Demographic, neuropsychologic and routine radiologic data
We included 19 consecutive euthymic, elderly patients with bipolar disorder (11 with type I and 8 with type II bipolar disorder) and 47 healthy controls in this prospective study. All participants were right-handed and had normal or corrected-to-normal visual acuity. None reported a history of sustained head injury or neurologic or psychiatric disorders.
There were no significant differences in age, education and Charlson Cormorbidity Index41
score between patients with bipolar disorder and controls. There was a significantly higher percentage of women among the patients than the controls. The scores on the Geriatric Depression Scale (< 4/15 in all patients) and the Young Mania Rating Scale (< 3/44 in all patients) confirmed the patients’ euthymic mood state. Duration of illness was quite long in all patients, and the mean age at onset was 39.4 (standard deviation [SD] 15.3) years (). No patients had early-onset bipolar disorder (first episode before age 18 yr). Older patients had significantly worse performances than controls in the simple reaction time test, Letter–Number Sequencing subtest, cued recall on the CR48 and delayed recall on the CERAD Word List Memory test. There were no significant differences in the other neuropsychologic tests between the 2 groups ().
Demographic and clinical characteristics of patients with bipolar disorder and healthy controls
Neuropsychologic test results of patients with bipolar disorder and healthy controls
Regarding medication, 79% of patients received mood stabilizers, 26% received antidepressants, 37% received benzodiazepines and 26% received neuroleptics. In 15% of patients, there was no prescription for mood stabilizers, antidepressants, benzodiazepines or neuroleptics ().
Medication load of patients with bipolar disorder*
In a routine radiologic assessment of vascular burden, periventricular hyperintensity scores were comparable between patients with bipolar disorder and controls (Z = 1.40, p = 0.10). This was also the case for hyperintensities in neocortical white matter (Z = 0.39, p = 0.61), basal ganglia (Z = 0.39, p = 0.54) and infratentorial areas (Z = 1.0, p = 0.30).
Grey matter VBM analysis of T1 data
When the entire grey matter was considered, there were no TFCE-corrected suprathreshold differences between the bipolar disorder and control groups. Restricting the analysis to the fronto-basal cortex and basal ganglia revealed significantly higher grey matter concentration in controls than in patients with bipolar disorder in the right anterior insula, head of caudate nucleus, nucleus accumbens, ventral putamen and frontal orbital cortex (, ). The inverse comparison of the bipolar disorder group versus controls yielded no significant voxels. These differences persisted after controlling for medication effect (data not shown).
Fig. 1 Voxel-based morphometry (VBM) and tract-based spatial statistics (TBSS) analysis of the bipolar disorder group versus healthy controls illustrates the spatial distribution of threshold-free cluster enhancement–corrected significant differences (more ...)
Significant clusters for the grey matter voxel-based morphometry analysis between patients with bipolar disorder and healthy controls*
White matter TBSS analysis of DTI data
Fractional anisotropy was significantly decreased in patients with bipolar disorder compared with controls in the ventral and central portion of the corpus callosum (, ). The analysis of longitudinal, radial and mean diffusivity revealed no TFCE-corrected suprathreshold voxels, although the differences were just below threshold (between p < 0.06 and p < 0.08, corrected) and had a similar distribution to the reported fractional anisotropy analysis. These differences persisted after controlling for medication effect (data not shown). The additional ROI analyses revealed significant differences between groups in the body of the corpus callosum (fractional anisotropy, p = 0.033; mean diffusivity, p = 0.018; radial diffusivity, p = 0.015). There were no significant group differences in DTI parameters (fractional anisotropy and longitudinal, radial and mean diffusivity) for the bilateral uncinate fasciculus, anterior cingulum and posterior cingulum.
Significant clusters for the white matter tract-based spatial statistics analysis between patients with bipolar disorder and healthy controls*
Correlation analyses between MRI data, duration of illness and neuropsychologic tests
The correlation analysis for the duration of illness yielded no TFCE-corrected suprathreshold voxels (VBM, fractional anisotropy and longitudinal, radial and mean diffusivity). There was, however, a clear trend between increasing duration of illness and decreasing fractional anisotropy in a distributed network at the level of the corpus callosum (persisting up to p < 0.2, TFCE-corrected). Conversely, there was no trend between increasing duration of disease and increasing fractional anisotropy (p < 0.95, TFCE-corrected).
Concerning the correlation analyses between MRI data (VBM, fractional anisotropy and longitudinal, radial and mean diffusivity) and neuropsychologic tests, there were no TFCE-corrected suprathreshold differences.