Introduction
Cognitive impairment affects approximately 40% to 70% of people with multiple sclerosis (MS) at some point during the course of their disease. Reference Amato, Prestipino and Bellinvia1–Reference Strober, Rao, Lee, Fischer and Rudick4 It has also been documented before diagnosis in people with radiologically isolated syndrome and clinically isolated syndrome. Reference Lebrun, Blanc, Brassat, Zephir and de Seze5,Reference Anhoque, Domingues, Teixeira and Domingues6 The most common cognitive domains affected in MS include information processing speed (IPS) and episodic memory. Impairment in executive function, verbal fluency and visuospatial processing have also been well documented. Reference Strober, Rao, Lee, Fischer and Rudick4,Reference Sumowski, Benedict and Enzinger7–Reference Van Schependom, D’Hooghe and Cleynhens9 In our MS Cognitive clinic, in London, Ontario, Canada, people with MS most commonly complain about issues with word finding, recalling names or multi-tasking, which is often due to impaired processing speed, as well as recall or difficulty learning new tasks, which can be due to impaired processing speed or episodic memory. Cognitive impairment is commonly identified in our clinic within the first year of diagnosis. Reference DiGiuseppe, Blair and Morrow10 A number of tests have been designed to measure these aspects of cognition and provide a quantitative assessment of cognitive function. Reference Sumowski, Benedict and Enzinger7 However, various patient-specific confounders, including age, educational level, socioeconomic status, cognitive reserve and the presence of comorbid conditions, such as depression, sleep disturbances or substance abuse, can also impact cognition and must be considered as part of the assessment. Reference Kalb, Beier and Benedict11–Reference Kavaliunas, Danylaitė Karrenbauer, Binzer and Hillert15 Despite current recommendations for cognitive screening, routine assessments for changes in cognitive function do not always occur. Reference Kalb, Beier and Benedict11,Reference Pardini16,Reference Freedman, Devonshire and Duquette17 Consequently, important opportunities to detect disease activity beyond physical disability are missed.
The objectives of this review are to describe the relationships between cognitive dysfunction and other MS outcomes; to characterise how cognitive dysfunction affects the lives of people with MS; to review the evidence on the effects of disease-modifying therapies (DMTs) and other interventions on cognitive function in people with MS; and to provide clinical practice insights regarding screening and management of cognitive impairment in people with MS.
Relationship Between Cognitive Dysfunction and Other MS Outcomes
Cognitive changes often accompany radiological activity, and several magnetic resonance imaging (MRI) endpoints have been associated with cognitive dysfunction in people with MS. For example, measures of cortical and deep grey matter atrophy have been associated with impaired IPS. Reference Batista, Zivadinov and Hoogs18–Reference Amato, Portaccio and Stromillo26 Schoonheim et al. Reference Schoonheim, Popescu and Rueda Lopes24 monitored people with relapsing multiple sclerosis (RMS) from the time of diagnosis (or closely before) through to 6 years. They found subcortical atrophy was significantly correlated with cognitive impairment on the Brief Repeatable Battery for Neurological disease (BRB-N), particularly in men; these correlations were predicted by thalamic volume. Reference Schoonheim, Popescu and Rueda Lopes24 Another study including people with RMS or progressive forms of MS showed that grey matter volume and T1 lesion volume, in combination with age and baseline disability, accounted for 39.4% of the variance of change in Symbol Digit Modalities Test (SDMT) performance over 10 years. Reference Jacobsen, Zivadinov and Myhr27
