Clinical Neurology and Neurosurgery 127 (2014) 54–58
Contents lists available at ScienceDirect
Clinical Neurology and Neurosurgery journal homepage: www.elsevier.com/locate/clineuro
Association of cognitive impairment and lesion volumes in multiple sclerosis – A MRI study Murat Yildiz a,∗ , Barbara Tettenborn a , Ernst-Wilhelm Radue b , Kerstin Bendfeldt b , Stefan Borgwardt b,c a
Department of Neurology, Kantonsspital St. Gallen, St. Gallen, Switzerland Medical Image Analysis Center, University of Basel, Basel, Switzerland c Department of Psychiatry, University of Basel, Basel, Switzerland b
a r t i c l e
i n f o
Article history: Received 3 January 2014 Received in revised form 18 August 2014 Accepted 24 September 2014 Keywords: Multiple sclerosis Cognitive impairment Lesion volume MRI
a b s t r a c t Cognitive impairment (CI) can be demonstrated in 40–65% of multiple sclerosis (MS) patients, sometimes starting from the early stages of the disease. The objective of this study was a community-based investigation of FLAIR-hyperintense lesion volumes (LV) and their association with CI in patients with relapsing remitting (RR) MS. The neurocognitive assessment was conducted with the brief cognitive screening instrument, MUSIC. Magnetic resonance imaging (MRI) scans were obtained with a 1.5 Tesla Sigma Magnetom MRI scanner. We conducted a stepwise multiple regression analysis to assess the relative contribution of the main clinical, demographic and MRI-variables in predicting cognitive impairment. We recruited 78 patients with RRMS. The mean MUSIC score was 20.6 ± 5.4. Forty five percent of patients (n = 35, mean score 15.1 ± 3.3) had CI and 55% (n = 43, mean score 24.4 ± 2.5) had no sign of CI. In the correlation analysis of the MUSIC subtests only the interference test correlated negatively with the LV (r = −0.23). Multivariate linear regression analysis using MUSIC as the dependent continuous variable revealed LV and disability severity as independent factors associated with MUSIC (p value of the regression model < 0.001; adjusted R-square = 0.11). The results of the present study suggest an association between white matter damage and CI in MS. We could demonstrate an association between attention difficulties and the LV in MS patients. Trial Registration: ClinicalTrials.gov Identifier: NCT01250665 and NCT01250678. © 2014 Published by Elsevier B.V.
1. Introduction Neurocognitive impairment (CI) has been identified in 40–65% of multiple sclerosis (MS) patients [1], and are already evident at the early stages of the disease [2]. Although in the past the relevance of CI was underestimated by clinicians, it is now accepted as a critical aspect of disease progression in MS [3]. Cognitive impairment is a leading cause of disability in MS and it can have profound socio-economic consequences for patients and their families [4–6]. The most common CIs found in MS patients are difficulties in information retention and processing speed, verbal abilities, sustained attention and executive functions [7].
∗ Corresponding author at: Department of Neurology, Cantonal Hospital of Saint Gallen, Rorschacher Strasse 95, Switzerland. Tel.: +41 71 494 11 11; fax: +41 71 494 2895. E-mail address: [email protected] (M. Yildiz). http://dx.doi.org/10.1016/j.clineuro.2014.09.019 0303-8467/© 2014 Published by Elsevier B.V.
