the assumption of a mean 1Q of 100 and a standard deviation of 15 IQ points. Stroop Color/Word Test’ (three subtests). This procedure consists of three tasks, each timed at 45 seconds: reading names of colors printed in black on white; identifying colors of capital xs printed in red. blue. and green; and identifying colors in which color names are printed. while suppressing the word itself. Raw scores are the number of stimuli correctly read or identified in each 45-second task. These are converted to standardized T-scores by procedures described in the manual. (Patients were screened for color blindness before test was administered. 1 Serial 7s. I6 The most common administration procedure is to ask the subject to subtract 7 from 100 and to continue subtracting 7s until stopped by the examiner. ‘Attention and
INCIDENCE
229
OF OCCULT HEAD INJURY IN SCI, Wilmot
concentration, as well as simple calculation skills, are reflected in performance. Raven Matrices (short form). l3 This test consists of a series of 60 plates containing large designs and a series of small designs, one of which the subject is instructed to choose to fit into a cutout in the larger design. Constructed as a culture fair test of general intellectual ability, it requires the subject to detect organizing principles based on pattern matching, series completion, numeric and spatial relationships, and analogous reasoning to obtain the correct response. Norms are available for ages 8 to 65 years. Basic scores are converted to percentiles which may be converted to WAIS-equivalent IQs based on properties of the normal distribution. Test/retest reliability correlations are in the range of 0.70 to 0.90.5 Validity studies have confirmed its usefulness in assessing visuospatial reasoning.*q3 Associate Learning subtests of Wechsler Memory ScaZe’*(two subtests). This procedure exposes subjects to 10 pairs of words (6 pairs with easily associated words, 4 pairs difficult to associate), to be learned in 3 trials. Scoring according to manual instructions gives final credit for only half of the easy pairs but all of the hard pairs learned across the three trials, Delayed recall is tested after 30 minutes of intervening activity. The original standardization was based on 200 subjects ranging in age from 20 to 50 years. Mean scores on Associate Learning have been reported as 15.72 with standard deviation (SD) of 2.8 1 for the age group 20 to 29 years’*; 15.48 with SD 3.48 for ages 30 to 39 years9; and 13.91 with SD 3.12 for ages 40 to 49 years. ‘* These means should be compared with a total possible score of 21. Shipley Hartford5 (two scales). This test is comprised of two subtests, Vocabulary and Abstract Thinking, each timed at 10 minutes. The scale is based on the clinical and experimental observation that mental deterioration is accompanied by discrepancies in measures of vocabulary and the capacity for abstract (conceptual) thinking. The SILS (Shipley Institute of Living Scale) was standardized on 1,046 individuals, including students ranging from fourth grade through college level. No validity studies are reported in the manual. Reliability coefficients, based on 322 army recruits, were found to be 0.87 for the Vocabulary subtest, 0.89 for the Abstraction subtest, and 0.92 for the two combined. Two neuropsychologists shared the responsibility of testing study patients. Inter-rater reliability was checked by having both psychologists test three patients simultaneously but independently. Ratings on a 1 to 6 impairment scale (1 = no impairment, 6 = profound) revealed only one discrepancy of 1 point in 30 ratings between psychologists. The conversion table for transforming raw scores to impairment rating scores on each test is shown in table 1. These criteria were determined based on published standardized values for each scale. Premorbid data were collected on each patient from family and significant others, as well as patient report and scholastic records to determine whether premorbid cognitive dysfunction was present. A patient was considered to have a poor premorbid history if he/she had (1) dropped out of grade school or high school, (2) failed two or more subjects in one semester, or (3) been formally diagnosed as learning disabled or educationally handicapped. The admitting physician’s history and physical examination, including mental status examination, was obtained from the
medical chart to determine whether head injury was suspected. Pre-admission medical records also were reviewed when available. Any hypoxia, epidural or subdural hematoma, or other relevant complication occurring after acute episode but previous to testing was recorded from chart review. Documentation was obtained from the medical chart of any medications the patient was taking at the time of neuropsychologic testing that might affect central nervous system (CNS) functioning, and therefore test results. RESULTS The breakdown of study patients by criteria for selection was as follows: forty-seven patients (70%) met at least criterion one, that is, quadriplegia, high deceleration accident. Fifteen patients (22%) had loss of consciousness at injury. Neurologic indicators were present in 12 (18%) and respiratory support was required in 11 (16%) of the patients. Twenty-five of the 67 patients met more than one criterion. Three methods of assessing impairment of cognition were used: 1. Neuropsychologic impairment was defined by a mean rating of 1.5 or greater across the 11 tests on the 1 to 6 impairment scale. A patient would necessarily score mildly impaired on five of the 11 measures to be considered minimally impaired. See the figure for frequencies of patients in each category of impairment for the combined 11 neuropsychologic tests. Of the 67 patients, 24 (36%) were rated normal and 43 (64%) of the patients were rated as mildly impaired or worse. Most impaired ratings were in the mild category of 1.5 to 2.4 (70%). No patient averaged beyond 4.1 (severe). 