[Skip to Content]
Access to paid content on this site is currently suspended due to excessive activity being detected from your IP address 18.207.134.98. Please contact the publisher to request reinstatement.
[Skip to Content Landing]
Figure.
Mean Symbol Digit Modalities Test (SDMT) Scores for Patients With Pediatric-Onset (POMS) and Adult-Onset (AOMS) Multiple Sclerosis
Mean Symbol Digit Modalities Test (SDMT) Scores for Patients With Pediatric-Onset (POMS) and Adult-Onset (AOMS) Multiple Sclerosis

Lines indicate mean; shaded area, SD; and data points, mean per age. The SDMT scores range from 0 to 120, with higher scores indicating greater information-processing efficiency.

Table 1.  
Comparison of Demographic and Clinical Characteristics of the POMS and AOMS Groups
Comparison of Demographic and Clinical Characteristics of the POMS and AOMS Groups
Table 2.  
Multivariable Linear Mixed-Effect Model of SDMT Score in POMS vs AOMS Groups
Multivariable Linear Mixed-Effect Model of SDMT Score in POMS vs AOMS Groups
Table 3.  
Multivariable Linear Mixed-Effect Model of SDMT Score in POMS vs AOMS Groups With an Interaction for Time Included
Multivariable Linear Mixed-Effect Model of SDMT Score in POMS vs AOMS Groups With an Interaction for Time Included
Table 4.  
Unadjusted and Adjusted Odds of Impairment in Information-Processing Efficiency in POMSa
Unadjusted and Adjusted Odds of Impairment in Information-Processing Efficiency in POMSa
1.
Hakim  EA, Bakheit  AMO, Bryant  TN,  et al.  The social impact of multiple sclerosis—a study of 305 patients and their relatives.  Disabil Rehabil. 2000;22(6):288-293. doi:10.1080/096382800296755PubMedGoogle ScholarCrossref
2.
Amato  MP, Portaccio  E, Goretti  B,  et al; TuSCIMS Study Group.  Relevance of cognitive deterioration in early relapsing-remitting MS: a 3-year follow-up study.  Mult Scler. 2010;16(12):1474-1482. doi:10.1177/1352458510380089PubMedGoogle ScholarCrossref
3.
Cortese  M, Riise  T, Bjørnevik  K,  et al.  Preclinical disease activity in multiple sclerosis: a prospective study of cognitive performance prior to first symptom.  Ann Neurol. 2016;80(4):616-624. doi:10.1002/ana.24769PubMedGoogle ScholarCrossref
4.
Waldman  A, Ness  J, Pohl  D,  et al.  Pediatric multiple sclerosis: clinical features and outcome.  Neurology. 2016;87(9)(suppl 2):S74-S81. doi:10.1212/WNL.0000000000003028PubMedGoogle ScholarCrossref
5.
Chabas  D, Green  AJ, Waubant  E.  Pediatric multiple sclerosis.  NeuroRx. 2006;3(2):264-275. doi:10.1016/j.nurx.2006.01.011PubMedGoogle ScholarCrossref
6.
Amato  MP, Krupp  LB, Charvet  LE, Penner  I, Till  C.  Pediatric multiple sclerosis: cognition and mood.  Neurology. 2016;87(9)(suppl 2):S82-S87. doi:10.1212/WNL.0000000000002883PubMedGoogle ScholarCrossref
7.
Smith  A.  Symbol Digit Modalities Test: Manual. Los Angeles: Western Psychological Services; 1982.
8.
Charvet  LE, Beekman  R, Amadiume  N, Belman  AL, Krupp  LB.  The Symbol Digit Modalities Test is an effective cognitive screen in pediatric onset multiple sclerosis (MS).  J Neurol Sci. 2014;341(1-2):79-84. doi:10.1016/j.jns.2014.04.006PubMedGoogle ScholarCrossref
9.
Parmenter  BA, Weinstock-Guttman  B, Garg  N, Munschauer  F, Benedict  RHB.  Screening for cognitive impairment in multiple sclerosis using the Symbol Digit Modalities Test.  Mult Scler. 2007;13(1):52-57. doi:10.1177/1352458506070750PubMedGoogle ScholarCrossref
10.
Piehl  F, Holmén  C, Hillert  J, Olsson  T.  Swedish natalizumab (Tysabri) multiple sclerosis surveillance study.  Neurol Sci. 2011;31(suppl 3):289-293. doi:10.1007/s10072-010-0345-yPubMedGoogle ScholarCrossref
11.
James  G, Witten  D, Hastie  T, Tibshirani  R.  An Introduction to Statistical Learning. Vol 112. New York, NY: Springer; 2013. doi:10.1007/978-1-4614-7138-7
12.
Berrigan  LI, Fisk  JD, Walker  LAS,  et al.  Reliability of regression-based normative data for the oral Symbol Digit Modalities Test: an evaluation of demographic influences, construct validity, and impairment classification rates in multiple sclerosis samples.  Clin Neuropsychol. 2014;28(2):281-299. doi:10.1080/13854046.2013.871337PubMedGoogle ScholarCrossref
13.
Benedict  RHB, DeLuca  J, Phillips  G, LaRocca  N, Hudson  LD, Rudick  R; Multiple Sclerosis Outcome Assessments Consortium.  Validity of the Symbol Digit Modalities Test as a cognition performance outcome measure for multiple sclerosis.  Mult Scler. 2017;23(5):721-733. doi:10.1177/1352458517690821PubMedGoogle ScholarCrossref
14.
Pfeifenbring  S, Bunyan  RF, Metz  I,  et al.  Extensive acute axonal damage in pediatric multiple sclerosis lesions.  Ann Neurol. 2015;77(4):655-667. doi:10.1002/ana.24364PubMedGoogle ScholarCrossref
15.
