Mode of partition of callosal area and regional distribution of magentic resonance imaging lesions. The central point represents the center of gravity (mathematically calculated for each subject).
Pelletier J, Suchet L, Witjas T, Habib M, Guttmann CRG, Salamon G, Lyon-Caen O, Chérif AA. A Longitudinal Study of Callosal Atrophy and Interhemispheric Dysfunction in Relapsing-Remitting Multiple Sclerosis. Arch Neurol. 2001;58(1):105-111. doi:10.1001/archneur.58.1.105
To determine if callosal atrophy and interhemispheric dysfunction can be detected in the early stages of relapsing-remitting multiple sclerosis (MS) and to evaluate their progression in relation to the disability and evolution of lesions seen on magnetic resonance imaging during a 5-year period.
We compared 30 patients who had clinically definite early-onset replasing-remitting MS and mild disability with control subjects. Regional and segmental callosal size and extent of white matter abnormalities on magnetic resonance imaging, as well as performance on tasks exploring interhemispheric transfer of motor, auditory, and sensory information were assessed. Patients with MS were evaluated at baseline and after 5 years. Physical disability was determined at both times using the Expanded Disability Status Scale score.
Patients with MS were seen with significant callosal atrophy and functional impairment of interhemispheric transfer at baseline that worsened during the 5-year study. A significant correlation was found between the magnitude of disability and the severity of morphological and functional callosal involvement at baseline. This association persisted at year 5. Baseline clinical characteristics such as age and prestudy relapse rate were unrelated to callosal size or interhemispheric performance. However, the number of baseline T2-weighted lesions was correlated with callosal involvement and this relation persisted at year 5.
Patients who had relapsing-remitting MS in the early stages of the disease and mild disability had significant callosal involvement that progressed over time. The relationship between disability, T2-weighted lesions load, and degree of morphological and functional callosal impairment confirm the potential value of using callosal dysfunction as a surrogate marker of disease progression in MS.
CLINICAL AND magnetic resonance imaging (MRI) features that are predictive of disease progression in the early stages of relapsing-remitting (RR) multiple sclerosis (MS) remain uncertain. It was recently shown that T2-weighted brain lesion (T2 lesion) volume in patients with clinically isolated symptoms suggestive of MS was predictive of clinical disease progression.1,2 However, neuronal involvement and axonal loss have been demonstrated in early RR MS by magnetic resonance spectroscopy and histopathological studies.3- 6 Some studies have shown strong correlations between disability and spinal cord,7 cerebellar,8 and cerebral atrophy9- 11 in MS. In particular, Simon et al11 recently reported that patients with RR MS and moderate disability have measurable amounts of cerebral atrophy that progress yearly and that the course of cerebral atrophy was influenced by prior inflammatory activity of MS evaluated by the presence of gadolinium-enhancing brain lesions as seen on MRI. Cerebral atrophy observed in patients with MS could reflect an irreversible process based on axonal loss already present at an early stage of MS.3,12 However, if brain atrophy increases more rapidly in patients with high levels of disability, the progression of atrophy seems unrelated to clinical variables such as relapse rate or changes in the Expanded Disability Status Scale (EDSS) score.13 Finally, in early MS, the relationship between indexes of atrophy and level of disability, as well as their predictive value for clinical disease progression, remains uncertain. Nevertheless, whole-brain atrophy, as estimated by brain parenchymal fraction measurements on MRI, could represent a sensitive index for axonal loss and for evaluating the potential effects of treatments.13 In particular, using this approach, a recent report suggested a protective impact of interferon β-1a on the progression of cerebral atrophy.13 Another valuable method to evaluate axonal loss could be provided by studying callosal anatomy and interhemispheric function.