2. Clinical judgment of the testing psychologist was recorded. After giving the test battery, but before averaging results, the psychologist indicated in his/her judgment whether the patient had cognitive dysfunction. ‘These clinical assessments did not always agree with mean test scores. Concurrence rate was 8 1%. All six of the ratings by psychologists indicating no impairment that were discrepant with test scores had average impairment scores of 1.5 (borderline impaired). The seven clinical ratings indicating abnormality that were discrepant with test scores had average impairment scores ranging from 1.1 to 1.4. 3. Admission physicals and histories were reviewed for indications of cognitive dysfunction. Seventeen percent (4/24) of those patients whose test scores were normal were considered brain damaged by the physician. All four had good premorbid academic histories. Twenty-three percent (10/43) of those testing impaired were suspected of brain injury on admitting examination. Of those falling within the mildly impaired category, 20% (6/30) of the patients had been reported brain damaged within the physical examination and history. Of those with impairment ratings of three or greater (moderate to severe), five of the seven patients were undiagnosed. Overall, the physician’s admitting opinion was discordant with the neuropsychologic assessment in 55% of cases. All neuropsychologic test raw scores correlated significantly @>O.OOl) with the mean impairment rating. Table 2 presents a correlation matrix showing the relationship among the 11 neuropsychologic evaluations. The Raven Progressive Matrices correlated most highly with all other scales and the ShipArch Phys Med Rehabil Vol66, April 1965
INCIDENCEOF OCCULTHEADINJURYIN SCI, Wilmot
230
Table 2: Correlations Among Neuropsychologic Tests and With Overall Impairment Rating
GOAT
Quick
Shipley Hartford Verbal
Shipley Hartford Abstract
stroop 1
Stroop 2
Stroop 3
Serial IS (errors)
Raven
Wechsler Word
We&+ ler Delay
GOAT Quick
.213*
Shipley Hartford Verbal
,025
.717**
Shipley Hartford Abstract
.315**
.335**
Stroop 1
.269*
.268*
- .104
swoop 2
.335**
.2.56*
,071
.514**
swoop 3
.253*
.284*
,091
.525**
Serial 7s (errors)
- .211
- .286*
.400**
- .392**
Raven
.306**
.605**
Wechsler Word
.585**
.258*
- .004
Wechsler Delay
.618**
.302**
.064
- .622**
- .544**
Overall Impairment
*
<0.05.
.569**
- .449**
.278*
- .591**
.651** .467** - ,228
- ,219
.495**
.346**
.495**
.335**
- .626**
.267*
.268*
.333**
.I 12
- ,236
.290**
.309**
.250*
.433**
.272*
-.174
.256*
- .739**
- .500**
- .648**
- .599**
.636**
- .670**
.793** ~~.571**
- .621**
**
ley Hartford Verbal subscale correlated least. All tests significantly discriminated between those with good and poor premorbid histories except the Stroop, test 1 (a measure of reading speed), and the Wechsler memory subtests. No one test stands out as the strongest comprehensive indicator of overall impairment. Twenty-two patients (33%) were found to have poor premorbid histories. Approximately half of the 22 patients had not graduated from high school, and half had had a formal diagnosis of learning disability or emotional handicap. Mean impairment ratings for those with poor vs good premorbid history were significantly different (t=4.80, p>O.OOOl). A categoric comparison of premorbid history by impairment rating is given in table 3. Only 3 of 24 patients (13%) with a normal rating on neuropsychologic testing had a poor premorbid history. In contrast, 24145 (53%) of those with good premorbid history scored as impaired. This is the group that most clearly indicates the probable incidence of head injury in the SC1 population as a whole. The 24 patients with good premorbid histories who scored impaired on neuropsychologic testing were compared to patients with good premorbid histories and normal scores on number of days to testing. Mean days to testing were 53 and 80 respectively, a difference that was not statistically significant (t = 1.85), with a clear trend toward earlier testing in the impaired group. Criteria for selection for the study also were Table 3: Comparison of Premorbid History and Cognitive Impairment Post Injury Impairment Premorbid
history
Good Poor Chi Square=7.01; Arch
.666** - ,221
rating
< 1.5
3 1.5
21 3
24 19
p
Phys Med Rehabll Vol66, April 1965
compared. Proportionally more of the first group had loss of consciousness (42% vs 25%) at the time of injury, although this difference was not significant (z = 1.27). Further analyses involved the following: 1. A correlation was run on days to testing with overall impairment rating and with each scale score to investigate whether the brain damage may start resolving after a period of time. There was no significant correlation of days to testing with performance on any of the tests individually except Stroop 2 (color identification: r = 0.27, pcO.03) in which longer number of days to testing was related to lesser impairment. No relationship was found with overall impairment rating (1.= - 0.1). 2. Those patients who were taking medications known or believed to affect the CNS (n = 35) were analyzed separately from those not taking CNS medications. No differences in performance were found between groups for any of the tests individually or on overall impairment (t= -0.66). 3. A correlation was computed for length of stay in rehabilitation with the overall impairment score and individual scales. No significant relationship was observed. A similar correlation was run for only the subgroup of patients diagnosed as head injured with good premorbid histories (n = 24) and rehabilitation stay. Again, no significant relationship was found. 4. Level of lesion. A Pearson correlation between degree of overall impairment and neurologic level for complete lesions (n = 41) was not significant (r = 0.18). nor was the correlation of Frankel class and impairment rating. DISCUSSION Almost two-thirds (43/67 or 64%) of the SC1 patients given neuropsychologic evaluations scored as impaired; most of these (30/43 or 70%) fell within the mildly impaired category. Despite the fact that treating physicians were aware of the study, only ten (23%) of those testing impaired were diagnosed
