You are here
Longitudinal associations between MRI and cognitive changes in very early MS
Multiple Sclerosis and Related Disorders, January 2016, Pages 47 - 52
Cognitive dysfunction in multiple sclerosis (MS) has been primarily examined in patients with advanced disease. Our objective was to study the longitudinal associations between brain magnetic resonance imaging (MRI) metrics and neuropsychological outcomes in patients with early MS.
Relapsing MS patients within 12 months of onset were enrolled in a neuroprotection trial of riluzole versus placebo with up to 36 months of follow-up. MRI metrics included percent brain volume changes measured by SIENAX normalized measurements [normalized brain parenchymal volume (nBPV), normalized normal-appearing white and gray matter volume (nNAWMV and nGMV)] and T2 lesion volume (T2LV). A neuropsychological battery was performed annually. Mixed model regression measured time trends and associations between imaging and neuropsychological outcomes, adjusting for sex, age and education level.
Forty-three patients (mean age 36 years; 31 females) were enrolled within 7.5±4.9 months of disease onset. 11.6% of patients with baseline cognitive assessment met conservative criteria for cognitive impairment. Compared to placebo, riluzole had no significant effect on neuropsychological performance; thus, both groups were combined for the association analyses. Baseline T2LV predicted subsequent changes in PASAT (p=0.006) and SDMT (p=0.002) scores. Longitudinal changes of T2LV were associated with changes in CVLT-II (p<0.001).
These findings suggest that cognitive impairment is relatively common in patients with very early MS. Baseline and longitudinal changes in the lesion load may be associated with some of the most frequently identified changes in cognitive function in MS.
- We studied MRI and cognitive change up to 36 months in patients with very early MS.
- Powerful statistical model to examine the association between MRI and cognition.
- Only baseline lesion load predicted changes in cognitive scores.
- No longitudinal association between change in atrophy measures and cognition.
- Highlights the importance of lesion load in patient with early disease.
Keywords: Multiple sclerosis, MRI scans, Neuropsychological tests, Atrophy.
Cognitive impairment affects from 40% to 70% of patients with MS and has a profound impact on patients' psychosocial functioning and quality of life (Beatty et al., 1995). Processing speed, attention, explicit and episodic memory, and executive functioning deficits are common among MS patients with cognitive difficulties (Chiaravalloti and DeLuca, 2008). Cognitive dysfunction may affect patients even in early stages of the disease and can even be present in asymptomatic patients who eventually go on to develop MS (Potagas et al, 2008 and Sinay et al, 2015). Patient's age, severity of physical disability and disease duration cannot reliably predict cognitive disability, and thus identification of reliable predictors of cognitive impairment is needed (Beatty et al., 1995).
Measures of brain atrophy on MRI are sensitive and reproducible measures of tissue loss and neurodegeneration (Barkhof et al, 2009 and Simon, 2006). An association between these MRI outcomes and cognitive deficit in MS patients has been reported in cross-sectional studies of patients with long-standing disease (Hoffmann et al, 2007 and Hohol et al, 1997). However, much less is known regarding the predictive value of MRI measures of atrophy and disease activity for change in cognitive outcomes in patients with very early MS (Deloire et al., 2011). Identifying sensitive and reproducible imaging markers of cognitive deterioration could help with the design of future neuroprotective trials (Maghzi et al., 2013). In the current study, we investigated the longitudinal associations between various MRI parameters, including lesion load and brain atrophy, and several clinically relevant neuropsychological outcomes in patients with very early MS.
2.1. Study design
This is an exploratory analysis of a randomized, double-blind, placebo-controlled trial assessing the possible neuroprotective effects of riluzole in combination with intramuscular IFNB-1a that was conducted at two centers [University of California, San Francisco (UCSF) and Oregon Health & Science University (OHSU)] (Waubant et al., 2014). Patients with a history of a clinically isolated syndrome or relapsing-remitting (RR) MS with symptom onset within the previous 12 months, with at least two silent T2 hyperintensities in the deep white matter on a clinical brain or cervical MRI scan were offered to participate in the study (Polman et al, 2005 and McDonald et al, 2001). Other inclusion criteria included: (1) age between 18 and 55 years; (2) no prior exposure to disease-modifying therapies; and (3) no MS exacerbation or use of glucocorticoids within the four weeks preceding randomization and first MRI scan and cognitive testing. Patients were randomized to receive 50 mg riluzole or placebo twice daily. After three months on the study drug, subjects also initiated weekly intramuscular IFNB-1a. The core study was 24 months, and the first half of the subjects completing this two year period were offered continuation for an additional 12 months. The institutional review boards at both institutions approved the study and written informed consent was obtained from the participants.