In a cross-sectional study including people with RMS or secondary progressive MS, cortical lesion load detected by double inversion recovery was significantly correlated with SDMT. Reference Shaaban, Elmongui, Razek and Belal28 The presence of rimmed lesions may also signify cognitive involvement, with cross-sectional data from 192 people with MS demonstrating significantly worse scores on the SDMT and the Paced Auditory Serial Addition Test (PASAT) in people with at least four rimmed lesions versus those with none. Reference Absinta, Sati and Masuzzo29 Finally, damage to normal-appearing white matter has been identified as an important component of cognitive impairment, with findings from one study showing a significant relationship between heterogeneity in the normal-appearing white matter and SDMT performance in people with MS. Reference Abel, Vavasour and Lee30
No single MRI measure can definitively identify cognitive impairment or fully monitor changes in cognition over time. This was illustrated in a study that followed people with a baseline diagnosis of RMS or primary progressive MS (PPMS). In people with stable RMS, declines in cognitive function, using the BRB-N, correlated with increasing lesion volume over 5 years. In those who converted from RMS to progressive MS during that time, cognitive changes were predicted by deep grey matter atrophy. Reference Eijlers, Dekker and Steenwijk31 In people with PPMS, cognitive decline was only correlated with the rate of cortical atrophy. Reference Eijlers, Dekker and Steenwijk31
Stratifying people with MS into different cognitive phenotypes based on severity and functional domains of impairment can pinpoint brain regions that may be substrates for that type of cognitive dysfunction. In a cross-sectional study evaluating MRI features in patient groups defined by levels of cognitive impairment, people who were categorised as having a mild, multidomain phenotype demonstrated cortical atrophy as the most common MRI characteristic, whereas those with a severe executive/attention phenotype had greater T2 lesion volume. Reference De Meo, Portaccio and Giorgio32 People categorised with severe multidomain cognitive dysfunction showed severe atrophy in all brain regions examined; importantly, some people with this phenotype had a short disease duration and were not severely physically impaired. Reference De Meo, Portaccio and Giorgio32 However, despite the fact that MRI measures are clearly associated with, and may predict the development of, cognitive impairment in persons with MS, they remain elusive in many clinical practices, including in our MS Cognitive clinic. Thus, these possible biomarkers are mainly used in research studies, although in time they may become mainstream and available clinically.
Significant relationships also have been observed between physical disability and cognition. Transient worsening of cognition, particularly on the SDMT, has been observed in the context of acute relapse. Reference Benedict, Morrow and Rodgers33–Reference Morrow, Jurgensen, Forrestal, Munchauer and Benedict38 It also may be possible to predict disability outcomes based on the evaluation of cognitive impairment using simple tests of IPS. For example, in a study following 45 people with RMS over time, lower baseline SDMT and Selective Reminding Test scores were correlated with worse outcomes on the Expanded Disability Status Scale (EDSS) at 5 and 7 years, respectively. Reference Deloire, Ruet, Hamel, Bonnet and Brochet39
Impact of Cognitive Dysfunction on Functional Outcomes and Quality of Life
The impact of cognitive dysfunction is widespread and affects numerous aspects of daily life. Significant limitations have been observed in the workplace, including lower income Reference Kalb, Beier and Benedict11,Reference Kavaliunas, Danylaite Karrenbauer and Gyllensten40,Reference Rao, Leo, Ellington, Nauertz, Bernardin and Unverzagt41 and higher rates of unemployment. Reference Rao, Leo, Ellington, Nauertz, Bernardin and Unverzagt41–Reference Povolo, Blair, Mehta, Rosehart and Morrow44 In a study of 97 people with clinically definite MS, 28.9% had documented and paid disability benefits, and 45.4% reported a reduction in hours/work responsibilities. Moreover, declines in processing speed (as measured by the SDMT) and verbal memory (as measured by the California Verbal Learning Test, 2nd edition [CVLT2]) were found to be the most consistent predictors of clinically meaningful decline. Reference Morrow, Drake, Zivadinov, Munschauer, Weinstock-Guttman and Benedict43 Declines in executive function independently predict employment deterioration (stopping employment or reduced work hours) in people with RMS. Reference van Gorp, van der Hiele and Heerings45 Conversely, work-related improvements (such as increased work status, capacity or compensation) reported by people with MS have been associated with improved cognitive functioning. Reference Wojcik, Jaworski and Dwyer46