Autopsy and MRI studies indicate that atrophy of corpus callosum (CC) is a common finding in patients with MS.14- 24 Demyelinating lesions of callosal or subcallosal areas found in most patients with MS could explain CC atrophy.17,20 However, some studies demonstrated dysfunction of interhemispheric transfer in patients with MS, particularly on verbal dichotic listening tasks showing a left-ear dichotic suppression.18,22,25- 27 Functional impairment of interhemispheric transfer in MS has also been correlated to the degree of callosal atrophy and to the severity and diffusion of white matter changes identified by MRI.18,22,23 Moreover, it has been recently shown that even in cases where high-resolution MRI had failed to demonstrate cerebral demyelination, interhemispheric dysfunction and, to a lesser degree, callosal atrophy may be present.22 This unexpected result could be interpreted as witnessing infraradiological involvement of callosal fibers that could be therefore viewed as an early marker of axonal and neuronal damage.
To determine the potential value of callosal atrophy and interhemispheric transfer impairment as a sensitive marker of MS progression, we conducted a 5-year prospective study in a group of patients with RR MS enrolled at the early stage of MS who had mild disability based on EDSS scores. The aims of this study were to (1) assess whether callosal atrophy and interhemispheric impairment could be detected in patients with RR MS early in the disease, (2) determine if functional and morphological callosal impairments progress during a 5-year period, (3) evaluate the potential relationship between functional and morphological callosal measures and disability, and (4) determine the influence of T2 lesions on callosal involvement as seen on MRI.
Of the 90 patients with MS who were initially examined and enrolled in our original study,22 30 were still available and were considered suitable for our 5-year follow-up study. Inclusion criteria were (1) clinically definite RR MS as determined by the criteria of Poser et al28 with a clinically documented MS duration of less than 3 years at baseline, (2) no long-term treatment for MS during the follow-up period, (3) clinical remission at the time of evaluation, (4) no major signs of cerebellar, motor, or sensitive involvement of the upper limbs at baseline examination, and (5) hemispheric white matter hyperintensities on T2-weighted MRI consistent with the diagnosis of MS29 at baseline MRI. All patients' disabilities were scored using the EDSS30 at both times by the same evaluating neurologist (J.P.).
Twenty-five sex-, age-, and handedness-matched normal subjects without known neurological disorders or history of alcoholism or other drug abuse were recruited as controls and underwent neuropsychological testing. This group had a mean age of 29 years (age range, 20-40 years). Normal callosal morphology was defined by a previous study of 53 healthy volunteers who had undergone the same MRI procedure as used for this study.31
Interhemispheric transfer of auditory (verbal dichotic listening task), sensory (crossed tactile finger localization), and motor (finger-tapping task) information was evaluated in all patients and controls at baseline and 5 years later for the group with MS only. The detailed procedure has been previously described.22 Twenty-eight patients had normal brainstem auditory evoked potentials. For the remaining 2 patients, it was verified individually that abnormal brainstem auditory evoked potentials could not account for the asymmetry in dichotic listening performance. Data analyses were based on functional transfer (FT) indexes (mean difference in errors between the 2 ears for verbal dichotic listening task; mean difference in errors between intermanual and monomanual conditions for sensory transfer task, and mean ratio between monomanual and bimanual conditions for motor transfer task).22
The design of this study was performed in 1991 and we used the same MRI unit (Magneton; Siemens, Erlangen, Germany) operating at a field strength of 1.5 T at baseline and at year 5. The image resolution was 0.89 mm in-plane for 5-mm-thick sections (acquisition matrix, 256 × 256 pixels; field of view, 23 cm) and imaging was done in the axial and sagittal planes. The axial section thickness was 5 mm and computed tomographic scans were performed on patients using a T2 sequence with a repetition time of 2700 milliseconds and an echo-delay time of 20 to 90 milliseconds. All patients were also evaluated by T1-weighted sagittal partial saturation images using a repetition time of 600 milliseconds and an echo-delay time of 20 milliseconds (slice thickness, 5 mm).