2.2. Neuropsychological evaluation
The neuropsychological assessment included 6 tests whose scores addressed the following cognitive domains: (1) attention, working memory, and processing speed—Symbol Digit Modalities Test (SDMT), Paced Auditory Serial Addition Test (PASAT) (Rao, 1990); (2) visuospatial functioning—Judgment of Line Orientation (JLO) (Benton et al., 1994); (3) verbal episodic learning and recall—California Verbal Learning Test, Second Edition (CVLT-II) (Delis, 2000) total learning; (4) nonverbal episodic learning and recall—Brief Visuospatial Memory Test-Revised (BVMT-R) total learning (Benedict et al., 1996); and (5) executive functioning—Delis Kaplan Executive Function System (DKEFS) Trail Making Test (TMT) number-letter switching (Delis et al., 2001). The above neuropsychological battery is very similar to the neuropsychological tests included in the minimal assessment of cognitive function in multiple sclerosis (MACFIMS), with the addition of the DKEFS TMT and omission of the DKEFS Sorting Test and Controlled Oral Word Association Test (Benedict et al., 2002). We used raw scores of these tests (as opposed to their z scores) in modeling the association of MRI measures and cognitive function. Higher scores in all the cognitive tests (except DKEFS TMT number-letter switching) indicate better performance.
To limit minimize practice effect, at each visit, we applied alternate forms of each neuropsychological test (if available). The testing battery was administered by a clinical neuropsychologist at the baseline, months 12, 24, and in half of patients, at month 36. Published age-stratified normative data were used to estimate the impairment ratings (z score) for each of the 6 test scores. A test score was considered abnormal if it fell at least 1.5 standard deviation (SD) below the normative mean. Patients were considered cognitively impaired if at least one third of their test scores fell 1.5 SD below the normative mean. This is one of the most common methods of defining cognitive abnormality impairment in the literature (Fischer et al., 2014). We also calculated an overall cognitive impairment index by averaging the z scores (Fischer et al., 2014) of the 6 cognitive tests and assessed the association of this index with the MRI variables.
2.3. Brain imaging measurements
Details of the MRI protocol have been described previously (Srinivasan et al., 2006). Scans were performed on 3.0 T MRI scanners equipped with an 8-channel phased array coil (General Electric, Milwaukee, WI). A three-dimension volumetric inversion recovery spoiled gradient-echo T1-weighted isotropic sequence (3D IRSPGR, 1×1×1 mm3, 180 slices) was acquired for all brain volume measurements. MRI scans were obtained at the baseline and months 12, 24 and in half of patients, at month 36.
Changes in brain volume between two time-points were obtained from 3D IRSPGR, images using Structural Image Evaluation, using Normalization of Atrophy (SIENA) version 2.6 (Waubant et al., 2014). Output was converted into percentage brain volume change (PBVC) between pairs of scans. Brain volume metrics for each time-point which were obtained utilizing SIENAX included normalized normal-appearing white matter volume (nNAWMV), normalized gray matter volume (nGMV) and normalized brain parenchymal volume (nBPV). The T2 lesion volume (T2LV) was determined using a semi-automated contouring technique (Zhang et al, 2001 and Smith et al, 2002). T1 and T2 lesion masks were derived from manual segmentation of T1-visible white matter lesions on the 3D IRSPGR, using methods described previously (Mowry et al., 2009). T1 lesion masks were incorporated into the SIENAX program to prevent voxel misclassification.
2.4. Statistical analysis
Descriptive statistics for patient characteristics were presented either as percentages (%) or using mean +/− standard deviation (SD). Spearman correlations were used to assess the cross-sectional association of clinical and imaging outcomes at baseline. To assess the longitudinal nature of the data, a mixed model regression was used with the changes over baseline in the imaging variables as the predictor and the cognitive variables as the outcome. The mixed model allowed subject-specific intercepts and slopes to accommodate the repeated measures nature of the data and possible time trends during the course of the study. This model can isolate the between individual changes (using the baseline value of the predictor) from the within individual changes (using the difference between each value of the predictor and its baseline value). As age, sex and level of education are strong determinants of many neuropsychological scores; we adjusted our models for age, sex and years of education. For assessing the effect of riluzole (compared to placebo) on neuropsychological measures we used a similar mixed model regression, with the changes over baseline in the cognitive scores as the outcome and allocation group as the predictor.
All analyses were conducted in SAS Version 9.3 (SAS Institute, Cary, NC) and Stata Version 13.1 (Stata Corp, College Station TX). As the analyses of treatment effect showed no effect of riluzole (compared to placebo) on primary and secondary outcome measures (including the cognitive scores); we combined both treatment groups in the analyses of association between MRI and neuropsychological outcomes (Waubant et al., 2014). To decrease the chance of false positive findings due to multiple comparisons, we considered a nominal p value of ≤0.01 as statistically significant.