Ozakbas  S, Kaya  D, Idiman  E.  Early onset multiple sclerosis has worse prognosis than adult onset multiple sclerosis based on cognition and magnetic resonance imaging.  Autoimmune Dis. 2012;2012:563989.PubMedGoogle Scholar
16.
Fenu  G, Lorefice  L, Loi  L,  et al.  Adult brain volume in multiple sclerosis: the impact of paediatric onset.  Mult Scler Relat Disord. 2018;21:103-107. doi:10.1016/j.msard.2018.03.004PubMedGoogle ScholarCrossref
17.
Kail  R, Salthouse  TA.  Processing speed as a mental capacity.  Acta Psychol (Amst). 1994;86(2-3):199-225. doi:10.1016/0001-6918(94)90003-5PubMedGoogle ScholarCrossref
18.
Park  DC, Reuter-Lorenz  P.  The adaptive brain: aging and neurocognitive scaffolding.  Annu Rev Psychol. 2009;60:173-196. doi:10.1146/annurev.psych.59.103006.093656PubMedGoogle ScholarCrossref
19.
Eckert  MA.  Slowing down: age-related neurobiological predictors of processing speed.  Front Neurosci. 2011;5:25. doi:10.3389/fnins.2011.00025PubMedGoogle ScholarCrossref
20.
Akbar  N, Banwell  B, Sled  JG,  et al.  Brain activation patterns and cognitive processing speed in patients with pediatric-onset multiple sclerosis.  J Clin Exp Neuropsychol. 2016;38(4):393-403. doi:10.1080/13803395.2015.1119255PubMedGoogle ScholarCrossref
21.
MacAllister  WS, Christodoulou  C, Milazzo  M, Krupp  LB.  Longitudinal neuropsychological assessment in pediatric multiple sclerosis.  Dev Neuropsychol. 2007;32(2):625-644. doi:10.1080/87565640701375872PubMedGoogle ScholarCrossref
22.
Amato  MP, Goretti  B, Ghezzi  A,  et al; Multiple Sclerosis Study Group of the Italian Neurological Society.  Cognitive and psychosocial features of childhood and juvenile MS.  Neurology. 2008;70(20):1891-1897. doi:10.1212/01.wnl.0000312276.23177.faPubMedGoogle ScholarCrossref
23.
Amato  MP, Goretti  B, Ghezzi  A,  et al; Multiple Sclerosis Study Group of the Italian Neurological Society.  Cognitive and psychosocial features in childhood and juvenile MS: two-year follow-up.  Neurology. 2010;75(13):1134-1140. doi:10.1212/WNL.0b013e3181f4d821PubMedGoogle ScholarCrossref
24.
Amato  MP, Goretti  B, Ghezzi  A,  et al; MS Study Group of the Italian Neurological Society.  Neuropsychological features in childhood and juvenile multiple sclerosis: five-year follow-up.  Neurology. 2014;83(16):1432-1438. doi:10.1212/WNL.0000000000000885PubMedGoogle ScholarCrossref
25.
Till  C, Racine  N, Araujo  D,  et al.  Changes in cognitive performance over a 1-year period in children and adolescents with multiple sclerosis.  Neuropsychology. 2013;27(2):210-219. doi:10.1037/a0031665PubMedGoogle ScholarCrossref
26.
Charvet  LE, O’Donnell  EH, Belman  AL,  et al; US Network of Pediatric MS Centers.  Longitudinal evaluation of cognitive functioning in pediatric multiple sclerosis: report from the US Pediatric Multiple Sclerosis Network.  Mult Scler. 2014;20(11):1502-1510. doi:10.1177/1352458514527862PubMedGoogle ScholarCrossref
27.
Baruch  NF, O’Donnell  EH, Glanz  BI,  et al.  Cognitive and patient-reported outcomes in adults with pediatric-onset multiple sclerosis.  Mult Scler. 2016;22(3):354-361. doi:10.1177/1352458515588781PubMedGoogle ScholarCrossref
28.
Ruano  L, Branco  M, Portaccio  E,  et al.  Patients with paediatric-onset multiple sclerosis are at higher risk of cognitive impairment in adulthood: an Italian collaborative study.  Mult Scler. 2018;24(9):1234-1242. doi:10.1177/1352458517717341PubMedGoogle ScholarCrossref
29.
Kavaliunas  A, Danylaite Karrenbauer  V, Gyllensten  H,  et al.  Cognitive function is a major determinant of income among multiple sclerosis patients in Sweden acting independently from physical disability.  Mult Scler J. 2019;25(1):104-112. doi:10.1177/1352458517740212PubMedGoogle ScholarCrossref
30.
Strober  LB, Christodoulou  C, Benedict  RHB,  et al.  Unemployment in multiple sclerosis: the contribution of personality and disease.  Mult Scler. 2012;18(5):647-653. doi:10.1177/1352458511426735PubMedGoogle ScholarCrossref
31.
Christodoulou  C, Krupp  LB, Liang  Z,  et al.  Cognitive performance and MR markers of cerebral injury in cognitively impaired MS patients.  Neurology. 2003;60(11):1793-1798. doi:10.1212/01.WNL.0000072264.75989.B8PubMedGoogle ScholarCrossref
32.
Deloire  MS, Bonnet  MC, Salort  E,  et al.  How to detect cognitive dysfunction at early stages of multiple sclerosis?  Mult Scler. 2006;12(4):445-452. doi:10.1191/1352458506ms1289oaPubMedGoogle ScholarCrossref
33.
Wijnands  JMA, Kingwell  E, Zhu  F,  et al.  Health-care use before a first demyelinating event suggestive of a multiple sclerosis prodrome: a matched cohort study.  Lancet Neurol. 2017;16(6):445-451. doi:10.1016/S1474-4422(17)30076-5PubMedGoogle ScholarCrossref
Original Investigation
June 17, 2019