as having
a head injury or cognitive problem by any treating physician previous to neuropsychologic assessment. Only six (20%) of those patients who scored mildly impaired were diagnosed as having a head injury or cognitive problem. Even five of the seven most severely impaired were not diagnosed as such. Yet the implications for treatment and education change considerably given this information. For example, knowledge of cognitive deficits would impact on the patient in several ways: prescription of sedatives, simplification and redundancy in educating patients, provision of repeated orienting information, and greater understanding of irritability, erratic behavior, lack of motivation, and confusion, in patients. An interesting intervening factor in establishing incidence of concomitant head injury in SC1 trauma is premorbid status. Only three patients who had poor premorbid histories scored as normal on neuropsychologic testing, whereas almost half (19 or 44%) of impaired patients had evidence of poor premorbid academic performance. This indicates that impairment ratings for these patients may have been due to premorbid status rather than traumatic head injury. This is substantiated by the fact that neuropsychologic scores were significantly lower for those with poor as opposed to good premorbid histories. Even so, more than one-third (24/67 or 36%) of all patients are suspect for having obtained a head injury at the time of the SCI. This figure does not include those with poor premorbid histories, at least some of whom also may have sustained a head injury in the accident. Level of lesion did not appear to affect degree of cognitive impairment, nor did Frankel classification. Although most patients in the study and on the unit were quadriplegic, paraplegic patients were not significantly less likely to obtain abnormal scores. The lack of relationship between length of rehabilitation stay and overall impairment score is not surprising. Although one might expect that cognitively impaired patients might have a longer stay, the intervening variables dictating length of stay are innumerable and most certainly more influential than mild cognitive problems. Future research will require the following: (1) establishment of local norms on a condensed neuropsychologic test battery to verify that the mean 1.5 division between normal and impaired status is justified, (2) follow-up assessment of impaired individuals having good premorbid histories to ascertain whether neurologic indicators endure months after injury, (3) consensual validation of impairment by an objective measure of neuropathology such as multimodal cortical-evoked potential testing, and (4) compilation of such follow-up data as complications, employment status, and domestic problems, for comparative analysis. Beyond that, a treatment regimen must be established suited to the needs of these cognitively impaired persons, especially considering that their well-being depends on their ability to provide adequate self-care after rehabilitation discharge.
ADDRESS REPRINT REQUESTS TO: Karyl M. Hall, EdD Northern CA Regional Spinal Injury System Institute For Medical Research
2260 Clove Drive San Jose, CA 95128
References 1.
2.
3.
4.
5.
6. 7.
8. 9. 10. 11.
12.
13. 14. 15.
16. 17.
18. 19.
Acknowledgments: Appreciation
is extended to Jim Belvel, MS, and Penelope Hogg, MA, for their help in neuropsychologic testing. Sheilah Sanders assisted in data analysis.
20.
Ammons RB, Ammons CH: Quick test (QT): provisional manual. Psycho1 Rep 11:ll l-161, Monograph Supplement I-VII, 1962 Archibald YM, Wepman JM, Jones LV: Performance on nonverbal cognitive tests following unilateral cortical injury to right and left hemisphere. J Nerv Mental Dis 145:25-36. 1967 Colombo A, De Renzi E, Faglioni P: Occurrence of visual neglect in patients with unilateral cerebral disease. Cortex 12:221231, 1976 Davis D, Bohlman H, Walker AE, Fisher R, Robinson R: Pathological findings in fatal craniospinal injuries. J Neurosurg 34:603613, 1971 Eichorn D: The Raven Progressive Matrices (Review). In Frankenburg WK, Camp BW (eds): Pediatric Screening Tests. Springfield, IL, Charles Thomas, 1975 Golden CJ: Stroop Color and Word Test: Manual for Clinical and Experimental Uses. Chicago, Stoelting Company, 1978 Guttman L: Section 3: Surgical Management. Initial treatment of traumatic paraplegia and tetraplegia. In Harris P (ed): Spinal Injuries. Proceedings of Symposium. Royal College of Surgeons, Edinburgh, June 7-8, 1963, 1967, pp 80-92 Harris P: Associated injuries in traumatic paraplegia and tetraplegia. Paraplegia 5:215-220, 1968 Hulicka IM: Age differences in Wechsler Memory Scale scores. J Genet Psycho1 109: 135-145, 1966 Jennett B, Teasdale G: Management of Head Injuries. Philadelphia, Davis, 1981 Levin HS, O’Donnell VM, Grossman RG: Galveston Orientation and Amnesia Test: practical scale to assess cognition after head injury. J Nerv Mental Dis 167:675-684, 1979 Porter RW: Some problems in management of spinal cord injured patient with associated head or facial trauma. Proceedings of 19th Veterans Administration Spinal Cord Injury Conference, 1973. Feb 1977, pp 188-196 Raven JC: Guide to Standard Progressive Matrices. London, HK Lewis, 1960 Rime1 RW, Giordani B, Barth JT, Boll TJ, Jane JA: Disability caused by minor head injury. Neurosurgery 9:221-228, 1981 Shipley WC: Shipley-Institute of Living Scale for Measuring Intellectual Impairment, Manual of Directions. and Scoring Key. Los Angeles, Western Psychological Services, 1946 Taylor MA: Neuropsychiatric Mental Status Examination. New York, SP Medical & Scientific Books, 1981, pp 215, 265 Wagner KA, Kopaniky DR. Esposito L: Head and spinal cord injured patients: impact of combined sequalae, (abstract). Arch Phys Med Rehabil 64:519, 1983 Wechsler D: Standardized memory scale for clinical use. J Psychol 19:87-95, 1945 Weiss MH: Head trauma and spinal cord injuries: diagnostic and therapeutic criteria. Crit Care Med 2:3 1 l-3 16, 1974 Young JS: Hospital Study Report. Model Systems’ SC1 Digest 1: 1 l-32, Fall 1979