3.1. Patient characteristics
Baseline characteristics of patients who were recruited in the study are shown in Table 1. As all 43 patients enrolled in the trial had at least one cognitive assessment during the study, they all contributed to the longitudinal analysis of association between MRI and cognitive measures. Five patients (11.6%) did not complete the 24-month study and were lost to follow-up after months 3, 9, 15, 18 and 21 respectively, but contributed to the analysis. Overall, only 5 patients (2 in the riluzole group and 3 in the placebo group) switched to another disease-modifying medication during the study, as treating physicians determined there was clinical or radiological disease activity on intramuscular weekly IFN-beta 1a. Relapse rate was 0.22 per year in the placebo and 0.15 in the riluzole group (p=0.27).
|Demographics and clinical characteristics|
|Mean age in years±SD||36±9.32|
|Mean disease duration in months±SD||7.52±4.93|
|Median EDSS (range)||2.0 (0.0–5.5)|
|Mean Education in years± SD||15.4±3.0|
|Mean nBPV (cm3)±SD||1640±118|
|Mean nNAWMV (cm3)±SD||730±61.9|
|Mean nGMV (cm3)±SD||908±70.7|
|Mean T1 lesion volume (cm3)±SD||3.9± 5.1|
|Mean T2 lesion volume (cm3)±SD||4.3±5.6|
|Neuropsychological scores (Mean±SD)|
|Verbal episodic learning and recall|
|CVLT-II total learning||57.9±9.33|
|Nonverbal episodic learning and recall|
|BVMT-R total learning||3.4±2.13|
|Information processing speed|
|Symbol digit modalities test||57.9±9.42|
|DKEFS TMT-number-letter sequencing in seconds||67.3±23|
|Judgment of line orientation||12.3±1.99|
EDSS: Expanded disability status scale; nBPV: Normalized brain parenchymal volume; nNAWMV: Normalized normal appearing white matter volume; nGMV: Normalized gray matter volume; CVLT-II: California verbal learning test version II; DKEFS: Delis–Kaplan Executive Function System; TMT: trail making test; CWI: color word interference; DF: design fluency; BVMT-R: brief visuospatial memory test-revised; SDMT: symbol digit modalities test
3.2. Cognition status at baseline
Out of 43 patients with baseline cognitive assessment, 5 patients (11.6%) met criteria for cognitive impairment, defined as having at least one-third of completed test scores ≥1.5 standard deviations below published normative data (Julian et al., 2013).
There was no significant difference in gender, age, educational level, EDSS and baseline MRI measures between cognitively normal subjects and those impaired at baseline (Table 2).
|Cognitively normal||Cognitively impaired||P value|
|Mean age in years±SD||35.2±8.3||39.2±17.1||0.39b|
|Median EDSS (range)||2 (0–5.5)||2 (1–2.5)||0.81c|
|Mean education in years±SD||15.8±3.2||13.8±1.8||0.19c|
|Mean nBPV (cm3)±SD||1651.8± 114.7||1594.1± 143.2||0.32c|
|Mean nNAWMV (cm3)±SD||730.8± 60.7||724.2± 75.5||0.83c|
|Mean nGMV (cm3)±SD||915.2± 67.3||864.4± 85.0||0.14c|
|Mean T2 lesion volume (cm3)±SD||4.4± 5.8||3.7± 3.7||0.81c|
a Fisher exact test.
b Wilcoxon rank-sum.
c Student t-test; EDSS: Expanded disability status scale; nBPV: Normalized brain parenchymal volume; nNAWMV: Normalized normal appearing white matter volume; nGMV: Normalized gray matter volume.
3.3. Effect of riluzole on neuropsychological scores
Using a mixed model regression, riluzole had no significant effect on any neuropsychological outcomes compared to placebo (data not shown).
3.4. Cross-sectional association of baseline cognitive and MRI outcomes
Estimated correlations were small and not statistically significant (p>0.01) between baseline MRI variables and neuropsychological test scores (data not shown).
3.5. Baseline MRI outcomes as predictors of longitudinal changes of cognitive outcomes
Table 3 shows the association between baseline MRI metrics and longitudinal changes in cognitive outcomes. Baseline T2LV predicted changes in PASAT (p=0.006) and SDMT (p=0.002) scores.