Long-term Cognitive Outcomes in Patients With Pediatric-Onset vs Adult-Onset Multiple Sclerosis

Author Affiliations
  • 1Department of Clinical Neuroscience, Karolinska Institutet, Stockholm, Sweden
  • 2Centre for Molecular Medicine, Karolinska Hospital, Stockholm, Sweden
  • 3Department of Psychology, St Francis Xavier University, Antigonish, Nova Scotia, Canada
  • 4Department of Psychiatry, Dalhousie University, Halifax, Nova Scotia, Canada
  • 5Nova Scotia Health Authority, Halifax, Nova Scotia, Canada
JAMA Neurol. 2019;76(9):1028-1034. doi:10.1001/jamaneurol.2019.1546
Key Points

Question  What are the long-term cognitive outcomes of pediatric-onset multiple sclerosis, and are they different from those observed in the more common adult-onset multiple sclerosis?

Findings  This nationwide cohort study included 5704 persons with definite multiple sclerosis (300 of whom had pediatric-onset disease) and 46 429 prospectively collected cognitive test scores. Patients with pediatric-onset multiple sclerosis exhibited reduced information-processing efficiency and were more likely to experience cognitive impairment in this domain than their counterparts with adult-onset disease, independent of age or disease duration.

Meaning  Pediatric-onset multiple sclerosis appears to render persons particularly susceptible to impairments in information processing in adulthood; children and adolescents who develop multiple sclerosis should be monitored closely for cognitive changes and helped to manage the potential challenges that early-onset multiple sclerosis poses for cognitive abilities later in life.

Abstract

Importance  Cognitive impairment in multiple sclerosis (MS) can lead to reduced quality of life, social functioning, and employment. Few studies have investigated cognitive outcomes among patients with pediatric-onset MS (POMS) over the long term.

Objective  To compare long-term information-processing efficiency between patients with POMS and adult-onset MS (AOMS).

