Anosognosia in Alzheimer's disease: A neuropsychological approach

Anosognosia is often found in Alzheimer's disease (AD), but its relationship with cognitivebehavioral changes is not well established. Objective To verify if anosognosia is related to cognitive-behavioral disturbances, and to regional brain dysfunction as evaluated by neuroimaging. Methods We included AD patients with Mini-Mental State Examination (MMSE) scores of 12 through 24, and Clinical Dementia Rating (CDR) scores of 1 or 2. Dementia diagnosis was based on DSM-IV and NINCDS-ADRDA criteria.We used Self-Consciousness Questionnaire (SCQ) and Denial of Illness Scale (DIS), and following neuropsychological counterproofs: WAIS-R digit span, Rey auditory verbal learning, verbal fluency test (category: animals), Cummings' neuropsychiatric inventory (NPI) and Cornell scale for depression in dementia (CSDD). Results We studied 21 patients (12 men, 9 women) with AD (14 mild, 7 moderate), age 72.4±8.5 years, education 4.9± 4.2 years, and MMSE score 18.2±5. SCQ and DIS did not correlate to age, education, or regional cerebral perfusion defects, but they tended to correlate to disease duration (and only SCQ also to MMSE). SCQ and DIS were correlated neither to CSDD, NPI, CDR, nor to any neuropsychological test. Significant correlations were found between SCQ and DIS, as well as between SCQ domain of “moral judgment” and MMSE. Conclusion SCQ and DIS were not correlated to age, education, disease duration, cognitive-behavioral measures, dementia severity, or regional cerebral perfusion defects, but were correlated to each other, suggesting SCQ and DIS evaluate similar mental functions.

Anosognosia (Gk. "gnosis", knowledge; "nosos", disease) has been defined broadly as "apparent unawareness, misinterpretation, or explicit denial of illness" 1 , or as impaired insight for behavioral and cognitive problems. Other authors2 have approached anosognosia as an impairment of self-consciousness (SC). SC is then conceived as the process by which the subject becomes the object of their own awareness by realizing that they are perceiving something (their own body or the outside world) or reflecting on their own history (autobiography) or projects 2,3 . SC is thus a complex mental function dependent on memory and other cognitive functions, comprising different aspects or degrees, which have to be taken into account in evaluation 2 .
Anosognosia has often been reported in Alzheimer´s disease (AD), with a prevalence ranging between 15% and 25% 4,5 . AD involves pervasive changes of attention, perception, memory, humor, and personality, and whose relation with anosognosia is not well understood.
Even the relationship between anosognosia and dementia severity has been inconsistent, partly because other confounding variables (e.g., depression) have not been taken into account. Smith et al. 6 have found a positive correlation between dementia severity and degree of anosognosia after controlling for depressive symptoms, while other authors 7 have found similar correlation even without controlling for these symptoms.
As regards the role of depression, various authors 5,6,8-10 have found it to be negatively correlated to anosognosia in AD patients, while others have not 11 . Migliorelli et al. 9 found that patients with depressive symptoms (dysthymia) had lower anosognosia scores (i.e., greater insight) than patients with AD who had either major depression or no depression. Therefore, it is important to distinguish between major depression and depressive symptoms when analyzing anosognosia 13 .
The relationship between anosognosia and regional brain dysfunction is not well established. Anosognosia has been associated to right-hemisphere lesions involving the parietal and temporal lobes, thalamus, and basal ganglia 13,14 , as well as the frontal lobes [15][16][17] . According to Starkstein & Robinson 18 in a study of stroke patients, the presence of neglect and frontal-subcortical dysfunction may constitute important predisposing factors. Furthermore, Reed et al. 4 and Starkstein et al. 19 have found anosognosia associated with a deficiency in the perfusion of the dorsolateral portion of the right frontal lobe. Thus, Lopez et al. 15 and Stuss & Benson 20 suggest that the frontal lobes play a relevant role in self-consciousness and monitoring of cognitive tasks. As such, anosognosia would result from a deficit in auto-monitoring due to frontal lobe dysfunction. However, Starkstein et al. 12 detected no significant difference between AD patients with and without anosognosia, as regards their performance in tests of executive function.
The aim of this study was to verify relationships between anosognosia and cognitive deficits, depressive symptoms, behavioral disturbances, and regional cerebral blood flow in patients with mild to moderate AD. Our hypotheses were that anosognosia would be more frequent and severe (1) the longer the duration of the disease, (2) the more widespread the brain lesions, (3) the more severe the cognitive deficits and dementia, and (4) the lower the depressive scores.