Arch Phys Med Rehabil Vol55, April 1985
Occult Head Injury: Its Incidence in Spinal Cord Injury Conal B. Wilmot, MD, D. Nathan Cope, MD, Karyl M. Hall, EdD, Mary Acker, PhD Santa Clara Valley Medical Center, San Jose. CA 95128 ABSTRACT. Wilmot CB, Cope DN, Hall KM, Acker M: Occult head injury: its incidence in spinal cord injury. Arch Phys Med Rehabil66:227-231, 1985. l This study investigated the suspicion that a significant proportion of individuals having spinal cord injury (SCI) also sustain a concomitant undiagnosed occult head injury during the trauma accident. The criteria for high risk of head injury included the following: (1) quadriplegia with high energy deceleration accident, (2) loss of consciousness at time of injury, (3) brainstem or cortical neurologic indicators, or (4) respiratory support required at time of injury. In this study, 67 patients admitted to the rehabilitation unit were given a neuropsychologic evaluation a median of 48 days after injury. Motor free scales used were the Galveston Orientation and Amnesia Test (GOAT), Quick Test, Raven Progressive matrices, serial 7s, Shipley Hartford, Stroop Color/Word Interference, and the Wechsler Memory Scale Associate Learning Tests. Forty-three of the 67 patients. (64%) scored mildly to profoundly impaired on the test battery. Evidence of poor premorbid academic history was present in 19 (44%) of those with impaired performance on the neurologic evaluation and in only three (13%) of those scoring unimpaired. Consequently, 56% (24/43) of the impaired had no previous record of scholastic difftculties, presumably acquiring cognitive impairment at the time of injury. The implications of this high incidence of impaired cognitive functioning for treatment of individuals with SC1 are significant. KEY WORDS: Brain injuries: Cognition disorders: Head injuries; Spinal cord injuries
At the Northern California regional spinal cord injury (SCI) center, the combination of acute SC1 with head injury leading to subacute neurologic complications has been sufficiently frequent to warrant a formal investigation. Consequent to the cognitive deficits of head injury, significant management and rehabilitation training problems can develop which might better be managed with an appreciation of the occult head injury. Rime1 and associates’4 studied a large sample of individuals with minor head injury having unconsciousness 20 minutes or less. They found that three months after injury, 59% of patients complained of memory loss and 34% with previous employment were now unemployed. Results of neurologic evaluation were normal in nearly all patients. In contrast, neuropsychologic examination was sensitive to problems of attention, concentration, memory, and judgment in these patients. There have been several published reports referring to the high incidence of coexisting multiple injuries in primary diagnoses of head injury and SCI.10.‘7 Weiss” reasoned that anyone receiving a blow severe enough to render him/her unconscious is a candidate for a possible cervical spine injury. PorterI found that in 62 (29%) of 216 nonacute SC1 mortalities there was evidence of significant associated injury to the brain or skull on pathologic examination. These findings concur with those of Guttman7 and Harris.’ Davis and associates4 reported that 46% of autopsied spinal cord injuries had. demonstrable pathology of the brain. Most cases were upper as compared to lower cervical lesions. Cases with lesions below TI were not examined. However, Porter” points out that almost half of his brain injured sample had lower spinal cord injuries (T5-L4). Young2’ recorded major head injuries coincident with paraplegia at 16% and with quadriplegia, 7%. Wagner and associates I7 found that of those with head injuries, complete quadriplegic individuals had severe head injuries as clinically assessed at 24 hours, whereas complete paraplegic individuals had mild injuries. Patients with incomplete lesions also sustained head injuries but these more often were mild in both paraplegic and quadriplegic individuals.
Still to be determined is the actual incidence of head injury in a typical acute SC1 population and whether screening procedures such as neuropsychologic testing would be a sensitive and cost-effective means of diagnosing and assessing the problem. It is our clinical impression that patient management and incidence of neurosurgical complications could both be improved with a greater awareness of the coexistence of these two conditions. The purposes of this study were to: (1) establish an incidence of cognitive dysfunction in the SC1 population; (2) document the degree of cognitive impairment and relate findings to level and completeness of SC1 lesion; (3) investigate the impact of head injury on length of hospital stay; (4) investigate which clinical factors and/or tests best indicate head injury; (5) establish the prevalence of premorbid problems as determined by scholastic performance for comparison with those diagnosed head injured after injury; and (6) assess adequacy of standard admitting history, physical examination, and mental status examination in identifying patients with cognitive dysfunction.