Design, Setting, and Participants  This population-based longitudinal cohort study accessed the Swedish MS Registry (SMSreg), which collates information from all 64 neurology clinics in Sweden. Registered cases with definite MS in the SMSreg with an onset before April 15, 2018, and at least 2 Symbol Digit Modalities Test (SDMT) scores recorded were included. Only persons aged 18 to 55 years and with duration of disease of less than 30 years at the time of SDMT administration were included, to ensure comparable ranges between patients with POMS and AOMS. Of 8247 persons with an SDMT recorded in the SMSreg, 5704 met inclusion criteria, 300 (5.3%) of whom had POMS. Data were collected from April 1, 2006, through April 15, 2018 and analyzed from April through August 2018.

Exposures  Pediatric-onset MS (onset <18 years of age) vs AOMS (onset ≥18 years of age).

Main Outcomes and Measures  Information-processing efficiency measured every 6 or 12 months by the SDMT. Linear mixed-effects models were used to compare all available SDMT scores between patients with POMS and those with AOMS. Persons with cognitive impairment (ever vs never) were identified using regression-based norms and compared between POMS and AOMS groups using logistic regression.

Results  Of the 5704 participants, 4015 were female (70.4%), and 5569 had a relapsing-onset disease course (97.6%). Most participants were exposed to a disease-modifying therapy (DMT) during follow-up (98.8%). Median age at baseline for the POMS group was 25.6 years (interquartile range, 21.0-31.7 years) and for the AOMS group, 38.3 years (interquartile range, 31.4-45.2 years). A total of 46 429 unique SDMT scores were analyzed. After adjustment for sex, age, disease duration, disease course, total number of SDMTs completed, oral or visual SDMT form, and DMT exposure, the SDMT score for patients with POMS was significantly lower than that of patients with AOMS (β coefficient, −3.59 [95% CI, −5.56 to −1.54]). The SDMT score for patients with POMS declined faster than that of patients with AOMS (β coefficient, −0.30 [95% CI, −0.42 tp −0.17]). The odds of cognitive impairment were also significantly elevated in the POMS group (odds ratio, 1.44; 95% CI, 1.06-1.98).

Conclusions and Relevance  In adulthood, patients with POMS demonstrated a more rapid reduction in information-processing efficiency over time and were more likely to experience cognitive impairment than patients with AOMS, independent of age or disease duration. Further investigation is required to understand the mechanisms by which early MS onset influences cognitive outcomes.

Introduction

Cognitive impairment is common in persons with multiple sclerosis (MS) and can lead to reduced quality of life, social functioning, and employment.1 Recent evidence suggests that cognitive deficits, once considered a consequence of decreased neurologic ability, can occur early in MS2 and even before the onset of other recognized clinical signs and symptoms.3

Pediatric-onset MS (POMS), occurring at younger than 18 years, represents 2% to 10% of all cases of MS.4 Although POMS and adult-onset MS (AOMS) mirror each other in many ways, POMS has distinctive features. For example, patients with POMS are significantly less likely to develop primary progressive MS, and they exhibit increased inflammation and disease activity early in their disease relative to those with AOMS.5 As noted in a recent review,6 cognitive dysfunction has consistently been found in about one-third of patients with POMS in cross-sectional studies, but longitudinal data on cognitive outcomes is lacking; in particular, few studies have explored cognitive outcomes of patients with POMS who have entered adulthood. We aimed to explore changes in cognitive function over time in adults with POMS relative to adults with AOMS.

Methods

Individuals were identified from the Swedish MS Registry (SMSreg), a national database established in 2001 that collates clinical information from all 64 neurology clinics in Sweden. To be included in our study, participants had to have clinically definite MS, as defined by their neurologist, with an onset before April 15, 2018. The date of MS onset was determined as the date of first recorded clinical manifestation of MS. Persons with an MS onset at younger than 18 years were classified as having POMS, whereas persons with onset at 18 years or older were classified as having AOMS. The study was approved by the Regional Ethical Review Board of Stockholm, and patients provided written informed consent for the collection of their clinical information.

Cognition was measured via the Symbol Digit Modalities Test (SDMT).7 The SDMT is a measure of information-processing efficiency and has been validated as an assessment of cognitive functioning in POMS8 and AOMS.9 Scores range from 0 to 120, with higher scores indicating greater information-processing efficiency. The SDMT was initially introduced for the evaluation of cognitive performance as part of the Immunomodulation and MS Epidemiology (IMSE) study,10 a postmarketing phase 4 surveillance study of newly introduced second-line disease-modifying therapies (DMTs) in Sweden. The IMSE study began in 2006 with the introduction of natalizumab and continues today. All patients with definite MS who initially received or switched to therapy with natalizumab, fingolimod, alemtuzumab, teriflunomide, dimethyl fumarate, rituximab, or ocrelizumab are eligible to participate. In some clinics in Sweden, the SDMT is a standard component of MS clinical care, measured in patients regardless of their participation in the IMSE studies. These scores were also available for analysis through the SMSreg.