People with MS with cognitive impairment may have a reduced capacity to care for themselves and manage their daily lives. Reference Rao, Leo, Ellington, Nauertz, Bernardin and Unverzagt41 Deficits in verbal learning, memory and verbal fluency, for example, have been linked with reasoning impairments that may limit an individual’s ability to make informed decisions about their medical treatment. Reference Kalb, Beier and Benedict11 The ability to drive and manage money, and other important measures of independence, can also be compromised. Reference Kalb, Beier and Benedict11,Reference Morrow, Classen and Monahan47,Reference Schultheis, Garay, Millis and Deluca48
IPS deficits have been linked with symptoms of anxiety and depression in people with MS. Reference Vissicchio, Altaras and Parker49,Reference Niino, Mifune and Kohriyama50 Moreover, these individuals appear to have reduced social support. Compared with non-impaired people with MS, those with cognitive dysfunction tend to engage in fewer social activities, as measured by the Katz Adjustment Scale. Reference Rao, Leo, Ellington, Nauertz, Bernardin and Unverzagt41 Reduced IPS (as measured by the SDMT and PASAT) was strongly correlated with lower scores on the Medical Outcomes Study Social Support Survey in people with MS. Reference Eizaguirre, Vanotti and Merino51
The collective impact of cognitive dysfunction and its downstream effects on quality of life (QoL) in people with MS is substantial. Reference Eizaguirre, Vanotti and Merino51,Reference Barker-Collo52 Even in individuals with comparable levels of physical disability, those with cognitive impairments experience greater reductions in QoL than those who are non-impaired. Reference Kavaliunas, Danylaite Karrenbauer and Gyllensten40,Reference Rao, Leo, Ellington, Nauertz, Bernardin and Unverzagt41 Cognitive decline can predict these effects in a clinically meaningful way and can be used for advising people regarding what to expect over time. Reference Morrow, Drake, Zivadinov, Munschauer, Weinstock-Guttman and Benedict43
Interventions: Impact of DMTs on Cognitive Outcomes
The goal of preserving cognitive impairment has been the mainstay of treatment in persons with MS due to the lack of evidence demonstrating that medications can restore cognitive function. Medications used to treat dementia and Alzheimer’s disease, such as donepezil and memantine, have not been shown to be of any benefit. Reference Peyro Saint Paul, Creveuil and Heinzlef53,Reference Krupp, Christodoulou and Melville54 Further, treatment with amphetamines has shown some promise, but clear evidence of long-term benefit has not yet been published. Reference Harel, Appleboim, Lavie and Achiron55–Reference Morrow and Rosehart57
There have been some indications in the literature that DMTs may provide cognitive benefits for people with MS, presumably through relapse prevention and mitigation of lesion development. There is insufficient evidence that specific DMTs improve cognitive outcomes, and these changes to date have been modest. Current evidence does not support changing DMTs on the basis of cognitive assessment alone. Reference Kalb, Beier and Benedict11,Reference Freedman, Devonshire and Duquette17,Reference Landmeyer, Bürkner and Wiendl58,Reference DeLuca, Chiaravalloti and Sandroff59 A summary of phase 3 clinical studies of MS DMTs that included cognitive assessment is shown in Table 1. Reference Weinstock-Guttman, Galetta and Giovannoni60–Reference Langdon, Tomic and Penner64
Table 1: Phase 3 DMT trials with data evaluating cognition
PASAT, paced auditory serial addition test; SDMT, symbol digit modalities test; IFN, interferon; RMS, relapsing multiple sclerosis.