Quantification of CC morphology was carried out from the T1-weighted midsagittal MRI. Global and segmental morphology of the CC was assessed with automated measurements based on a manual midsagittal outline of the CC. The midcallosal area was partitioned into the following 6 subregions: 3 anterior (A1, A2, and A3) and 3 posterior (P1, P2, and P3) regions.22,31 The extent of T2 white matter lesions was assessed on axial sections using a semiquantitative method with a 5-point grading scale (total lesion index).32 Regional distribution of lesions within and outside of the CC was analyzed for an anterior (A3 + A2), middle (A1 + P1), and posterior (P2 + P3) region using the same 5-point scale. Each axial image was subdivided into these 3 regions by projecting the boundaries calculated on the midsagittal image (Figure 1). Magnetic resonance imaging scans were graded independently by 2 blinded physicians (L.S. and T.W.). Interrater agreement was high (0.87 for lesion grading; 0.96 for CC measurements).
Analysis of variance (ANOVA) was performed to characterize the overall relationship between callosal functional performances and morphological measures on both data series (at baseline and at year 5). Multiple stepwise regression analysis was computed to assess the relationship between FT impairment, CC atrophy, and baseline and 5-year measures. Patients' and controls' FT scores, as well as global and segmental callosal morphology measures of patients with MS and controls were compared using the t test. Spearman rank correlation coefficients were calculated to characterize the relationship between clinical, functional, and MRI measures. The t tests were used to compare baseline and year 5 changes for callosal morphology and functional performances. All statistical analyses were performed on a Macintosh personal computer (Apple Computer Inc, Cupertino, Calf) using StatWorks (version 1.1, 1985; MacIntosh Inc, Apple Computer Inc) and CLR ANOVA (1992; Midvision Software, Abacus Concept Inc, Berkeley, Calif) softwares.
Demographic and clinical characteristics of the MS population are summarized in Table 1. As determined by the inclusion criteria, patients were at the early stage of MS (mean disease duration, 2.4 years) with a low EDSS score (mean EDSS score, 2.1).
As previously reported, patients with MS were significantly impaired compared with controls for all modalities explored (left ear extinction, hands-tactile condition, and increased time in bimanual tapping [Table 2]).
The comparison of mean callosal areas between the population with MS and the controls showed significant atrophy in the group with MS for global callosal area, as well as for all 6 subregions (P<.001) (Table 3). We found no significant effect or interaction of gender and hand preference on callosal area measurements in patients with MS.
The global ANOVA performed on the baseline results showed a significant interaction between performance on FT tasks for all modalities explored and EDSS score (F = 4.25, P<.01). A significant interaction was also found between CC morphology and EDSS score (F = 5.89, P<.001) and between T2 MRI lesion grading and EDSS score (F = 4.39, P<.01). The stepwise regression confirmed that the EDSS score was predominantly linked to dichotic listening (F = 10.625) and sensory tranfer tests (F = 9.34) for interhemispheric tranfer, to A3 and P3 areas for callosal morphology (F = 12.6), and to total index for MRI white matter lesions. A significant correlation was found between the magnitude of relative left ear impairment and all callosal measurements carried out (Table 4). Similarly, significant correlations were noted between severity of transfer impairment for tactile and motor tasks and callosal atrophy except for the splenium. As we previously reported, each FT was predominantly associated with atrophy of one part of the CC (anterior region for motor transfer, middle region for tactile localization task, and posterior region for dichotic listening test) (Table 4). Correlation between regional grading of T2 MRI white matter abnormalities and functional impairment of interhemispheric transfer tests demonstrated a global relation between the dichotic listening task score or the sensory transfer test score and each callosal measure (P<.01). In contrast, a preferential relationship between the motor transfer test and lesions localized in the middle and anterior regions was found (Table 5). Results of correlation analysis between callosal atrophy and MRI lesion load showed an association between the total lesion index and all measurements of callosal areas. More interestingly, regional analysis of MRI lesions demonstrated a preferential relationship between lesion indexes for the anterior, middle, and posterior regions of callosal and subcallosal areas and atrophy of the anterior, middle, and posterior callosal subregions (Table 6).
There were no significant correlations between baseline CC measures and FT and age (P = .09) or prestudy relapse rate (P = .12). However, baseline EDSS scores were correlated with baseline total callosal atrophy (P<.04), FT impairment (P<.01), and extent of T2 lesions (P = .02) (Table 7).