SUBJECTS AND METHODS Subjects All SC1 patients admitted to the regional spinal injury center within one year after injury who met any of the following additional criteria for high risk were candidates for inclusion in the study: (1) quadriplegic with high energy deceleration accident (any significant impact); (2) history of or documented loss of consciousness at time of injury; (3) significant brain-
This research was supported in part by grant GO08202589 to the Northern California Regional Spinal Injury System (Project 128EH20015) from the Rehabilitation Services Administration. U.S. Department of Education. Washington, DC. Submitted for publication February 8. 1984. Accepted in revised form July I I, 1984.
Arch Phys Med Rehabil Vol66,
April 1965
INCIDENCE OF OCCULT HEAD INJURY IN SCI, Wilmot
228
Table 1: Scoring for Neuropsychologic
Test Battery Imoairment
Test l.GOAT
levels
1
2
3
4
5
6
None
Mild
Moderate
Severe
Extreme
Profound
90+
x0-w
70-79
ho-641
95+
K-Y-!
75-M
K-71
50+
-IO-JY
x1-39
X-ZY
4-Y
7
(100minus error points 2. Quick I Q 3-5. Stroop Color Word (T-scores) (Do not use interference score for these impairment
categories) O- I
7
i-3
5-h
7. Raven IQ
95+
x5-94
75-M
65-74
8. Wechsler Word Pairs IQ
95 +
US-94
75-x-1
65-74
0+
l-8
5-h
3-4
6. Serial 7s (errors)
9. Wechsler Delayed Word Pairs (number correct) 10. Shipley Hartford Vocabulary (number correct)
xt
77.75 _.. _.
18-21
I-1-17
11. Shipley Hartford Abstraction (numbercorrect)
77+
‘2-26
16-21
IO- I5
stem or cortical neurologic indicators; and (4) respiratory support required at time of accident. During the study (February to December 1983). 7 1 of 77 trauma patients admitted to the rehabilitation unit met the criteria. Sixty-seven (55 males, 12 females) actually participated. Four were not tested; none refused to participate. Initially. we planned to compare “high risk” patients with those who did not meet criteria for high risk of head injury, but it quickly became evident that almost all SC1 patients qualified as high risk, given the criteria used for this study. Study patients had an average age of 29 years (range: 12 to 69 years). Average length of hospital stay was 116 days (median 115 days). Sample distribution by level of lesion was: 18 high quadriplegics (C2-4), 33 low quadriplegics (C5-8). eight upper thoracic (Tl-7), and eight lower thoracic (T8 and below). Of the 67 patients, 41 (61%) had complete lesions. Patients were given the neuropsychologic evaluation as soon as possible after they were medically stable and could com-
”I
Number of patients in each category of impairment. Arch Phys
Med Rehabil
Vol 66, April 1985
plete the 90-minute testing battery with the psychologist. Patients were tested an average of 63 days after injury with a range of ten to 246 days (median 48 days). Some patients had been hospitalized for a considerable period of time when the study began. thus lengthening average days to testing for the total group.
Methods Tests used for neuropsychologic evaluation excluded procedures dependent upon motor functions. Tests used were: Galveston Orientation and Amnesia Test (GOAT).” This test consists of ten items addressing orientation to person. place. and time. and memory for events. Duration of post-traumatic amnesia, as defined by persistence of defective GOAT scores. was longer in patients with CT-scan evidence of diffuse or bilateral brain injury as compared to cases with focal unilateral lesions. Initial standardization was based on 50 patients who had sustained mild closed head injury. Quick Test. ’ This procedure. which measures vocabulary and simple verbal reasoning, requires the test subject to identify which of four pictures on a card best represents words given by the examiner. Six equivalent forms are available, with 50 words on each form. Mental age norms for levels 1.5 through 19.0 years are provided. along with adult percentiles and Wechsler-type IQs based on the assumption of a mean 1Q of 100 and a standard deviation of 15 IQ points. Stroop Color/Word Test’ (three subtests). This procedure consists of three tasks, each timed at 45 seconds: reading names of colors printed in black on white; identifying colors of capital xs printed in red. blue. and green; and identifying colors in which color names are printed. while suppressing the word itself. Raw scores are the number of stimuli correctly read or identified in each 45-second task. These are converted to standardized T-scores by procedures described in the manual. (Patients were screened for color blindness before test was administered. 1 Serial 7s. I6 The most common administration procedure is to ask the subject to subtract 7 from 100 and to continue subtracting 7s until stopped by the examiner. ‘Attention and
INCIDENCE
229
OF OCCULT HEAD INJURY IN SCI, Wilmot
concentration, as well as simple calculation skills, are reflected in performance. Raven Matrices (short form). l3 This test consists of a series of 60 plates containing large designs and a series of small designs, one of which the subject is instructed to choose to fit into a cutout in the larger design. Constructed as a culture fair test of general intellectual ability, it requires the subject to detect organizing principles based on pattern matching, series completion, numeric and spatial relationships, and analogous reasoning to obtain the correct response. Norms are available for ages 8 to 65 years. Basic scores are converted to percentiles which may be converted to WAIS-equivalent IQs based on properties of the normal distribution. Test/retest reliability correlations are in the range of 0.70 to 0.90.5 Validity studies have confirmed its usefulness in assessing visuospatial reasoning.