Natalizumab
Findings from two phase 3 trials (AFFIRM and SENTINEL) in people with MS showed that those who received natalizumab were significantly less likely to experience cognitive deficit (defined as worsening of 0.5 SD on the PASAT-3, confirmed for 12 weeks) than those treated with placebo; however, no difference was observed between those who received natalizumab + interferon (IFN) β-1a versus IFN β-1a alone. Reference Weinstock-Guttman, Galetta and Giovannoni60
Ozanimod
In the phase 3 SUNBEAM study in people with RMS, a numerically greater mean change in SDMT Z score was observed with ozanimod treatment compared with IFN β-1a, suggesting the potential benefits of ozanimod on IPS. Reference Comi, Kappos and Selmaj61 In exploratory post hoc analyses, ozanimod treatment was more likely to lead to clinically meaningful improvement (≥4 points) in SDMT than IFN β-1a. Reference DeLuca, Schippling and Montalban62
Fingolimod
Pooled data from the phase 3 FREEDOMS and FREEDOMS II trials showed that people with MS who received fingolimod had significantly greater improvements on their PASAT score over time compared with those receiving placebo. Reference Kappos, Radue, Chin, Ritter, Tomic and Lublin63 These results were consistently observed from 6 months of treatment through to 36 months when people receiving placebo were switched to fingolimod. Ten-year extension data showed continued benefits of fingolimod treatment. Reference Langdon, Tomic and Penner64
Interventions: Impact of Cognitive Rehabilitation/Behavioural Techniques
Solid evidence from well-designed studies of cognitive rehabilitation/behavioural techniques and physical exercises that may improve cognition in people with MS is limited, but has been increasing. Reference Kalb, Beier and Benedict11,Reference Goverover, Chiaravalloti, O’Brien and DeLuca65 In a randomised controlled trial (RCT) comparing a 12-week, remotely supervised, computer-based adaptive cognitive remediation program to an active computer game control, people with MS who participated in the cognitive remediation program improved on a neuropsychological composite measure compared with those assigned to the control group. Reference Charvet, Yang and Shaw66 The MEMREHAB trial was another RCT that demonstrated the use of the modified Story Memory Technique, a 10-session behavioural intervention, is effective for improving memory and learning in people with MS. Reference Chiaravalloti, Moore, Nikelshpur and DeLuca67 Promising results were observed in an RCT evaluating the effects of a progressive aerobic exercise program on IPS, Reference Langeskov-Christensen, Hvid and Jensen68 although more evidence is needed in this area before exercise programs can be considered part of recommended treatment. Reference Kalb, Beier and Benedict11,Reference Langeskov-Christensen, Hvid and Jensen68 Despite data demonstrating the benefits of cognitive rehabilitation/behavioural techniques, rehabilitation programs may be difficult to access due to a lack of trained clinicians, travel and/or time restrictions, or financial barriers. Reference Kalb, Beier and Benedict11,Reference Charvet, Yang and Shaw66
In our MS Cognitive clinic, our main approach is the use of behavioural techniques that can be used at work, school, or home to compensate for any impact noted by cognitive impairment in the lives of our patients with MS. Although there are limited options to improve cognitive function in persons with MS, using appropriate aids and techniques – the equivalent of using a cane to help with ambulation – can have a significant impact on quality of life.
Rationale for Routine Screening for Cognitive Dysfunction
The substantial negative impact of cognitive impairment on the lives of people with MS underscores the importance of regular screening and management of cognitive symptoms. Reference Strober, Rao, Lee, Fischer and Rudick4,Reference Kalb, Beier and Benedict11 Although there is insufficient evidence at this time to recommend specific DMTs to address cognitive impairment, awareness of these symptoms could indirectly impact treatment decisions, for example, by avoiding treatments with complex administration or reporting schedules. Reference Freedman, Devonshire and Duquette17 Moreover, recognition of changes in cognition may alert clinicians to ongoing disease activity that is not captured using more traditional assessments, such as the EDSS, which emphasise ambulation and other aspects of physical disability. In fact, recent evidence suggests that the EDSS alone underestimates the degree of disability before and during relapses and that incorporating the SDMT and the fatigue severity scale into the EDSS score improves the accuracy of disability assessment by accounting for cognitive changes during relapse. Reference Morrow, Conway and Fuchs69 The concept of isolated cognitive relapse, described as a transient worsening of performance on cognitive tests that are accompanied by the observance of gadolinium-enhancing brain lesions, but in the absence of any physical relapse symptoms, has been suggested by multiple studies. Reference Pardini16 A standard protocol for neuropsychological evaluation that can reliably detect such isolated cognitive relapses is necessary and would first require the presence of baseline cognitive assessment. Reference Pardini16