|Dependent variable||Predictor variable||Estimate||95% confidence interval||p-Value|
|PASAT||nBPV (per 100 units)||0.01||−0.06 to 0.08||0.75|
|nGMV (per 100 units)||0.03||−0.09 to 0.16||0.63|
|nNAWMV (per 100 units)||0.01||−0.12 to 0.16||0.79|
|T2LV||−0.03||−0.05 to −0.01||0.004|
|SDMT||nBPV (per 100 units)||0.00||−.07 to 0.08||0.92|
|nGMV (per 100 units)||0.03||−0.10 to 0.16||0.64|
|nNAWMV (per 100 units)||0.00||−0.17 to 0.14||0.90|
|T2LV||−0.03||−0.05 to 0.01||0.021|
|CVLT-II total learning||nBPV (per 100 units)||0.09||0.03 to 0.16||0.002|
|nGMV (per 100 units)||0.15||0.04 to 0.21||0.005|
|nNAWMV (per 100 units)||0.17||0.04 to 0.30||0.012|
|T2LV||0.02||−0.01 to 0.04||0.14|
|BVMT-R total learning||nBPV (per 100 units)||0.00||−0.04 to 0.04||0.99|
|nGMV (per 100 units)||0.00||−0.07 to 0.07||0.94|
|nNAWMV (per 100 units)||0.00||−0.08 to 0.08||0.98|
|T2LV||0.01||−0.01 to 0.02||0.33|
|DKEFS TMT-number-letter switching||nBPV (per 100 units)||−0.09||−0.24 to 0.05||0.21|
|nGMV (per 100 units)||−0.19||−0.45 to 0.06||0.14|
|nNAWMV (per 100 units)||−0.15||−0.45 to 0.15||0.32|
|T2LV||0.00||−0.05 to 0.05||0.94|
|JOL||nBPV (per 100 units)||0.00||−0.01 to 0.03||0.41|
|nGMV (per 100 units)||0.02||−0.02 to 0.05||0.37|
|nNAWMV (per 100 units)||0.01||−0.04 to 0.05||0.55|
|T2LV||0.00||−0.01 to 0.01||0.42|
|Average z score||nBPV (per 100 units)||0.00||−0.01 to 0.02||0.41|
|nGMV (per 100 units)||0.01||−0.01 to 0.05||0.31|
|nNAWMV (per 100 units)||0.01||−0.02 to 0.05||0.62|
|T2LV||0.00||−0.01 to 0.00||0.14|
nBPV: normalized brain parenchymal volume; nGMV: normalized gray matter volume; nNAWMV: normalized normal appearing white matter volume; T1LV: T1 lesion volume, T2LV: T2 lesion volume; CVLT-II: California verbal learning test version II; DKEFS: Delis-Kaplan Executive Function System; TMT: trail making test; CWI: color word interference; DF: design fluency; BVMT-R: brief visuospatial memory test-revised.
3.6. Changes of MRI and cognitive outcomes over time
Using mixed-effects regression models, the changes in neuropsychological scores, T2LV, or nNAWMV over 2 years were small and statistically non-significant. On average, nBPV decreased by 1.9 cm3 per month (95% CI: 0.3–3.4; p=0.017) and nGMV decreased by 1.3 cm3 per month (95% CI: 0.4–2.2, p=0.004) during the study.
3.7. Longitudinal MRI changes as predictors of cognitive changes
Table 4 shows the association between longitudinal changes in MRI and cognitive outcomes. Longitudinal changes of T2LV were associated with changes in CVLT-II total learning; each cm3 increase in T2LV was associated with a 2.4 point decrease in CVLT-II total score (p<0.001).
|Dependent variable||Predictor variable||Estimate||95% confidence interval||p-Value|
|PASAT||nBPV (per 100 units)||0.0||−0.7 to 0.7||0.99|
|nGMV (per 100 units)||0.0||−1.5 to 1.1||0.76|
|nNAWMV (per 100 units)||0.3||−1.0 to 1.7||063|
|T2LV||−0.7||−1.7 to 0.4||0.19|
|SDMT||nBPV (per 100 units)||0.2||−0.4 to 0.8||0.52|
|nGMV (per 100 units)||0.3||−0.9 to 1.4||0.65|
|nNAWMV (per 100 units)||0.5||−0.6 to 1.7||0.37|
|T2LV||−1.0||−1.9 to −0.1||0.04|
|CVLT-II total learning||nBPV (per 100 units)||0.5||−0.2 to 1.2||0.15|
|nGMV (per 100 units)||0.8||−0.5 to 2.1||0.24|
|nNAWMV (per 100 units)||1.0||−0.4 to 2.3||0.17|
|T2LV||−1.9||−2.9 to −0.8||0.001|
|BVMT-R total learning||nBPV (per 100 units)||−0.1||−0.7 to 0.5||0.73|
|nGMV (per 100 units)||−0.2||−1.3 to 0.8||0.66|
|nNAWMV (per 100 units)||−0.1||−1.2 to 0.9||0.79|
|T2LV||−0.9||−1.7 to −0.1||0.023|
|DKEFS TMT-number-letter switching||nBPV (per 100 units)||−0.1||−2.2 to 2.1||0.94|
|nGMV (per 100 units)||−1.5||−5.6 to 2.5||0.45|
|nNAWMV (per 100 units)||−1.5||−2.7 to 5.6||0.49|
|T2LV||0.4||−2.6 to 3.4||0.81|
|JOL||nBPV (per 100 units)||0.1||−0.1 to 0.4||0.11|
|nGMV (per 100 units)||0.3||−0.1 to 0.6||0.17|
|nNAWMV (per 100 units)||0.3||−0.1 to 0.7||0.10|
|T2LV||0.1||−0.2 to 0.4||0.48|
|Average z score||nBPV (per 100 units)||0.00||−0.01 to 0.01||0.19|
|nGMV (per 100 units)||0.00||−0.01 to 0.01||0.45|
|nNAWMV (per 100 units)||0.00||−0.01 to 0.01||0.10|
|T2LV||−0.07||−0.13 to −0.01||0.015|
nBPV: normalized brain parenchymal volume; nGMV: normalized gray matter volume; nNAWMV: normalized normal appearing white matter volume; T1LV: T1 lesion volume, T2LV: T2 lesion volume; NLV: normalized lesion volume; CVLT-II: California verbal learning test version II; DKEFS: Delis-Kaplan Executive Function System; TMT: trail making test; CWI: color word interference; DF: design fluency; BVMT-R: brief visuospatial memory test-revised.