For our study, the date of the first SDMT assessment was considered the baseline visit. Because this was a longitudinal study of changes in cognition, persons with less than 2 SDMT scores recorded in the registry were excluded. To create a comparable age range between the POMS and AOMS groups, persons younger than 18 years at baseline were excluded, as were persons older than 55 years at the time of testing. Further, to ensure a comparable range of disease duration between groups, persons with a disease duration of longer than 30 years were excluded. The SDMT was completed as an oral test in most clinics, although Stockholm residents completed the written version of the test. The same version of the SDMT (identical pattern of symbols and digits) was used for all participants and all tests. The SDMT test was administered at baseline (DMT initiation date), at 6 months, and annually thereafter for persons enrolled in IMSE studies.10 In the clinical setting, the SDMT is administered on an approximately annual basis to coincide with a patient’s annual clinic visit. Individuals were followed up until the date of their last recorded SDMT.

Additional information available from the SMSreg included sex, birth date, date of MS onset, date of diagnosis, clinical course at MS onset (relapsing or primary progressive), MS center, region of Sweden, and DMT use (product and start and stop dates). Disease-modifying therapies were categorized as first-line therapies, including interferon betas, glatiramer acetate, teriflunomide, and dimethyl fumarate, or as second-line therapies, including fingolimod, rituximab, alemtuzumab, mitoxantrone hydrochloride, ocrelizumab, and natalizumab.

Statistical Analyses

Data were analyzed from April to August 2018. We compared the baseline clinical and demographic characteristics of the POMS and AOMS groups using the Pearson χ2 test or the Fisher exact test for categorical variables and the unpaired, 2-tailed t test or the Wilcoxon rank sum test for continuous variables.

Linear mixed-effects models were used to compare SDMT scores between groups. This approach allows for the simultaneous analysis of all available SDMT scores while accounting for clustering within persons and MS centers. Covariates that we considered included age (at baseline and time varying), disease duration (at baseline and time varying), disease course at onset, DMT exposure (first- or second-line vs no treatment; time varying), type of SDMT (written or oral), and total SDMTs completed by each person. Time-varying covariates were estimated at the time of each SDMT assessment. We compared models, including interaction terms using the Akaike information criterion and reported the most parsimonious model, with results presented as β coefficients with 95% CIs. To assess for collinearity, we measured the variance inflation factor in all models using a predetermined threshold of 5 as suggestive of multicollinearity.11 As a complementary analysis, we matched all cases of POMS with 5 cases of AOMS with the nearest disease duration and reran the final models to ensure our results were not unduly influenced by the differences in disease duration between groups.

In addition, we used a modified version of published regression-based norms for the oral SDMT and classified individuals as cognitively impaired or unimpaired.12 The original method incorporated sex, age, and educational level; because we did not have information regarding educational level, a modified algorithm was developed using the same cohort as in the original publication.12 Details of this process are available in the eMethods in the Supplement, including details of the regression model with and without educational level (eTable). We used logistic regression with the outcome ever cognitively impaired or never cognitively impaired during the follow-up period to compare the POMS and AOMS groups. All analyses were adjusted for the total amount of follow-up time contributed by each individual. Stockholm residents were excluded from this analysis because they completed the written form of the SDMT. Findings were reported as odds ratios with 95% CIs.

Results

Of 8247 persons with an SDMT recorded in the SMSreg before April 15, 2018, 5704 met inclusion criteria (300 with POMS and 5407 with AOMS; 1689 male [29.6%] and 4015 female [70.4%]). A total of 46 429 unique SDMT scores were analyzed. The median number of SDMTs contributed was 5 (IQR, 3-9); the mean (SD) baseline SDMT score was 51.0 (12.2); and the median follow-up time was 3.0 years (IQR, 1.6-5.2 years). Of the total cohort, 1644 (28.8%; 1552 with AOMS and 92 with POMS) completed the written form of the SDMT. The mean (SD) baseline SDMT score was lower among persons who completed the written form compared with the oral form (50.0 [12.0] vs 51.0 [12.7]; P = .003). No differences in sex ratio, disease course, region of residence, or baseline SDMT score were found between the POMS and AOMS groups. By definition, the patients with POMS were younger at onset and diagnosis. Compared with the AOMS group, the POMS group was also younger at the study baseline (median age, 25.6 [IQR, 21.0-31.7] vs 38.3 [IQR, 31.4-45.2] years; P < .001) and had a longer disease duration (median, 10.2 [IQR, 5.3-15.7] vs 5.2 [IQR, 1.7-11.0] years; P < .001) at baseline. The POMS group was more likely to have received only second-line therapy (226 [75.3%] vs 3421 [63.3%] for AOMS), whereas AOMS had a higher proportion of patients receiving first-line therapy (1048 [19.4%] vs 32 [10.7%] for POMS) during follow-up (P < .001) (Table 1). A total of 3647 participants in both groups (63.9%) had been exposed to a second-line DMT, while only 67 (1.2%) of the full cohort were never exposed to a DMT during follow-up.