*q3 Associate Learning subtests of Wechsler Memory ScaZe’*(two subtests). This procedure exposes subjects to 10 pairs of words (6 pairs with easily associated words, 4 pairs difficult to associate), to be learned in 3 trials. Scoring according to manual instructions gives final credit for only half of the easy pairs but all of the hard pairs learned across the three trials, Delayed recall is tested after 30 minutes of intervening activity. The original standardization was based on 200 subjects ranging in age from 20 to 50 years. Mean scores on Associate Learning have been reported as 15.72 with standard deviation (SD) of 2.8 1 for the age group 20 to 29 years’*; 15.48 with SD 3.48 for ages 30 to 39 years9; and 13.91 with SD 3.12 for ages 40 to 49 years. ‘* These means should be compared with a total possible score of 21. Shipley Hartford5 (two scales). This test is comprised of two subtests, Vocabulary and Abstract Thinking, each timed at 10 minutes. The scale is based on the clinical and experimental observation that mental deterioration is accompanied by discrepancies in measures of vocabulary and the capacity for abstract (conceptual) thinking. The SILS (Shipley Institute of Living Scale) was standardized on 1,046 individuals, including students ranging from fourth grade through college level. No validity studies are reported in the manual. Reliability coefficients, based on 322 army recruits, were found to be 0.87 for the Vocabulary subtest, 0.89 for the Abstraction subtest, and 0.92 for the two combined. Two neuropsychologists shared the responsibility of testing study patients. Inter-rater reliability was checked by having both psychologists test three patients simultaneously but independently. Ratings on a 1 to 6 impairment scale (1 = no impairment, 6 = profound) revealed only one discrepancy of 1 point in 30 ratings between psychologists. The conversion table for transforming raw scores to impairment rating scores on each test is shown in table 1. These criteria were determined based on published standardized values for each scale. Premorbid data were collected on each patient from family and significant others, as well as patient report and scholastic records to determine whether premorbid cognitive dysfunction was present. A patient was considered to have a poor premorbid history if he/she had (1) dropped out of grade school or high school, (2) failed two or more subjects in one semester, or (3) been formally diagnosed as learning disabled or educationally handicapped. The admitting physician’s history and physical examination, including mental status examination, was obtained from the
medical chart to determine whether head injury was suspected. Pre-admission medical records also were reviewed when available. Any hypoxia, epidural or subdural hematoma, or other relevant complication occurring after acute episode but previous to testing was recorded from chart review. Documentation was obtained from the medical chart of any medications the patient was taking at the time of neuropsychologic testing that might affect central nervous system (CNS) functioning, and therefore test results. RESULTS The breakdown of study patients by criteria for selection was as follows: forty-seven patients (70%) met at least criterion one, that is, quadriplegia, high deceleration accident. Fifteen patients (22%) had loss of consciousness at injury. Neurologic indicators were present in 12 (18%) and respiratory support was required in 11 (16%) of the patients. Twenty-five of the 67 patients met more than one criterion. Three methods of assessing impairment of cognition were used: 1. Neuropsychologic impairment was defined by a mean rating of 1.5 or greater across the 11 tests on the 1 to 6 impairment scale. A patient would necessarily score mildly impaired on five of the 11 measures to be considered minimally impaired. See the figure for frequencies of patients in each category of impairment for the combined 11 neuropsychologic tests. Of the 67 patients, 24 (36%) were rated normal and 43 (64%) of the patients were rated as mildly impaired or worse. Most impaired ratings were in the mild category of 1.5 to 2.4 (70%). No patient averaged beyond 4.1 (severe). 2. Clinical judgment of the testing psychologist was recorded. After giving the test battery, but before averaging results, the psychologist indicated in his/her judgment whether the patient had cognitive dysfunction. ‘These clinical assessments did not always agree with mean test scores. Concurrence rate was 8 1%. All six of the ratings by psychologists indicating no impairment that were discrepant with test scores had average impairment scores of 1.5 (borderline impaired). The seven clinical ratings indicating abnormality that were discrepant with test scores had average impairment scores ranging from 1.1 to 1.4. 3. Admission physicals and histories were reviewed for indications of cognitive dysfunction. Seventeen percent (4/24) of those patients whose test scores were normal were considered brain damaged by the physician. All four had good premorbid academic histories. Twenty-three percent (10/43) of those testing impaired were suspected of brain injury on admitting examination. Of those falling within the mildly impaired category, 20% (6/30) of the patients had been reported brain damaged within the physical examination and history. Of those with impairment ratings of three or greater (moderate to severe), five of the seven patients were undiagnosed. Overall, the physician’s admitting opinion was discordant with the neuropsychologic assessment in 55% of cases. All neuropsychologic test raw scores correlated significantly @>O.OOl) with the mean impairment rating. Table 2 presents a correlation matrix showing the relationship among the 11 neuropsychologic evaluations. The Raven Progressive Matrices correlated most highly with all other scales and the ShipArch Phys Med Rehabil Vol66, April 1965
INCIDENCEOF OCCULTHEADINJURYIN SCI, Wilmot
230
Table 2: Correlations Among Neuropsychologic Tests and With Overall Impairment Rating
GOAT
Quick
Shipley Hartford Verbal
Shipley Hartford Abstract
stroop 1
Stroop 2
Stroop 3
Serial IS (errors)
Raven
Wechsler Word
We&+ ler Delay
GOAT Quick
.213*
Shipley Hartford Verbal
,025
.717**
Shipley Hartford Abstract
.315**
.335**
Stroop 1
.269*
.268*
- .104
swoop 2
.335**
.2.56*
,071
.514**
swoop 3
.253*
.284*
,091
.525**
Serial 7s (errors)
- .211
- .286*
.400**
- .392**
Raven
.306**
.605**
Wechsler Word
.585**
.258*
- .004
Wechsler Delay
.618**
.302**
.064
- .622**
- .544**
Overall Impairment
*
<0.05.