At year 5, the RR course of MS persisted in 23 patients with MS (80%) and no difference was found between baseline and year 5 for annual relapse rate. Patients who had relapses were treated with corticosteroids. Seven patients were on a secondary progressive course of the disease at year 5, but none of them was being treated with immunosuppressive drugs or serial infusions of corticosteroids during the study.
A significant progression of CC atrophy and FT impairment was noted between baseline and follow-up evaluation (Table 2 and Table 3). During the 5-year study, patients with MS had a significant increase of CC atrophy for each measure explored (P = .001 for total area, P<.02 for anterior subregions, and P<.01 for posterior subregions) and they presented a significant progression of impairment on the dichotic listening task (P = .005), sensory transfer test (P = .003), and motor transfer task (P = .01).
No significant correlations were found between baseline CC and FT and age (P = .09) or prestudy relapse rate (P = .12). Because of the semiquantitative method used to assess T2 MRI lesions, it was impossible to compare change in T2 white matter extent between baseline and year 5. Global ANOVA performed at follow-up showed a persistent significant interaction between performance on dichotic listening and sensory tasks and EDSS score (F = 3.82, P<.01), but no significant interaction was found for motor transfer test (F = 1.28, P = .12). A significant interaction was also found between CC morphology and EDSS scores (F = 4.21, P = .006) and between T2 MRI lesion grading and EDSS score (F = 4.02, P = .002). Multiple regression analysis showed no relationship between CC or FT and 5-year relapse rate or the number of corticosteroid courses used during the study. A significant relationship persisted at year 5 between CC atrophy and FT impairment except for the same subregions than baseline results (splenium for sensory and motor transfer tests and midanterior subregion for dichotic listening test) (Table 4). These results confirmed that CC atrophy and FT impairment follow a progressive and parallel evolution during the course of the disease. Correlation studies showed that the extent of T2 MRI abnormalities continued to be associated with FT impairment (Table 5) and to a lesser degree with CC atrophy (Table 6). Moreover, the extent of MRI lesions at year 5 was correlated with the final EDSS score and a significant correlation persisted between CC atrophy and FT impairment and EDSS level evaluated at year 5 (Table 7). Finally, progression of disability evaluated with EDSS was significantly related with progression of CC atrophy and FT impairment and with T2 lesion load.
At first, our results confirmed that patients with MS are seen with significant callosal atrophy compared with controls. As previously reported in other studies,15,17,20- 23 callosal atrophy is frequently observed in MS, either in advanced disease with a high level of disability or in patients with MS who have mild disability. Since the patients with MS included in this study were at an early stage of the RR MS phase and had a mild disability, callosal atrophy seems to be an early morphological marker of CC involvement. These results are similar to those reported recently by Simon et al.11 In their longitudinal study of brain atrophy, patients with MS included had a mild disability status (mean EDSS score±SD, 2.3 ± .82) and presented at baseline a significant decrease of CC area and brain width, as well as significant increases of third ventricle and lateral ventricle widths. A similar result was obtained using the brain parenchymal fraction that measures whole-brain atrophy.13 Moroever, our study provides further evidence of impaired auditory, motor, and sensory interhemispheric transfer and of a correlation between sensory interhemispheric transfer and regional callosal atrophy. This result argues in favor of the concept of topographical distribution of callosal fibers here demonstrated in vivo in a population with MS. Functional and morphological involvement of CC could be explained by demyelinating lesions of callosal or subcallosal areas and their possible consequences of axonal loss and wallerian degeneration.15
The relationship between callosal atrophy and degree of white matter lesions evaluated by MRI is unclear. Some previous studies have reported a strong significant interaction of T2 lesions with CC atrophy,16,22 with others finding only a weak relationship.17,23 These conflicting results could be explained by the heterogeneity of methods used to assess demyelinating lesions by MRI. Simon et al11 showed recently that the degree of CC atrophy was related to T2 lesion MRI volume as well as to third and lateral ventricle atrophy. On the contrary, they found no effect of gadolinium-enhancing lesions on callosal atrophy while the number of enhancing lesions at baseline was predictive of progression of third ventricle atrophy. In the light of this