From a research perspective, accrual of longitudinal cognitive data in people with MS will provide a valuable resource that may be used to help advise people with MS and their caregivers on appropriate expectations for disease progression, as well as important context for the development of clinical trials aimed at identifying pharmacologic and other strategies to treat cognitive decline. Reference Strober, Rao, Lee, Fischer and Rudick4 The first step, however, is recognition, and thus it is recommended that neurologists promote awareness of this common symptom, how best to test for it, how to overcome barriers to testing that may exist in the clinic, and include cognition as an outcome measure in clinical trials. Reference Strober, Rao, Lee, Fischer and Rudick4,Reference Kalb, Beier and Benedict11,Reference Landmeyer, Bürkner and Wiendl58,Reference Strober, DeLuca and Benedict70
Recommended Screening Tools
Clinical practice recommendations that include guidance on the assessment of cognitive impairment in people with MS have been published by the National MS Society and the Canadian MS working group. Reference Kalb, Beier and Benedict11,Reference Freedman, Devonshire and Duquette17 In general, early baseline screening with the SDMT or a similarly validated test (Table 2) Reference Kalb, Beier and Benedict11,Reference Tombaugh71–Reference O’Brien, Gaudino-Goering, Shawaryn, Komaroff, Moore and DeLuca74 is recommended if the person is clinically stable. Reference Kalb, Beier and Benedict11,Reference Freedman, Devonshire and Duquette17 Reassessment with the same instrument is recommended on a regular basis (2–3 years) for screening for new-onset problems or for progression of cognitive impairment over time, detecting acute disease activity, and assessing treatment effects or relapse recovery. Reference Kalb, Beier and Benedict11,Reference Freedman, Devonshire and Duquette17 Awareness and treatment of comorbid factors that could confound test results, such as depression, anxiety, sleep disorders, and polypharmacy/cannabis use, are also important. Reference Kalb, Beier and Benedict11,Reference Feinstein, Meza, Stefan and Staines13 Indeed, testing for depression is recommended annually. Reference Kalb, Beier and Benedict11 In addition, individual patient factors such as education level, socio-economic status, and cognitive reserve can impact the rate of cognitive decline and test performance and should be taken into consideration. Reference Sumowski, Benedict and Enzinger7,Reference Sumowski, Rocca and Leavitt12,Reference Migeot, Calivar, Granchetti, Ibáñez and Fittipaldi14,Reference Kavaliunas, Danylaitė Karrenbauer, Binzer and Hillert15 Positive tests or evidence of significant cognitive decline should be followed up with more comprehensive testing. Reference Kalb, Beier and Benedict11,Reference Freedman, Devonshire and Duquette17
Table 2: Validated screening tools for cognitive assessment in MS
EDSS, expanded disability status scale; MS, multiple sclerosis.
A summary of validated tools for measuring cognitive dysfunction in MS is included in Table 2. The PASAT, a digit-summing activity that measures IPS as well as aspects of attention, has been widely used to assess cognitive impairment in people with MS. Reference Tombaugh71 Despite being a reliable and sensitive test, the PASAT is largely unpopular with people with MS and medical staff. It has been criticised for being affected by age and IQ and for creating psychological stress and agitation during administration. Reference Tombaugh71,Reference Sonder, Burggraaff, Knol, Polman and Uitdehaag75–Reference Aupperle, Beatty, Shelton Fde and Gontkovsky78 Furthermore, there is a significant practice effect with the PASAT as well as a ceiling effect with this test (a maximal score of 60), which limits longitudinal follow-up. Additionally, a low score does not confirm cognitive impairment. Reference Tombaugh71
