The Effect of Major Depression on Subjective and Objective Cognitive Deficits in Mild to Moderate Traumatic Brain Injury

Cognitive deficits following traumatic brain injury (TBI) include impairments in attention, memory and executive functioning.^1–7 Patients who perform poorly on various batteries of neurocognitive testing often complain of these difficulties.^8–10 However, the association between self-reported cognitive symptoms and performances on objective cognitive tests has not been consistently supported by empirical data. Some investigators have concluded that subjective cognitive complaints were not always related to neuropsychological test performance^11–13 but rather to emotional difficulties.^14,15 However, other studies have failed to find an association between objective measures of cognitive dysfunction and emotional distress.^16,17 These conflicting results may be due to differing methods of evaluating self-report symptoms and objective cognitive impairment, the heterogeneity of TBI subjects in terms of trauma severity and varying periods of assessment following head injury.

Despite the mounting evidence suggesting a strong association between clinically diagnosed major depression and poor performance on cognitive testing following mild to moderate TBI,^18–23 the few studies^14,16 that have investigated the relationship between self-reported cognitive difficulties, neuropsychological test performances and emotional complaints, have relied mainly on subjective measures of depression or on instruments that assess psychological profiles in order to detect comorbid depressive symptoms. Additionally, they have neglected to control for depression when investigating whether subjective cognitive complaints accurately reflected cognitive dysfunction on objective testing.

The focus of the current study was to investigate the relationship between cognitive impairment assessed by standard neuropsychological tests and subjective cognitive complaints based on a self-report questionnaire in mild to moderate TBI 6 months following head injury. The effect of clinically diagnosed major depression on subjective (self-report) and objective (neuropsychological testing) indices of cognitive impairment was also explored.

METHOD

Patient Selection

The study population consisted of 63 TBI patients who were recruited consecutively from a traumatic brain injury clinic at a tertiary care referral center. The subjects had either sustained a mild head injury²⁴ (Glasgow Coma Scale [GCS]=13–15; loss of consciousness [LOC] <20 minutes; posttraumatic amnesia [PTA] < 24 hours) or moderate TBI (GCS=9–12; PTA>24 hours but less than 1 week) and were between 18 to 60 years of age. All participants had a detailed neuropsychiatry examination at 6 months posthead injury before undergoing a battery of cognitive tests. The presence of posttraumatic “subjective cognitive” difficulties was determined by patients’ response on three questions of the Rivermead Post Concussion Symptoms Questionnaire (RPQ): “forgetfulness, poor memory”, “poor concentration” and “taking more time to think”. These selected items are the only questions from the RPQ that focus on cognition. The RPQ has been shown to be a valid measure of symptoms commonly experienced following head injury.²⁵

Written consent was provided from all subjects and institutional ethics board approval was obtained for this study.

Data Collection

Background information.

All subjects were seen 6 months following a mild to moderate TBI. The collected demographic and TBI-related data were: age, gender, marital and preinjury employment status, educational level, problem drinking (based upon the CAGE questionnaire²⁶ and amount of alcohol consumed per week), past psychiatric history, prior head injury and mechanism of injury. Psychotropic and/or analgesic medication use was also recorded given the CNS depressant properties of these drugs potentially impairing cognition. For head injury severity, the following were obtained: initial GCS recorded in the emergency room, LOC and PTA duration, and initial CT brain results.

Diagnosis of Major Depression

The study participants were interviewed with the mood disorder section of the Structured Clinical Interview for DSM–IV Disorders (SCID²⁷ for DSM–IV) to establish a diagnosis of major depression. The clinic’s neuropsychiatrist who conducted the interview was blind to the subjects’ cognitive data.

Neuropsychological Tests

The neuropsychological battery comprised 10 tests known to be sensitive to cognitive changes following TBI. They included indices of attention/working memory (Weschsler Adult Intelligence Scale-III [WAIS-III]²⁸—working memory, verbal memory (California Verbal Learning Test [CVLT-II]²⁹—total, long delay free recall and recognition hits), visuospatial learning and memory (Brief Visuospatial Memory Test—Revised [BVMT-R]³⁰—immediate and delay total recall), information processing speed and sustained and divided attention (Paced Auditory Serial Addition Task [PASAT]³¹ at increasingly quicker rate of number presentation: 2.0 and 1.2 sec.), executive functioning (Wisconsin Card Sorting Test [WCST]³²—total and perseverative responses). The Vocabulary subscale of the Wechsler Abbreviated Intelligence Scale (WASI)³³ was used to provide an estimate of the premorbid IQ and the Word Memory Test (WMT)³⁴ was administered to assess subject compliance and effort with testing.

Statistical Analysis

Univariate analyses.

Patients with and without subjective cognitive complaints were compared using two-tailed t tests for continuous variables, and χ² analyses for categorical variables. Fisher’s exact test was reported when appropriate. For the cognitive data, raw scores were used for analysis.

Multivariate analyses.

A multivariate analysis of covariance (MANCOVA) was performed using the variables found statistically significant (p≤0.05) on univariate analyses. Subject grouping was designated as the fixed factor and seven of the cognitive tasks were entered as dependent variables. Two separate MANCOVAs were then computed across groups for each of the cognitive measures, one adjusting for gender and injury severity as measured by the Glasgow Coma Scale (GCS) score and the other for depression (SCID for DSM–IV) in addition to gender and GCS score.

RESULTS

Demographic and Injury-Related Data

The mean age of the 63 patients enrolled in the study was 33 years (SD=11.7) and 55.6% were male. Thirty-four (54%) subjects reported subjective cognitive complaints on the RPQ. There were no differences in terms of demographic and injury-related variables between the two groups, except that those with subjective cognitive complaints had a higher GCS (t=−2.8, df=35, p=0.007) and were more likely to be women (χ²=6.2, df=1, p=0.01) (Table 1).