latter result, it is surprising to consider that the presence of enhancing lesions reflected inflammatory activity at the early stage of the RR MS but did not influence the degree of callosal atrophy. In contrast, T2 abnormalities that characterized probably more chronic demyelinating lesions with axonal loss were closely related to the degree of CC atrophy. Accordingly, our results clearly showed that atrophy of CC and FT impairment are both related to T2 MRI lesions and support the hypothesis that demyelinating lesions of callosal and pericallosal regions induce CC atrophy and interhemispheric impairment in patients with MS. The fact that CC atrophy could be detected at early stage of MS and in patients without MRI lesion could argue that CC atrophy represents an early marker of atrophy in MS.22 However, further studies of patients with MS seen at an early stage of MS with normal white matter appearance on standard MRI are needed to prove that involvement of CC could be related with early myelin and/or axonal loss. In this way, future studies using proton magnetic resonance spectroscopy and new techniques such as magnetization transfer MRI in patients with MS who have normal white matter appearance and are seen with isolated syndromes suggestive of MS could document this hypothesis.3,4,8
The relationship between atrophy measures and disability in MS has been evaluated in other studies, showing a significant link between cerebellar dysfunction and cerebellar atrophy,8 and between EDSS score and spinal cord atrophy.9,32 Consistent with a recent report by Simon et al11 of a significant link between CC, third and lateral ventricle atrophy, and EDSS score, we found a significant relationship between CC measures and EDSS score at baseline. In the same way, the degree of FT impairment for all modalities explored were correlated with the EDSS score. This significant interaction between clinical, functional, and morphological measures and their relation with T2 MRI lesions suggests that a destructive pathologic process is already present at the early stage of RR MS.
Longitudinal studies of brain atrophy in MS are rare.9,11,13 Results of our longitudinal evaluation showed that CC atrophy and FT impairment observed at baseline increased significantly at year 5 in the MS group. At follow-up, a significant interaction also persisted between functional scores on interhemispheric transfer and CC measures, and T2 lesion load was significantly related with degree of FT impairment and CC atrophy. Moreover and as already noted at baseline, the level of disability was linked to FT impairment and CC atrophy at year 5. Because of the small number of patients included in this study, it was impossible to compare FT and CC means in MS subgroups with high and low levels of clinical disease activity. The lack of interaction between callosal involvement and other clinical measures such as relapse rate, confirm the poor value of these clinical variables as prognostic markers in patients with RR MS.11,13,33 Moreover, because of the absence of relationship between treatment of relapses and CC atrophy or FT impairment, corticosteroids do not seem to be a contributing factor to callosal involvement. Finally, our results indicate that the major factor influencing CC atrophy and FT impairment either at baseline or at follow-up was T2 lesion load.
To further investigate the predictive value of callosal involvement in MS, another possibly fruitful avenue would be to explore the relationship between CC atrophy or FT impairment and neuropsychological dysfunctions. Although clinically apparent callosal disconnection has been rarely reported,34 some prior studies demonstrated that global cognitive dysfunction, as well as intellectual and memory disturbances were associated with a significant increase of ventricular width.35- 37 More recently, focal atrophy of the anterior part of the CC has been related to verbal fluency impairment.19,21 In particular, according to evidence showing that CC may play a key role in between-hemisphere facilitation that maintains bilateral cerebral arousal,38 future studies should focus on the relationship between attentional dysfunction, which is frequently reported in patients with MS,39,40 and CC atrophy and MR abnormalities41 to determine the natural history of callosal involvement in MS. Finally, the present results provide strong arguments for using anatomical and functional callosal measurements as key indexes for future evaluations of treatments susceptible of exerting a preventive effect on the natural history of MS.42- 45
Accepted for publication October 2, 2000.
This investigation was supported in part by grants from ARSEP (Association pour la Recherche sur la Sclérose en Plaques), and PHRC1994-CA3844-UF-1628, from the French National Ministry of Health, Paris, France.
Corresponding author: J. Pelletier, MD, PhD, Department of Neurology, CHU Timone, F-13385 Marseilles 5, France (e-mail:firstname.lastname@example.org).