.569**
- .449**
.278*
- .591**
.651** .467** - ,228
- ,219
.495**
.346**
.495**
.335**
- .626**
.267*
.268*
.333**
.I 12
- ,236
.290**
.309**
.250*
.433**
.272*
-.174
.256*
- .739**
- .500**
- .648**
- .599**
.636**
- .670**
.793** ~~.571**
- .621**
**
ley Hartford Verbal subscale correlated least. All tests significantly discriminated between those with good and poor premorbid histories except the Stroop, test 1 (a measure of reading speed), and the Wechsler memory subtests. No one test stands out as the strongest comprehensive indicator of overall impairment. Twenty-two patients (33%) were found to have poor premorbid histories. Approximately half of the 22 patients had not graduated from high school, and half had had a formal diagnosis of learning disability or emotional handicap. Mean impairment ratings for those with poor vs good premorbid history were significantly different (t=4.80, p>O.OOOl). A categoric comparison of premorbid history by impairment rating is given in table 3. Only 3 of 24 patients (13%) with a normal rating on neuropsychologic testing had a poor premorbid history. In contrast, 24145 (53%) of those with good premorbid history scored as impaired. This is the group that most clearly indicates the probable incidence of head injury in the SC1 population as a whole. The 24 patients with good premorbid histories who scored impaired on neuropsychologic testing were compared to patients with good premorbid histories and normal scores on number of days to testing. Mean days to testing were 53 and 80 respectively, a difference that was not statistically significant (t = 1.85), with a clear trend toward earlier testing in the impaired group. Criteria for selection for the study also were Table 3: Comparison of Premorbid History and Cognitive Impairment Post Injury Impairment Premorbid
history
Good Poor Chi Square=7.01; Arch
.666** - ,221
rating
< 1.5
3 1.5
21 3
24 19
p
Phys Med Rehabll Vol66, April 1965
compared. Proportionally more of the first group had loss of consciousness (42% vs 25%) at the time of injury, although this difference was not significant (z = 1.27). Further analyses involved the following: 1. A correlation was run on days to testing with overall impairment rating and with each scale score to investigate whether the brain damage may start resolving after a period of time. There was no significant correlation of days to testing with performance on any of the tests individually except Stroop 2 (color identification: r = 0.27, pcO.03) in which longer number of days to testing was related to lesser impairment. No relationship was found with overall impairment rating (1.= - 0.1). 2. Those patients who were taking medications known or believed to affect the CNS (n = 35) were analyzed separately from those not taking CNS medications. No differences in performance were found between groups for any of the tests individually or on overall impairment (t= -0.66). 3. A correlation was computed for length of stay in rehabilitation with the overall impairment score and individual scales. No significant relationship was observed. A similar correlation was run for only the subgroup of patients diagnosed as head injured with good premorbid histories (n = 24) and rehabilitation stay. Again, no significant relationship was found. 4. Level of lesion. A Pearson correlation between degree of overall impairment and neurologic level for complete lesions (n = 41) was not significant (r = 0.18). nor was the correlation of Frankel class and impairment rating. DISCUSSION Almost two-thirds (43/67 or 64%) of the SC1 patients given neuropsychologic evaluations scored as impaired; most of these (30/43 or 70%) fell within the mildly impaired category. Despite the fact that treating physicians were aware of the study, only ten (23%) of those testing impaired were diagnosed
as having
a head injury or cognitive problem by any treating physician previous to neuropsychologic assessment. Only six (20%) of those patients who scored mildly impaired were diagnosed as having a head injury or cognitive problem. Even five of the seven most severely impaired were not diagnosed as such. Yet the implications for treatment and education change considerably given this information. For example, knowledge of cognitive deficits would impact on the patient in several ways: prescription of sedatives, simplification and redundancy in educating patients, provision of repeated orienting information, and greater understanding of irritability, erratic behavior, lack of motivation, and confusion, in patients. An interesting intervening factor in establishing incidence of concomitant head injury in SC1 trauma is premorbid status. Only three patients who had poor premorbid histories scored as normal on neuropsychologic testing, whereas almost half (19 or 44%) of impaired patients had evidence of poor premorbid academic performance. This indicates that impairment ratings for these patients may have been due to premorbid status rather than traumatic head injury. This is substantiated by the fact that neuropsychologic scores were significantly lower for those with poor as opposed to good premorbid histories. Even so, more than one-third (24/67 or 36%) of all patients are suspect for having obtained a head injury at the time of the SCI. This figure does not include those with poor premorbid histories, at least some of whom also may have sustained a head injury in the accident. Level of lesion did not appear to affect degree of cognitive impairment, nor did Frankel classification. Although most patients in the study and on the unit were quadriplegic, paraplegic patients were not significantly less likely to obtain abnormal scores. The lack of relationship between length of rehabilitation stay and overall impairment score is not surprising. Although one might expect that cognitively impaired patients might have a longer stay, the intervening variables dictating length of stay are innumerable and most certainly more influential than mild cognitive problems. Future research will require the following: (1) establishment of local norms on a condensed neuropsychologic test battery to verify that the mean 1.5 division between normal and impaired status is justified, (2) follow-up assessment of impaired individuals having good premorbid histories to ascertain whether neurologic indicators endure months after injury, (3) consensual validation of impairment by an objective measure of neuropathology such as multimodal cortical-evoked potential testing, and (4) compilation of such follow-up data as complications, employment status, and domestic problems, for comparative analysis. Beyond that, a treatment regimen must be established suited to the needs of these cognitively impaired persons, especially considering that their well-being depends on their ability to provide adequate self-care after rehabilitation discharge.