Methods
We included patients with probable AD consecutively attended at our university hospital, aged 45 to 95 years, and presenting scores from 12 to 24 on Mini-Mental State Examination (MMSE) 21,22 , and scores 1 or 2 on Clinical Dementia Rating (CDR) 23 . Exclusion criteria were any clinically significant cardiac, pulmonary, hepatic or renal disease, chronic exposition to neurotoxic compounds, or previous head trauma with loss of consciousness. Major depressive disorder as a cause of dementia syndrome was excluded when not fulfilling DSM-IV criteria 24 for depression or when cognitive complaints had not improved after satisfactory treatment of depressive symptoms. All patients gave their informed consent to participate, in accordance with the rules of our Medical School Ethics Committee.
All patients provided a medical history, and underwent physical, neurological, and neuropsychological examination. Diagnosis of dementia was based on DSM-IV criteria 24 , as well as on NINCDS-ADRDA 25 for AD. Computed tomography (CT), magnetic resonance imaging (MRI), cerebral blood flow imaging (SPECT tomography using technetium-99m-HMPAO), electroencephalography, cerebrospinal fluid analysis, and relevant laboratory blood tests were performed to rule out other causes of dementia. SPECT images were interpreted by noting the location, extent and severity of the perfusion defects, and subsequently classified into the following perfusion patterns, according do Holman et al. 26 : A, normal; B, bilateral posterior temporal and/or parietal cortex defects; C, bilateral posterior temporal and/or parietal cortex defects with additional defects; D, unilateral posterior temporal and/or parietal cortex defects with or without additional defects; E, frontal cortex defects only; F, other large (>7 cm) defects; G, multiple small (≤7 cm) cortical defects.
Neuropsychological investigation comprised MMSE, WAIS-R Digit Span for attention 27 , Verbal Fluency test (VF; category: animals´ names in one minute) for executive functions 28 , Rey Auditory Verbal Learning test for memory (RAVLT) 28,29 , Neuropsychiatric Inventory (NPI) 30 , and Cornell Scale for Depression in Dementia (CSDD) 31 . CSDD was used to quantify depressive symptoms in our dementia patients.
Assessment of anosognosia was carried out using the Self-Consciousness Questionnaire 2 and the Denial of Illness Scale 18 (see Appendix). The Self-Consciousness Questionnaire (SCQ) has fourteen questions, four of them concerning "identity" (Nos. 1, 5, 6, 7), three "knowledge of cognitive disturbances" (metamemory or metacognition: Nos. 2, 3, 4), one "self-evaluation of the affec-tive state" (No. 8), two "knowledge about representation of the body" (Nos. 9, 10), one on "anticipation" (prospective memory: No. 11), one "capacities for introspection" (No. 12), and two "moral judgment" (Nos. 13,14). The answers were classified as relevant or correct (two points), incorrect (no points), or partly correct (one point). The higher the score, the greater the degree of self-consciousness. The score obtained for each of these aspects of consciousness was divided by the number of questions corresponding to each aspect, giving a total maximum score of 14 points. The answers were checked with the patient´s caregiver. The Denial of Illness Scale (DIS) consists of ten items judged by the examiner and used to classify the patients into those with mild, moderate, or severe anosognosia. In DIS, the higher the score (0, 1 or 2) on each item, patient refusal. SCQ scores were not related with age, education, or regional perfusion defects found on SPECT images, but tended to correlate with disease duration and MMSE scores, though did not reach statistical significance (r= -0.324 and r=0.317, respectively). In the analysis of correlation between SCQ scores and SPECT patterns (A, B, C, D, E, G; see Table 2), SCQ scores were classified into two subgroups: from 7 to 10.99, and from 11 to 14, since the minimum score was 7.75 and the maximum score, 14. Likewise, DIS scores did not show any correlation to age, education, MMSE, or SPECT data, but tended to correlate to disease duration (r=0.3208), though did not reach statistic significance. SCQ and DIS scores did not correlate with mood state (CSDD), behavioral changes (NPI), dementia severity (CDR 1 and 2), nor with any of the neuropsychological the greater the degree of anosognosia. SCQ and DIS were translated to Portuguese by the authors, since these scales had not been published or validated in Brazil. Data analysis by means of Statistica software 6.0 (StatSoft Inc., 2001) used contingency tables (chi-square) and Pearson coefficient for correlations between anosognosia items and neuropsychological tests (counterproofs). The significance level was 5% (two-tailed).