We demonstrate that even in very early MS, there are strong associations between baseline and longitudinal MRI, and cognitive changes.
Specifically, increasing lesion burden was a strong predictor of within subject changes in verbal learning and memory. Furthermore, baseline T2 lesion volume predicted between subject differences for information processing speed (IPS) up to 3 years. These findings, overall, suggest that memory and IPS might be more vulnerable to changes in global measures of disease activity on MRI early in the course of MS.
It has been reported that in univariate analyses, baseline T2 lesion load could predict change in memory and IPS scores over seven year (Deloire et al., 2011). T2 hyperintense lesions are the radiological hallmarks of disease activity in MS and although it is thought that the correlation between the overall T2 lesion burden and physical and cognitive disability is not strong, this clinic-radiological paradox disappears when appropriate analysis techniques are used (Hackmack et al., 2012). Changes in physical and cognitive performance in patients with early MS are usually small and detecting them with the techniques routinely used in clinic by neurologists is not possible. Even relatively sensitive tests (such as PASAT and SDMT) did not show statistically significant changes over time in our study of very early MS patients who received interferon beta 1-a intramuscularly once a week. Despite that, statistically significant associations between the small baseline T2 lesion load and IPS score changes (on both PASAT and SDMT tests) may reflect the detrimental effect of T2 lesion burden or location in patients with very early MS. This argues for the importance of instituting disease modifying therapies (with the goal of preventing new T2 lesions) in patients with early clinical or radiological signs of disease activity.
After adjusting analyses for possible confounders, we did not find any of the longitudinal association between MRI measures of atrophy and changes in neuropsychological scores in patients with CIS or very early MS that have been reported in the literature. For example, using a similar analysis model, a trend for longitudinal association of percent change of gray matter volume and cortical volume changes and BVMT-R delayed recall (a measure of non-verbal episodic learning and recall) was reported in CIS (Uher et al., 2014). In another study of CIS patients, number of T1 lesions, T1 lesion volume and gray matter tissue fraction at baseline predicted overall cognitive score, performance in a test of verbal response generation and inhibition and poor executive function after 7 years, respectively (Summers et al., 2008). In a study of patients with disease duration of 5 years, change in brain parenchymal fraction was an independent predictor of cognitive impairment (Zivadinov et al., 2001). It is possible that the decrease in the rate of brain atrophy by interferon beta 1-a in our study decreased our power to detect any longitudinal association between brain atrophy and cognitive changes.
We did not find a cross-sectional correlation between baseline volumetric measures on MRI and neuropsychological scores. Although this may in part be due to our small sample size, these findings are in line with reported lack of such correlations in patient with CIS (Uher et al., 2014). In another study, cognitive impairment (using less stringent definition than in our study) was reported in more than 50% of the patients within 3 months of MS onset; however, there was no cross-sectional association between cognitive impairment and T2, T1, and gadolinium-enhancing lesions on brain MRI (Achiron and Barak, 2003). This contrasts with previously reported cross-sectional associations of volumetric MRI measures and cognitive outcomes in patients with long-standing MS (Sanfilipo et al., 2006). One potential explanation for the lack of cross-sectional association between MRI and cognitive outcomes is the suboptimal sensitivity and specificity of atrophy and lesion volume measures and imperfect reproducibility of neuropsychological tests. These would tend to bias the results towards the null hypothesis (no correlation between predictors and outcomes).