Structural MRI studies have established WM lesions in multiple sclerosis (MS) and investigated their association with gray matter (GM) atrophy and clinical scores [8,9]. Recently, a number of magnetic resonance imaging (MRI) studies investigated neurobiological factors associated with CI in MS [10–16]. Previously, it was generally assumed that the association of CI and white matter (WM) lesions is rather modest in MS [16,17]. This may be explained in part due to the heterogeneity of the WM lesions with varying degrees of edema, inflammation, demyelination, remyelination, gliosis, and axonal loss [18]. Recent findings suggest a prominent role of WM lesions in the development of CI, especially the contribution of information-processing speed is higher than formerly assumed [19,12]. Cognitive impairment in MS due to WM damage may be an outcome of the interruption of crucial tracts between functional units which can be accompanied by insufficient compensatory reorganization of functional neuronal networks subsequently leading to CI [20]. The objective of this study was to investigate FLAIR-hyperintense lesion volumes (LV) in relation to CI in patients with relapsing remitting (RR) MS. By using
M. Yildiz et al. / Clinical Neurology and Neurosurgery 127 (2014) 54–58
an effectively applicable score for standard clinical practice (Multiple Sclerosis Inventory Cognition, MUSIC) we specifically aimed at translational application of the findings, stressing the significance of short assessment scales for their clinical usefulness in assessing cognitive impairment and relating it to MRI findings. In comparison to previous studies our cohort reflects the recent improvements in diagnosis and treatment of MS with shorter disease duration and milder disability. We hypothesized that the higher the LV the more severe the CI in RRMS. 2. Materials and methods 2.1. Subjects Patients were eligible with an age of 18–60 years and diagnosis of definite relapsing remitting multiple sclerosis (RRMS) according to the 2010 revisions to the McDonald criteria [21]. The study was approved by the local Ethics committee and written informed consent was attained from all participating patients. Disease duration was defined as time since first manifestation of MS. Neurological disability was assessed by Kurtzke expanded disability status scale (EDSS) [22]. Depressive symptoms were measured by the Beck depression inventory (BDI) [23]. Patients were tested for cognitive dysfunction and depressive symptoms in one visit between March and April 2011. We recruited 78 patients: 78 with diagnosis of RRMS at our outpatient clinic between 21 and 60 years of age at presentation. Forty-two patients were treated with natalizumab (ntz), n = 3 intramuscular IFN-1a 30 g once weekly, n = 20 s.c. IFN-1b 250 g every other day, n = 2 s.c. IFN-1a 22 g or n = 3 IFN-1a 44 g three times weekly and no treatment n = 8. 2.2. Neurocognitive assessment The neurocognitive evaluation was conducted with the brief cognitive screening instrument, MUSIC. It consists of 5 subtests to assess the cognitive core deficits in MS: memory and attention are mainly tested in “word list A” and “word list B” (also for assessing set-shifting capacity) for immediate recall and “word list A delayed” for delayed recall; the “verbal fluency” subtest is predominately measuring mental flexibility; information processing speed and inhibitory control are basically captured with a Stroop test called “interference test without interference” and “interference test with interference condition”, respectively. MUSIC also includes 3 items to examine fatigue [22]. The scale is ranging from 0 to 30 (higher scores indicating less cognitive impairment): 20–30, 16–19, 11–15 and ≤10 points indicate no cognitive dysfunction, mild cognitive dysfunction, moderate cognitive dysfunction and severe cognitive dysfunction, respectively. 2.3. Image acquisition MRI scans were obtained in the radiological institute of the Cantonal Hospital St. Gall (1.5 Tesla Sigma Magnetom Scanner, Siemens AG, Germany). They include axial pre- and post-gadolinium T1 weighted (T1w), sagittal T1w, sagittal and axial fluid attenuated inversion recovery (FLAIR) sequences and axial T2 weighted spinecho sequences. 2.4. Image analysis Marking and measurement of focal WM lesions was performed at the Medical Image Analysis Center, University Hospital in Basel, using commercial semi-automatic software (AMIRA 3.1.1; Mercury Computer Systems Inc.). Hyperintense lesions were identified
55