Results
We studied 21 patients (12 men, 9 women) with probable AD (14 mild and 7 moderate), age 72.4±8.5 years (mean±standard deviation), education 4.9±4.2 years, and score on MMSE 18.2±5.0 (Table 1). CT was performed in all patients and MRI in 12. SPECT could not be performed in 5 patients because of operational difficulties or tests (digit span, verbal fluency, Rey verbal learning). Also, there were no correlations between the different aspects of self-consciousness and any of these cognitive and behavioral variables. The only significant correlations were found between SCQ and DIS scores (r= -0.4482, t=2.185, df=19, p<0.05) and between the domain of "moral judgment" of SCQ and MMSE (r=0.4629, t= 2.2763, df=19, p<0.05).

Discussion
In agreement with other authors, wee have found no correlations between SCQ or DIS scores and age, education, and disease duration 2 . However, in contrast with these authors, we found no correlation either between SCQ or DIS, and dementia severity as measured by CDR or MMSE scores. As expected, there was a negative correlation between the scores of SCQ and those of DIS, suggesting they measure similar mental changes.
The analysis of each aspect of self-consciousness relative to the neuropsychological (MMSE, verbal learning, digit span, verbal fluency) or neuropsychiatric variables (CSDD, NPI) showed significant positive correlation only between the domain of "moral judgment" and MMSE, in agreement with Gil et al. 2 . This finding is difficult to interpret. It may simply be a statistical artefact without clinical meaning, since changes of moral judgment are usually related to frontal-orbital regions, and only one of our patients had bifrontal hypoperfusion, but without MRI signs of frontal atrophy. Most of our patients had atrophy and hypoperfusion in posterior regions (temporal, parietal, occipital), often bilaterally. Recent studies 32 have shown moral judgment ability to be a complex function, which requires a whole neurofunctional network, including posterior associative brain regions and integrating semantic-cultural knowledge and appropriate motivational and affective states.
We hypothesized that anosognosia would be worse the longer the disease duration, the more widespread the brain lesions, the more severe the dementia, and the lower the depressive scores. However, none of these hypotheses were confirmed by our findings, probably because our sample size was small and comprised mostly mild dementia cases (67%), without any patients with severe dementia (CDR 3). The lesion model we used (Alzheimer´s disease, with degeneration predominantly in temporal-parietal regions) did not allow us to verify the relevant role of frontal dysfunctions in anosognosia, as established by various authors 4,15,16 . Michon et al. 17 was indeed able to verify that the severity of anosognosia is related to signs of frontal dysfunction but not to the severity of dementia. Thus, in order to tackle these questions, we need to include a greater number of patients with Alzheimer´s disease with CDR 1 through 3, to compare to patients with frontotemporal dementia and normal control subjects, while also using neuroimaging methods robust enough to more precisely delimit the brain lesions or dysfunctions.