Major Depression

Based on the SCID for DSM–IV, major depression occurred in 18.5% of the group with subjective cognitive complaints but in none of those without self-reported cognitive difficulties. The difference between the two groups was significant (Fisher’s exact test, p=0.05) (Table 1).

Neurocognitive Deficits

While both groups had similar scores on their premorbid IQ and on the WMT (Table 1), patients with subjective complaints scored significantly more poorly on measures of working memory (WAIS-III) (t=2.7, df=60, p=0.01), verbal (CVLT-II recognition hits) (t=3.4, df=42, p=0.001) and visuospatial memory (BVMT-R immediate [t=2.9, df=59, p=0.005] and delayed total recall [t=3.2, df=54, p=0.002]), attention and information processing speed (PASAT 2.0 s) (t=2.2, df=57, p=0.03), and executive functioning (WCST total [t=2.4, df=51, p=0.02] and perseverative responses [t=−2.9, df=56, p=0.005]). There were no significant group differences on the CVLT-II total (t=1.7, df=61, p=0.09) and long delay free recall (t=1.4, df=53, p=0.18) or on the PASAT at 1, 2 presentation interval (t=1.8, df=56, p=0.07).

When using MANCOVA on cognitive variables found statistically significant during univariate analysis, with gender and GCS score serving as covariates, the differences between the groups remained significant on almost all cognitive measures: WAIS-III working memory (F=4.5, df=1,54, p=0.04), CVLT-II recognition hits (F=15.2, df=1,54, p<0.0001), BVMT-R immediate (F=5.9, df=1,54, p=0.02) and delayed total recall (F=8.2, df=1,54, p=0.006), and WCST perseverative responses (F=7.2, df=1,54, p=0.009). In a separate MANCOVA, adjusted for depression in addition to gender and GCS score, those with subjective cognitive complaints scored significantly more poorly only on CVLT-II recognition hits (F=7.1, df=1,54, p=0.01), BVMT-R delayed total recall (F=4.8, df=1,54, p=0.03) and WCST perseverative responses (F=4.3, df=1,54, p=0.04) (Table 2).

DISCUSSION

We demonstrated that patients with mild to moderate TBI, who reported cognitive difficulties at 6 months following injury, performed more poorly on objective neuropsychological testing than those without subjective cognitive complaints. This finding concurs with certain previous mild TBI studies^8–10 and has also been described in patients with HIV^35,36 and in subjects with multiple sclerosis.^37,38

The potential contribution of other factors such as prior head injury and psychotropic and/or analgesic medication intake to subjects’ performances on cognitive testing was also assessed. Lack of statistically significant differences between the two groups on all these variables eliminated the possibility of these factors influencing cognitive performance. Also, no differences were noted in terms of CT scan results, a finding that is significant given the association between neuropsychological deficits and CT scan abnormalities in patients with these injuries.³⁹

Our study also clarified contradictory findings from previous investigations^14,16 where the influence of psychological distress on cognition was evaluated based solely on self-report instruments of depression. Once major depression (according to DSM–IV criteria) was controlled for, the differences between those with and without subjective cognitive complaints disappeared on most cognitive tests, indicating a close association between objective measure of mood (SCID-IV) and several aspects of cognition.

However, depression did not account for all subjective reporting of cognitive difficulties in our study since, even after adjusting for low mood, significant differences remained between those with and without subjective cognitive complaints. Some of these individuals with self-reported cognitive difficulties may be part of the “miserable minority,” a term coined by Ruff et al.⁴⁰ to describe mild TBI patients whose recovery is compromised by psychological factors other than depression, such as premorbid personality traits. Recent data from functional imaging studies^41–45 suggest an alternative hypothesis—namely that, in certain patients with mild TBI, cognitive complaints may relate directly to cerebral dysfunction that can be unmasked by neuroimaging activation paradigms. Ruff et al.⁴¹ studied nine mild head injury patients with persistent complaints and measurable cognitive deficits, who underwent 18-Fluoro-deoxyglucose (FDG) PET, a mean of 29 months postaccident. Despite normal MRI/CT scans, the TBI subjects, when matched against healthy comparison subjects, demonstrated frontal and temporal hypometabolism while performing a visual sustained attention task. Alterations at the level of cerebral substrate have been documented even in acute cases of mild TBI (i.e., within 1 month of injury) using the fMRI approach. In response to increasing working memory (WM) processing load^42–43 as well as to tasks that probe episodic memory,⁴⁴ subjects with mild head injury have shown significant differences in brain activation patterns, relative to healthy comparison subjects, differences that may explain mild TBI patients’ persistent cognitive complaints and need to exert greater effort during cognitive tasks.⁴⁵

CONCLUSION

In summary, mild to moderate TBI patients with persisting subjective cognitive complaints have demonstrable evidence of cognitive dysfunction. In most, but not all patients, these objective cognitive difficulties are linked to comorbid major depression. This clinically important observation suggests that a detailed assessment of mood based on objective criteria should be considered in all patients who complain of cognitive problems following such injuries.

ACKNOWLEDGMENTS

Dr. Feinstein is supported by grants from the Canadian Institutes of Health Research (grant #36535)

The poster presentation of this study was given at the 15th Annual Meeting of the American Neuropsychiatric Association in Bal Harbour, Fl., February 21-24, 2004.

TABLE 1. Demographic, Injury and Psychological Variables Compared Between Patients With and Without Subjective Cognitive Complaints

Enlarge table

TABLE 2. Neurocognitive Test Performances Compared Between Patients With and Without Subjective Cognitive Complaints and Resulting MANCOVAs Adjusted For Gender, GCS Score, and Major Depression

Enlarge table