on the T2 and FLAIR images and manually outlined on the FLAIR images. Subsequently, volumes were calculated. 2.5. Statistical calculations Patient characteristics were summarized using mean ± standard deviation (or in the case of nonparametric data, median and range) or count and percentage, as appropriate. Categorical data were compared with the Chi-square test or Fisher’s exact test. By means of the Mann–Whitney U test, we compared the nonparametric data. The correlation between MUSIC and LV was estimated using the Pearson correlation coefficient, for which a 95% confidence interval was constructed using the Fisher Z transformation. We conducted a stepwise multiple regression analysis to assess the relative contribution of the main clinical, demographic and MRI-variables in predicting cognitive impairment. All tests have been performed two-sided. A statistical level of p < 0.05 was considered significant. Analyses were performed with Statistica Version 10, Tulsa, Oklahoma, USA. 3. Results The mean MUSIC score was 20.6 ± 5.4. The 5 subtests’ weighted scores of MUSIC were as follows: mean score 3.0 ± 1.1 (range 0–4; 4 maximum possible score) word list A; 3.8 ± 0.4 (2–4; 4) word list B; 1.8 ± 0.5 (1–3; 4) mental flexibility tested by verbal fluency; interference (Stroop) test: 3.4 ± 2.2 (0–6; 6) and 3.5 ± 2.0 (2–4; 5), picture naming without interference and with interference, respectively; 4.9 ± 2.2 (0–7; 7) word list A delayed for delayed recall. According to MUSIC 20 patients were categorized with mild cognitive dysfunction, 11 with moderate cognitive dysfunction and 4 with severe cognitive dysfunction, respectively. As the number of patients for each cognitive impairment subcategory was small we decided to form one group with cognitive impairments using the predefined
Contents lists available at ScienceDirect
Clinical Neurology and Neurosurgery journal homepage: www.elsevier.com/locate/clineuro
Association of cognitive impairment and lesion volumes in multiple sclerosis – A MRI study Murat Yildiz a,∗ , Barbara Tettenborn a , Ernst-Wilhelm Radue b , Kerstin Bendfeldt b , Stefan Borgwardt b,c a
Department of Neurology, Kantonsspital St. Gallen, St. Gallen, Switzerland Medical Image Analysis Center, University of Basel, Basel, Switzerland c Department of Psychiatry, University of Basel, Basel, Switzerland b
a r t i c l e
i n f o
Article history: Received 3 January 2014 Received in revised form 18 August 2014 Accepted 24 September 2014 Keywords: Multiple sclerosis Cognitive impairment Lesion volume MRI
a b s t r a c t Cognitive impairment (CI) can be demonstrated in 40–65% of multiple sclerosis (MS) patients, sometimes starting from the early stages of the disease. The objective of this study was a community-based investigation of FLAIR-hyperintense lesion volumes (LV) and their association with CI in patients with relapsing remitting (RR) MS. The neurocognitive assessment was conducted with the brief cognitive screening instrument, MUSIC. Magnetic resonance imaging (MRI) scans were obtained with a 1.5 Tesla Sigma Magnetom MRI scanner. We conducted a stepwise multiple regression analysis to assess the relative contribution of the main clinical, demographic and MRI-variables in predicting cognitive impairment. We recruited 78 patients with RRMS. The mean MUSIC score was 20.6 ± 5.4. Forty five percent of patients (n = 35, mean score 15.1 ± 3.3) had CI and 55% (n = 43, mean score 24.4 ± 2.5) had no sign of CI. In the correlation analysis of the MUSIC subtests only the interference test correlated negatively with the LV (r = −0.23). Multivariate linear regression analysis using MUSIC as the dependent continuous variable revealed LV and disability severity as independent factors associated with MUSIC (p value of the regression model < 0.001; adjusted R-square = 0.11). The results of the present study suggest an association between white matter damage and CI in MS. We could demonstrate an association between attention difficulties and the LV in MS patients. Trial Registration: ClinicalTrials.gov Identifier: NCT01250665 and NCT01250678. © 2014 Published by Elsevier B.V.
1. Introduction Neurocognitive impairment (CI) has been identified in 40–65% of multiple sclerosis (MS) patients [1], and are already evident at the early stages of the disease [2]. Although in the past the relevance of CI was underestimated by clinicians, it is now accepted as a critical aspect of disease progression in MS [3]. Cognitive impairment is a leading cause of disability in MS and it can have profound socio-economic consequences for patients and their families [4–6]. The most common CIs found in MS patients are difficulties in information retention and processing speed, verbal abilities, sustained attention and executive functions [7].
∗ Corresponding author at: Department of Neurology, Cantonal Hospital of Saint Gallen, Rorschacher Strasse 95, Switzerland. Tel.: +41 71 494 11 11; fax: +41 71 494 2895. E-mail address: [email protected] (M. Yildiz). http://dx.doi.org/10.1016/j.clineuro.2014.09.019 0303-8467/© 2014 Published by Elsevier B.V.