The SDMT is a short symbol/digit-substitution task that measures IPS. Reference Smith79 It demonstrates the best sensitivity to changes in cognition that correlate well with MRI measures of disease Reference Morrow, Jurgensen, Forrestal, Munchauer and Benedict38,Reference Strober, DeLuca and Benedict70,Reference Benedict, DeLuca, Phillips, LaRocca, Hudson and Rudick72 and with measures of work or activities of daily living. Reference Benedict, DeLuca, Phillips, LaRocca, Hudson and Rudick72 Compared with the PASAT, the SDMT is easier to administer and has a better longitudinal sensitivity to cognitive impairment. Reference Strober, DeLuca and Benedict70,Reference Drake, Weinstock-Guttman, Morrow, Hojnacki, Munschauer and Benedict76,Reference Lopez-Gongora, Querol and Escartin80,Reference Benedict, Cookfair and Gavett81 Raw score changes ranging from 3 to 4 points and changes ≥10% have been reported as clinically meaningful. Reference Freedman, Devonshire and Duquette17
The Multiple Sclerosis Neuropsychological Questionnaire is a 15-item questionnaire that involves questions about activities of daily living and is administered as both a patient report and an informant report. Reference Benedict, Munschauer and Linn82 The informant report is more strongly correlated with cognition and the self-report can be impacted by depressive disorder, so the use of the informant report as a screening tool, in addition to the self-report when possible, is recommended. Reference Benedict, Cox, Thompson, Foley, Weinstock-Guttman and Munschauer73,Reference O’Brien, Gaudino-Goering, Shawaryn, Komaroff, Moore and DeLuca74 The Brief International Cognitive Assessment for Multiple Sclerosis (BICAMS), a 15-minute test used in adults and children, assesses cognitive functions that are most likely to be impaired in MS using the SDMT, CVLT2, and Brief Visuospatial Memory Test-Revised. Reference Kalb, Beier and Benedict11,Reference Langdon, Amato and Boringa83,Reference Charvet, Shaw, Frontario, Langdon and Krupp84 Although BICAMS is not intended to replace a full neuropsychological assessment, it is a rapid tool that can be used in everyday practice that screens in more than one cognitive domain. Reference Langdon, Amato and Boringa83,Reference Niccolai, Portaccio and Goretti85
After a positive screening test indicating cognitive impairment in a person with MS, a more comprehensive battery of neurophysiological tests can be conducted, such as the Minimal Assessment of Cognitive Function in MS (MACFIMS), which includes tests for information processing, memory, visuospatial function, verbal function, and executive function, or the shorter BICAMS, which includes SDMT and measures of verbal learning and visuospatial functioning, as noted previously, included in the MACFIMS. Reference Kalb, Beier and Benedict11
Conclusion
Cognitive impairment is a prevalent symptom of MS with a substantial influence on patient outcomes and should be measured at baseline and monitored routinely (e.g. every 2–3 years) throughout the MS disease course. Reference Kalb, Beier and Benedict11,Reference Freedman, Devonshire and Duquette17 Establishment of validated MRI and clinical markers capable of predicting cognitive change would provide a critical context for patient management and the development of clinical trials. More data are needed regarding the assessment of cognitive relapses that occur in the absence of motor-sensory symptoms in terms of progression, prognosis, and management. Reference Pardini16,Reference Meli, Roccatagliata and Capello86 Further research is warranted concerning the efficacy of interventions to improve or preserve cognition in people with MS, and cognitive endpoints, particularly SDMT, should be included as endpoints in prospective clinical trials. Reference Kalb, Beier and Benedict11,Reference Rae-Grant, Day and Marrie87
Acknowledgements
All authors contributed to and approved this article; writing and editorial assistance were provided by Liz Rockstein, PhD, of Peloton Advantage, LLC, an OPEN Health company and was funded by Bristol Myers Squibb.
Conflicts of Interest
The following authors have identified actual or potential conflicts of interest. Sarah A. Morrow has received consulting fees, honoraria, and grants from Biogen IDEC, Celgene/Bristol Myers Squibb, EMD Serono, Novartis, Roche, and Sanofi-Genzyme; travel support for meetings from Biogen IDEC; and assistance from Open Health Group with preparation of this manuscript. Chantal Baldwin has received travel grants from Ipsen and CSL Behring and speaking honorarium from Seer Medical. Samir Alkabie has no conflicts of interest to report.
Statement of Authorship
SAM, CB, and SA contributed to all aspects of manuscript development, including evidence and data analysis, writing and revisions, and final approval of this manuscript.
Open access