Additionally, we found that a relatively substantial proportion of individuals with early MS met conservative criteria for cognitive impairment. Our findings are consistent with the few existing studies showing that a significant proportion of early MS patients show deficits on measures of processing and psychomotor speed, verbal and non-verbal learning and memory, and executive functioning (Uher et al, 2014 and Schulz et al, 2006). However, existing studies vary greatly in estimates of cognitive impairment, in part due to differences in how cognitive impairment was defined, how patients were selected and at what time testing was performed. While some studies of patients with CIS have reported estimates ranging from 12.3% to 13.6% (Uher et al, 2014 and Khalil et al, 2011), other studies have reported much higher estimates ranging up to 80% of the patient sample (Achiron and Barak, 2003, Feuillet et al, 2007, and Kocer et al, 2008). This disparity in the reported prevalence of abnormal cognitive function in patients with CIS or early MS may be accounted for by different definitions of cognitive impairment, use of different neuropsychological batteries across studies, and evolution of MS diagnostic criteria over time with overall earlier diagnosis of patients with inflammatory demyelinating diseases of the CNS in more recent studies.
We did not find any difference in the studied measures of brain atrophy or lesion volumes between patients with normal and abnormal cognitive performance at baseline. As only few patients met our definition of cognitive impairment, we may not have been powered to detect a moderate difference.
Riluzole, as compared to placebo, did not have any significant effect on studied neuropsychological outcomes. Considering the reported lack of beneficial effects of riluzole on the primary and secondary outcomes of this randomized controlled trial (Waubant et al., 2014), this finding was not unexpected.
The strengths of our study include frequent and systematic longitudinal measurement of a variety of MRI outcomes and extensive cognitive testing in a homogenous cohort of patients with very early MS in the setting of a clinical trial. Most reported studies of association between MRI measures and cognitive outcomes are cross-sectional and are performed in patients with long-standing MS. Those studies that investigated these associations in early MS in a longitudinal fashion used difference in MRI outcomes and cognitive performance between two time points in their statistical models (Deloire et al., 2011). We used mixed effects regression models that take advantage of each of the 3 to 4 concomitant MRI and neuropsychological evaluations available for each patient. This method of analysis better reflects changes over time. We also separated the effect of between subject and within subject change over time in our models. The present study has also several limitations including its limited sample size and follow-up duration and as such, we might have limited power to examine the temporal associations between radiological and neuropsychological outcomes. Due to the very early phase of MS under study, only a small proportion of patients had cognitive impairment. We also were unable to control for the practice effect in neuropsychological tests that did not have alternate forms (e.g. DKEFS TMT).
In conclusion, we demonstrate the presence of cognitive impairment in a subset of patients with very early MS and show strong associations between baseline lesion load and change in IPS and longitudinal changes in lesion load and verbal memory over up to 36 months. We did not find studied brain atrophy measures to be sensitive markers of change in cognitive scores in patients with very early MS taking a disease-modifying therapy. These findings also highlight the importance of measures of disease activity (lesion load) and instituting anti-inflammatory strategies early in the course of MS. Future studies should include longer follow-up and possibly examine additional atrophy modalities such as thalamic volume and cortical thickness in order to provide more thorough longitudinal data on these associations.
Bardia Nourbakhsh is a grantee of National MS Society. This research was conducted while B.N. was an American Brain Foundation and a Biogen Idec Postdoctoral Fellow. Amir-Hadi Maghzi was funded by the Multiple Sclerosis International Federation (www.msif.org) through a McDonald Fellowship. This research was performed as a research grant funded by the National MS Society (PI Waubant, RG3932-A-2).
Dr. Nourbakhsh is a a Biogen Idec Postdoctoral Fellow. He is a grantee of National MS Society.
Ms. Nunan-Saah and Dr. Maghzi report no disclosure.
Dr. Julian is employed by Genentech.
Dr. Spain report, Dr. Jin and Dr. Lazar report no disclosures.
Dr. Pelletier has received consulting fees from CNS Imaging Consultant, LLC, and research grants to his academic institution from Hoffmann-LaRoche, Biogen Idec, and Genzyme.
Dr. Waubant has received honorarium from Teva, Sanofi Aventis and Genentech for three educational lectures, and is on the advisory board for a trial of Novartis. Dr. Waubant has received free medication from Biogen Idec and Sanofi-Aventis for the trial from which these data were generated.