At younger than 30 years, SDMT scores between the POMS and AOMS groups were comparable; but after 30 years of age the trajectories began to diverge, as seen in the Figure (part A). At 35 years of age, the mean (SD) SDMT score for AOMS was 60.9 (17.1), and for POMS it was 51.1 (12.6). By 40 years of age, the mean (SD) AOMS score was 57.5 (16.7), whereas it was 46.4 (14.9) for the POMS group (Figure, A). When plotted against disease duration, the POMS and AOMS groups had similar SDMT trajectories (Figure, B).

In a simple longitudinal model including only pediatric- vs adult-onset status and disease duration, no difference in SDMT score occurred between groups (β coefficient, −0.60; 95% CI, −2.39 to 1.18; P = .51). Including sex and time-varying age and disease duration in the model led to a β coefficient of −3.56 (95% CI, −5.59 to −1.53) for pediatric- vs adult-onset cases. In a multivariable model that included sex, time-varying age and disease duration, disease course, total number of SDMTs completed, type of SDMT, and time-varying DMT exposure, the SDMT scores for patients with POMS were significantly lower than those of patients with AOMS (β coefficient, −3.59 [95% CI, −5.56 to −1.54]) (Table 2). The complementary matched analysis included 300 patients with POMS and 1500 with AOMS; together, the groups had a median disease duration of 10.3 years (IQR, 5.3-15.8 years). In a model adjusted for sex, age, disease duration, disease course, total number of SDMTs completed, type of SDMT, and DMT exposure, the β coefficient for POMS vs AOMS was −3.87 (95% CI, −5.85 to −1.90).

A model that incorporated an interaction term between time (from baseline to each SDMT) and pediatric- vs adult-onset status showed (model 1) a significant association of POMS with SDMT scores over time. After adjustment for sex, age, disease course, total number of SDMTs completed, type of SDMT, and DMT exposure, the POMS × time interaction term was significant (β coefficient, −0.29; 95% CI, −0.42 to −0.17). Similar results were found when adjusting for disease duration as opposed to age (model 2) (Table 3).

The logistic regression analyses, which excluded Stockholm residents, included 208 patients with POMS and 3852 with AOMS. The proportion of patients who ever met the definition for cognitive impairment in the domain of information-processing efficiency was 147 of 208 (70.7%) for POMS and 2305 of 3852 (59.8%) for AOMS. After adjustment for disease course, disease duration at baseline, DMT use at baseline, and duration of follow-up, the odds of being cognitively impaired in this domain were significantly elevated in the POMS group relative to the AOMS group (odds ratio, 1.44; 95% CI, 1.06-1.98) (Table 4).

Discussion

Using nationwide, prospectively collected information, we found that persons with an onset of MS at younger than 18 years exhibited greater decline in information-processing efficiency in adulthood than those who developed MS as adults. During the median follow-up time of 3.0 years in adulthood, 70.7% of patients with POMS and 59.8% of patients with AOMS met the definition for cognitive impairment in this domain. The SDMT scores for the POMS group were significantly lower than those for the AOMS group (β coefficient, −3.59), and on further analysis, we found that this effect was modified by time. Specifically, the SDMT scores of patients with POMS consistently declined faster than those of patients with AOMS (β coefficient, −0.30) independent of age, disease duration, or treatment status.

Prior validation of the SDMT in the context of MS identified vocational benchmarks associated with SDMT scores.13 Notably, the POMS sample reached the SDMT anchor for “employed but work challenged” (score of 55) at approximately 34 years of age whereas the AOMS sample reached this same threshold a mean of 16 years later, at 50 years of age, suggesting that the differences observed between the POMS and AOMS groups in this study may be meaningful.