Structural MRI studies have established WM lesions in multiple sclerosis (MS) and investigated their association with gray matter (GM) atrophy and clinical scores [8,9]. Recently, a number of magnetic resonance imaging (MRI) studies investigated neurobiological factors associated with CI in MS [10–16]. Previously, it was generally assumed that the association of CI and white matter (WM) lesions is rather modest in MS [16,17]. This may be explained in part due to the heterogeneity of the WM lesions with varying degrees of edema, inflammation, demyelination, remyelination, gliosis, and axonal loss [18]. Recent findings suggest a prominent role of WM lesions in the development of CI, especially the contribution of information-processing speed is higher than formerly assumed [19,12]. Cognitive impairment in MS due to WM damage may be an outcome of the interruption of crucial tracts between functional units which can be accompanied by insufficient compensatory reorganization of functional neuronal networks subsequently leading to CI [20]. The objective of this study was to investigate FLAIR-hyperintense lesion volumes (LV) in relation to CI in patients with relapsing remitting (RR) MS. By using
M. Yildiz et al. / Clinical Neurology and Neurosurgery 127 (2014) 54–58
an effectively applicable score for standard clinical practice (Multiple Sclerosis Inventory Cognition, MUSIC) we specifically aimed at translational application of the findings, stressing the significance of short assessment scales for their clinical usefulness in assessing cognitive impairment and relating it to MRI findings. In comparison to previous studies our cohort reflects the recent improvements in diagnosis and treatment of MS with shorter disease duration and milder disability. We hypothesized that the higher the LV the more severe the CI in RRMS. 2. Materials and methods 2.1. Subjects Patients were eligible with an age of 18–60 years and diagnosis of definite relapsing remitting multiple sclerosis (RRMS) according to the 2010 revisions to the McDonald criteria [21]. The study was approved by the local Ethics committee and written informed consent was attained from all participating patients. Disease duration was defined as time since first manifestation of MS. Neurological disability was assessed by Kurtzke expanded disability status scale (EDSS) [22]. Depressive symptoms were measured by the Beck depression inventory (BDI) [23]. Patients were tested for cognitive dysfunction and depressive symptoms in one visit between March and April 2011. We recruited 78 patients: 78 with diagnosis of RRMS at our outpatient clinic between 21 and 60 years of age at presentation. Forty-two patients were treated with natalizumab (ntz), n = 3 intramuscular IFN-1a 30 g once weekly, n = 20 s.c. IFN-1b 250 g every other day, n = 2 s.c. IFN-1a 22 g or n = 3 IFN-1a 44 g three times weekly and no treatment n = 8. 2.2. Neurocognitive assessment The neurocognitive evaluation was conducted with the brief cognitive screening instrument, MUSIC. It consists of 5 subtests to assess the cognitive core deficits in MS: memory and attention are mainly tested in “word list A” and “word list B” (also for assessing set-shifting capacity) for immediate recall and “word list A delayed” for delayed recall; the “verbal fluency” subtest is predominately measuring mental flexibility; information processing speed and inhibitory control are basically captured with a Stroop test called “interference test without interference” and “interference test with interference condition”, respectively. MUSIC also includes 3 items to examine fatigue [22]. The scale is ranging from 0 to 30 (higher scores indicating less cognitive impairment): 20–30, 16–19, 11–15 and ≤10 points indicate no cognitive dysfunction, mild cognitive dysfunction, moderate cognitive dysfunction and severe cognitive dysfunction, respectively. 2.3. Image acquisition MRI scans were obtained in the radiological institute of the Cantonal Hospital St. Gall (1.5 Tesla Sigma Magnetom Scanner, Siemens AG, Germany). They include axial pre- and post-gadolinium T1 weighted (T1w), sagittal T1w, sagittal and axial fluid attenuated inversion recovery (FLAIR) sequences and axial T2 weighted spinecho sequences. 2.4. Image analysis Marking and measurement of focal WM lesions was performed at the Medical Image Analysis Center, University Hospital in Basel, using commercial semi-automatic software (AMIRA 3.1.1; Mercury Computer Systems Inc.). Hyperintense lesions were identified
55
on the T2 and FLAIR images and manually outlined on the FLAIR images. Subsequently, volumes were calculated. 2.5. Statistical calculations Patient characteristics were summarized using mean ± standard deviation (or in the case of nonparametric data, median and range) or count and percentage, as appropriate. Categorical data were compared with the Chi-square test or Fisher’s exact test. By means of the Mann–Whitney U test, we compared the nonparametric data. The correlation between MUSIC and LV was estimated using the Pearson correlation coefficient, for which a 95% confidence interval was constructed using the Fisher Z transformation. We conducted a stepwise multiple regression analysis to assess the relative contribution of the main clinical, demographic and MRI-variables in predicting cognitive impairment. All tests have been performed two-sided. A statistical level of p < 0.05 was considered significant. Analyses were performed with Statistica Version 10, Tulsa, Oklahoma, USA. 3. Results The mean MUSIC score was 20.6 ± 5.4. The 5 subtests’ weighted scores of MUSIC were as follows: mean score 3.0 ± 1.1 (range 0–4; 4 maximum possible score) word list A; 3.8 ± 0.4 (2–4; 4) word list B; 1.8 ± 0.5 (1–3; 4) mental flexibility tested by verbal fluency; interference (Stroop) test: 3.4 ± 2.2 (0–6; 6) and 3.5 ± 2.0 (2–4; 5), picture naming without interference and with interference, respectively; 4.9 ± 2.2 (0–7; 7) word list A delayed for delayed recall. According to MUSIC 20 patients were categorized with mild cognitive dysfunction, 11 with moderate cognitive dysfunction and 4 with severe cognitive dysfunction, respectively. As the number of patients for each cognitive impairment subcategory was small we decided to form one group with cognitive impairments using the predefined