- Achiron and Barak, 2003 A. Achiron, Y. Barak. Cognitive impairment in probable multiple sclerosis. J. Neurol. Neurosurg. Psychiatry. 2003;74:443-446 Crossref
- Barkhof et al., 2009 F. Barkhof, P.A. Calabresi, D.H. Miller, et al. Imaging outcomes for neuroprotection and repair in multiple sclerosis trials. Nat. Rev. Neurol.. 2009;5:256-266 10.1038/nrneurol.2009.41 Crossref
- Beatty et al., 1995 W.W. Beatty, R.H. Paul, S.L. Wilbanks, et al. Identifying multiple sclerosis patients with mild or global cognitive impairment using the Screening Examination for Cognitive Impairment (SEFCI). Neurology. 1995;45:718-723 10.1212/WNL.45.4.718 Crossref
- Benedict et al., 1996 R.H.B. Benedict, L. Groninger, D. Schretlen, M. Dobraski, B. Shpritz. Revision of the brief visuospatial memory test: studies of normal performance, reliability and validity. Psychological Assessment. 1996;8:145-153 Crossref
- Benedict et al., 2002 R.H.B. Benedict, J.S. Fischer, C.J. Archibald, et al. Minimal neuropsychological assessment of MS patients: a consensus approach. Clin. Neuropsychol.. 2002;16:381-397 10.1076/clin.16.3.381.13859 Crossref
- Benton et al., 1994 A.L. Benton, A.B. Sivan, K. deS. Hamsher, et al. Contributions to Neuropsychological Assessment: A Clinical Manual. 2 edition (Oxford University Press, New York, 1994)
- Chiaravalloti and DeLuca, 2008 N.D. Chiaravalloti, J. DeLuca. Cognitive impairment in multiple sclerosis. Lancet Neurol.. 2008;7:1139-1151 10.1016/S1474-4422(08)70259-X Crossref
- Delis et al., 2001 D.C. Delis, E. Kaplan, J.H. Kramer. Delis–Kaplan executive function system. Psychol. Corp.. 2001;
- Delis, D.C., 2000, California Verbal Learning Test, Second Edition: CvLT-II ; Adult Version ; Manual. Pearson. Delis, D.C., 2000, California Verbal Learning Test, Second Edition: CvLT-II ; Adult Version ; Manual. Pearson.
- Deloire et al., 2011 M.S.A. Deloire, A. Ruet, D. Hamel, et al. MRI predictors of cognitive outcome in early multiple sclerosis. Neurology. 2011;76:1161-1167 10.1212/WNL.0b013e318212a8be Crossref
- Feuillet et al., 2007 L. Feuillet, F. Reuter, B. Audoin, et al. Early cognitive impairment in patients with clinically isolated syndrome suggestive of multiple sclerosis. Mult. Scler. Houndmills Basingstoke Engl.. 2007;13:124-127 Crossref
- Fischer et al., 2014 M. Fischer, A. Kunkel, P. Bublak, et al. How reliable is the classification of cognitive impairment across different criteria in early and late stages of multiple sclerosis?. J. Neurol. Sci.. 2014;343:91-99 10.1016/j.jns.2014.05.042 Crossref
- Hackmack et al., 2012 K. Hackmack, M. Weygandt, J. Wuerfel, et al. Can we overcome the “clinico-radiological paradox” in multiple sclerosis?. J. Neurol.. 2012;259:2151-2160 10.1007/s00415-012-6475-9 Crossref
- Hoffmann et al., 2007 S. Hoffmann, M. Tittgemeyer, D.Y. von Cramon. Cognitive impairment in multiple sclerosis. Curr. Opin. Neurol.. 2007;20:275-280 10.1097/WCO.0b013e32810c8e87
- Hohol et al., 1997 M.J. Hohol, C.R. Guttmann, J. Orav, et al. Serial neuropsychological assessment and magnetic resonance imaging analysis in multiple sclerosis. Arch. Neurol.. 1997;54:1018-1025
- Julian et al., 2013 L. Julian, D. Serafin, L. Charvet, et al. Cognitive impairment occurs in children and adolescents with multiple sclerosis: results from a United States network. J. Child Neurol.. 2013;28:102-107 10.1177/0883073812464816 Crossref
- Khalil et al., 2011 M. Khalil, C. Enzinger, C. Langkammer, et al. Cognitive impairment in relation to MRI metrics in patients with clinically isolated syndrome. Mult. Scler. Houndmills Basingstoke Engl.. 2011;17:173-180 10.1177/1352458510384009 Crossref
- Kocer et al., 2008 B. Kocer, T. Unal, B. Nazliel, et al. Evaluating sub-clinical cognitive dysfunction and event-related potentials (P300) in clinically isolated syndrome. Neurol. Sci. J. Italy Neurol. Soc. Italy Soc. Clin. Neurophysiol.. 2008;29:435-444 10.1007/s10072-008-1020-4 Crossref
- Maghzi et al., 2013 A.-H. Maghzi, A. Minagar, E. Waubant. Neuroprotection in multiple sclerosis: a therapeutic approach. CNS Drugs. 2013;27:799-815 10.1007/s40263-013-0093-7 Crossref
- McDonald et al., 2001 W.I. McDonald, A. Compston, G. Edan, et al. Recommended diagnostic criteria for multiple sclerosis: guidelines from the International Panel on the diagnosis of multiple sclerosis. Ann. Neurol.. 2001;50:121-127 Crossref
- Mowry et al., 2009 E.M. Mowry, A. Beheshtian, E. Waubant, et al. Quality of life in multiple sclerosis is associated with lesion burden and brain volume measures. Neurology. 2009;72:1760-1765 10.1212/WNL.0b013e3181a609f8 Crossref
- Polman et al., 2005 C.H. Polman, S.C. Reingold, G. Edan, et al. Diagnostic criteria for multiple sclerosis: 2005 revisions to the “McDonald Criteria.”. Ann. Neurol.. 2005;58:840-846 10.1002/ana.20703 Crossref
- Potagas et al., 2008 C. Potagas, E. Giogkaraki, G. Koutsis, et al. Cognitive impairment in different MS subtypes and clinically isolated syndromes. J. Neurol. Sci.. 2008;267:100-106 10.1016/j.jns.2007.10.002 Crossref
- RaoS. M., Manual for the Brief, Repeatable Battery of Neuropsychological Tests in Multiple Sclerosis. National Multiple Sclerosis Society. Rao S.M., Manual for the Brief, Repeatable Battery of Neuropsychological Tests in Multiple Sclerosis. National Multiple Sclerosis Society, 1990.