The differences observed in information-processing efficiency between the POMS and AOMS groups may be a consequence of the heightened inflammation and axonal loss that has been reported in POMS.14 Inflammation of the brain during critical developmental periods,4 including myelinogenesis in adolescence, may irreparably damage neural networks involved in cognition.15 This damage may also lead to the reduced brain and deep gray matter volumes in adulthood that have been reported in POMS relative to sex- and age-matched patients with AOMS independent of disease duration.16 Both groups showed the highest SDMT scores from 20 to 30 years of age, which is typical of the cognitive trajectory over the lifespan.17 We suspect that this phenomenon may also reflect practice effects, because we see both groups improve initially from their baseline SDMT. During the initial stages of POMS and AOMS among younger patients, regardless of disease duration, compensatory mechanisms in brain functional reorganization akin to brain “scaffolding”18 may allow for improved SDMT performance, particularly with practice. However, as the limits of such compensatory mechanisms are reached with increased age,19 those with POMS may become subject to the effects of brain aging at an accelerated rate, as seen in the Figure (part A). Such an interpretation is, of course, speculative and would require further study. Prior studies provide evidence for compensatory activity in POMS in the form of greater activation of certain brain regions associated with faster response times on a functional magnetic resonance imaging SDMT task compared with controls, interpreted as suggestive of an “adaptive mechanism that may contribute to limiting the impact of disease-related structural pathology.”20(p393)

Male sex and a primary progressive disease course were consistently associated with worse cognitive outcomes in both groups, and these patients should be monitored closely for cognitive changes. The association of DMT exposure with information-processing efficiency was less clear. Although there appeared to be a positive association with exposure to second-line therapies, any such interpretation should be cautious. This was not a study of DMT effectiveness, and we did not have information on DMT adherence, nor did we adjust for indication bias.

To date, few studies have prospectively examined cognitive functioning in persons with POMS.6 In 1 such study,21 12 children with MS from New York State were assessed at baseline and a mean of 22 months later. The authors noted that cognitive impairment was common and that approximately half of their sample demonstrated further decline at follow-up.21 Among 63 patients with MS with childhood and juvenile onset recruited from 11 MS centers across Italy,22 young age at onset was significantly associated with lower IQ scores. When this same cohort underwent retesting 2 years later, 75% (42 of 56) exhibited cognitive deterioration.23 Five years later, 56% of the 48 patients still enrolled showed cognitive deterioration from baseline.24 In a study from a single clinic in Toronto, Ontario, Canada,25 7 of 28 patients with POMS and 1 of 26 matched healthy controls exhibited cognitive deterioration during a 1-year period, whereas the healthy participants were more likely to show improvements over time. By contrast, a study of 67 patients with POMS from 9 centers across the United States26 who completed 2 cognitive assessments that included 9 tests a mean of 1.6 years apart found that the proportion of persons who were cognitively impaired remained stable (ie, 37% at baseline and 33% at follow-up). Despite this relative stability, however, these authors concurred with the opinion that POMS “may exact its greatest cost by inhibiting age-expected gains.”26(p7)

Two cross-sectional studies27,28 have directly compared cognition in POMS and AOMS, and both have reported findings concordant with our own. The first study included 51 patients with POMS and 550 with AOMS enrolled in a single MS center in Massachusetts and found that SDMT scores were 7.57 points lower for patients with POMS than for those with AOMS, after adjustment for age.27 The second study included 119 patients with POMS and 712 with AOMS recruited from 6 MS centers in Italy.28 After adjustment for age, the authors reported that the odds of impairment on at least 1 of 2 tests of information-processing efficiency, one of which was the SDMT, were 1.86 times higher among patients with POMS than among those with AOMS.28

Limitations

The SDMT is the most commonly used cognitive measure in adults with MS and is strongly associated with income,29 employment,30 and magnetic resonance imaging markers of disability,31 suggesting that it is an appropriate metric. Moreover, in a comparison with various other neuropsychological tests relative to a full neuropsychological test battery, the SDMT performed the best in identifying cognitive impairment in patients with MS, with a sensitivity of 76.5% and specificity of 87.0%.32 Because it can be administered efficiently and in any language, the SDMT offers a pragmatic method for monitoring cognition across various clinical settings. Despite its advantages, however, the SDMT is limited in addressing only information-processing efficiency, so we are unable to report on differences in other elements of cognition, such as memory or executive functioning. As with other neuropsychological tests, practice effects are common for the SDMT. This was apparent in our cohort, as many participants showed improved scores over time. This improvement may have also been influenced by the use of the same SDMT version at each testing. This ability to learn and improve over time, especially considering the periods of 6 to 12 months between testing, may reflect an effective cognitive ability in itself. Nonetheless, to ensure that our findings were not unduly influenced by practice effects, we only included persons with a minimum of 2 SDMTs recorded, and we adjusted models for the total number of SDMTs completed by each participant.