ADDRESS REPRINT REQUESTS TO: Karyl M. Hall, EdD Northern CA Regional Spinal Injury System Institute For Medical Research
2260 Clove Drive San Jose, CA 95128
References 1.
2.
3.
4.
5.
6. 7.
8. 9. 10. 11.
12.
13. 14. 15.
16. 17.
18. 19.
Acknowledgments: Appreciation
is extended to Jim Belvel, MS, and Penelope Hogg, MA, for their help in neuropsychologic testing. Sheilah Sanders assisted in data analysis.
20.
Ammons RB, Ammons CH: Quick test (QT): provisional manual. Psycho1 Rep 11:ll l-161, Monograph Supplement I-VII, 1962 Archibald YM, Wepman JM, Jones LV: Performance on nonverbal cognitive tests following unilateral cortical injury to right and left hemisphere. J Nerv Mental Dis 145:25-36. 1967 Colombo A, De Renzi E, Faglioni P: Occurrence of visual neglect in patients with unilateral cerebral disease. Cortex 12:221231, 1976 Davis D, Bohlman H, Walker AE, Fisher R, Robinson R: Pathological findings in fatal craniospinal injuries. J Neurosurg 34:603613, 1971 Eichorn D: The Raven Progressive Matrices (Review). In Frankenburg WK, Camp BW (eds): Pediatric Screening Tests. Springfield, IL, Charles Thomas, 1975 Golden CJ: Stroop Color and Word Test: Manual for Clinical and Experimental Uses. Chicago, Stoelting Company, 1978 Guttman L: Section 3: Surgical Management. Initial treatment of traumatic paraplegia and tetraplegia. In Harris P (ed): Spinal Injuries. Proceedings of Symposium. Royal College of Surgeons, Edinburgh, June 7-8, 1963, 1967, pp 80-92 Harris P: Associated injuries in traumatic paraplegia and tetraplegia. Paraplegia 5:215-220, 1968 Hulicka IM: Age differences in Wechsler Memory Scale scores. J Genet Psycho1 109: 135-145, 1966 Jennett B, Teasdale G: Management of Head Injuries. Philadelphia, Davis, 1981 Levin HS, O’Donnell VM, Grossman RG: Galveston Orientation and Amnesia Test: practical scale to assess cognition after head injury. J Nerv Mental Dis 167:675-684, 1979 Porter RW: Some problems in management of spinal cord injured patient with associated head or facial trauma. Proceedings of 19th Veterans Administration Spinal Cord Injury Conference, 1973. Feb 1977, pp 188-196 Raven JC: Guide to Standard Progressive Matrices. London, HK Lewis, 1960 Rime1 RW, Giordani B, Barth JT, Boll TJ, Jane JA: Disability caused by minor head injury. Neurosurgery 9:221-228, 1981 Shipley WC: Shipley-Institute of Living Scale for Measuring Intellectual Impairment, Manual of Directions. and Scoring Key. Los Angeles, Western Psychological Services, 1946 Taylor MA: Neuropsychiatric Mental Status Examination. New York, SP Medical & Scientific Books, 1981, pp 215, 265 Wagner KA, Kopaniky DR. Esposito L: Head and spinal cord injured patients: impact of combined sequalae, (abstract). Arch Phys Med Rehabil 64:519, 1983 Wechsler D: Standardized memory scale for clinical use. J Psychol 19:87-95, 1945 Weiss MH: Head trauma and spinal cord injuries: diagnostic and therapeutic criteria. Crit Care Med 2:3 1 l-3 16, 1974 Young JS: Hospital Study Report. Model Systems’ SC1 Digest 1: 1 l-32, Fall 1979
Arch Phys Med Rehabil Vol55, April 1985