- Sanfilipo et al., 2006 M.P. Sanfilipo, R.H.B. Benedict, B. Weinstock-Guttman, et al. Gray and white matter brain atrophy and neuropsychological impairment in multiple sclerosis. Neurology. 2006;66:685-692 10.1212/01.wnl.0000201238.93586.d9 Crossref
- Schulz et al., 2006 D. Schulz, B. Kopp, A. Kunkel, et al. Cognition in the early stage of multiple sclerosis. J. Neurol.. 2006;253:1002-1010 10.1007/s00415-006-0145-8 Crossref
- Simon, 2006 J.H. Simon. Brain atrophy in multiple sclerosis: what we know and would like to know. Mult. Scler. Houndmills Basingstoke Engl.. 2006;12:679-687 Crossref
- Sinay et al., 2015 V. Sinay, M. Perez Akly, G. Zanga, et al. School performance as a marker of cognitive decline prior to diagnosis of multiple sclerosis. Mult. Scler.. 2015;21:945-952 10.1177/1352458514554054 Crossref
- Smith et al., 2002 S.M. Smith, Y. Zhang, M. Jenkinson, et al. Accurate, robust, and automated longitudinal and cross-sectional brain change analysis. NeuroImage. 2002;17:479-489 Crossref
- Srinivasan et al., 2006 R. Srinivasan, C. Cunningham, A. Chen, et al. TE-averaged two-dimensional proton spectroscopic imaging of glutamate at 3 T. NeuroImage. 2006;30:1171-1178 10.1016/j.neuroimage.2005.10.048 Crossref
- Summers et al., 2008 M. Summers, J. Swanton, K. Fernando, et al. Cognitive impairment in multiple sclerosis can be predicted by imaging early in the disease. J. Neurol. Neurosurg. Psychiatry. 2008;79:955-958 10.1136/jnnp.2007.138685 Crossref
- Uher et al., 2014 T. Uher, J. Blahova-Dusankova, D. Horakova, et al. Longitudinal MRI and neuropsychological assessment of patients with clinically isolated syndrome. J. Neurol.. 2014;261:1735-1744 10.1007/s00415-014-7413-9 Crossref
- Waubant et al., 2014 E. Waubant, A.-H. Maghzi, N. Revirajan, et al. A randomized controlled phase II trial of riluzole in early multiple sclerosis. Ann. Clin. Transl. Neurol.. 2014;1:340-347 10.1002/acn3.60 Crossref
- Zhang et al., 2001 Y. Zhang, M. Brady, S. Smith. Segmentation of brain MR images through a hidden Markov random field model and the expectation-maximization algorithm. IEEE Trans. Med. Imaging. 2001;20:45-57 10.1109/42.906424 Crossref
- Zivadinov et al., 2001 R. Zivadinov, J. Sepcic, D. Nasuelli, et al. A longitudinal study of brain atrophy and cognitive disturbances in the early phase of relapsing-remitting multiple sclerosis. J. Neurol. Neurosurg. Psychiatry. 2001;70:773-780 Crossref
a Department of Neurology, University of California San Francisco, San Francisco, USA
b Department of Pediatrics, University of California San Francisco, San Francisco, USA
c Department of Internal Medicine, University of California San Francisco, San Francisco, CA, USA
d Department of Epidemiology and Biostatistics, University of California San Francisco, San Francisco, CA, USA
e Palo Alto University, Palo Alto, CA, USA
f Department of Neurology, Oregon Health and Science University, OR, USA
g Department of Neurology, University of Southern California, Los Angeles, CA, USA
⁎ Correspondence to: Department of Neurology, University of California San Francisco, 675 Nelson Rising Lane, Room 221F, Box 3206, 94158 San Francisco, CA, USA.
1 Conducted statistical analysis.
© 2015 Elsevier B.V., All rights reserved.