Our study has other limitations. The true date of MS onset was largely based on self-reported symptoms, and we now recognize that disease onset may occur many years before the first identifiable clinical signs.33 As such, some persons may have been misclassified as having POMS or AOMS. However, given the large study sample, we do not expect that this misclassification would have affected our results in a meaningful way. Further, this type of misclassification is expected to be nondifferential and therefore would have biased our estimates toward the null.

Most of the participants in our sample completed the SDMT as part of their involvement in the IMSE study to monitor newly approved MS drugs in Sweden. Therefore, our study population is representative primarily of treated patients. Nevertheless, the differences we observed are likely to be robust, because the inclusion of DMT use in the models did not alter the results. As previously stated, we lacked information on highest educational level achieved and thus were unable to examine for or adjust our models to account for the association of POMS with educational attainment. However, although earlier onset of MS may have influenced educational attainment, lower levels of education attained need not specifically reflect cognitive impairment and may indicate problems attending school owing to medical care needs and fatigue.2

Conclusions

This study provides, to our knowledge, the first population-based and longitudinal evidence of the cognitive consequences of POMS for adults. Adults who had developed POMS had significantly higher odds of impairment and greater decline over time in information-processing efficiency than their counterparts with AOMS, independent of age or disease duration. Cognition is a critical component of a person’s ability to work and engage in society, as well as their overall quality of life. Children and adolescents who develop MS should be monitored closely for cognitive changes and helped to manage the difficulties and challenges that MS poses on scholastic and work-related achievements, with a view to the long-term consequences of MS as they reach adulthood. Further investigation is required to better understand the biological and psychosocial mechanisms by which an early onset of MS influences cognitive outcomes and to establish appropriate interventions and treatment guidelines.

Back to top
Article Information

Accepted for Publication: January 21, 2019.

Corresponding Author: Jan Hillert, MD, PhD, Department of Clinical Neuroscience, Karolinska Institutet, Widerströmska Huset, Plan 5, Tomtebodavägen 18A, Stockholm, Sweden 17177 (jan.hillert@ki.se).

Published Online: June 17, 2019. doi:10.1001/jamaneurol.2019.1546

Author Contributions: Drs McKay and Manouchehrinia had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: McKay, Manouchehrinia, Hillert.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: McKay.

Critical revision of the manuscript for important intellectual content: Manouchehrinia, Berrigan, Fisk, Olsson, Hillert.

Statistical analysis: McKay, Manouchehrinia.

Obtained funding: McKay, Olsson, Hillert.

Administrative, technical, or material support: Manouchehrinia, Olsson, Hillert.

Supervision: Manouchehrinia, Hillert.

Conflict of Interest Disclosures: Dr McKay reported receiving research support from the Canadian Institutes of Health Research (CIHR) and European Committee for Treatment and Research in Multiple Sclerosis (ECTRIMS). Dr Manouchehrinia reported receiving speaker honoraria from Biogen. Dr Berrigan reported receiving research funding from the Canada Foundation for Innovation, the CIHR, and the Social Sciences and Humanities Research Council of Canada. Dr Fisk reported receiving research funding from the CIHR, Multiple Sclerosis Society of Canada, Nova Scotia Health Authority Research Fund, and Dalhousie Medical Research Foundation and consultation and distribution royalties from Mapi Research Trust. Dr Olsson reported receiving unrestricted research grants or honoraria for lectures or advisory boards from Biogen, Novartis, Merck & Co, Roche, and Sanofi Genzyme. Dr Hillert reported receiving honoraria for serving on advisory boards for Biogen, Sanofi Genzyme, and Novartis and speaker’s fees from Biogen, Novartis, Merck Serono, Bayer Schering Pharma AG, Teva Pharmaceutical Industries, Ltd, and Sanofi Genzyme and serving as principal investigator for projects or receiving unrestricted research support from Biogen Idec, Inc, Merck Serono, Teva Pharmaceutical Industries, Ltd, Sanofi Genzyme, and Bayer Schering Pharma AG. No other disclosures were reported.

Funding/Support: This study was supported by the Swedish Research Council and the Swedish Brain Foundation and by postdoctoral fellowship awards from the CIHR (Dr McKay) and ECTRIMS (Dr McKay).

Role of the Funder/Sponsor: The sponsors had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Additional Information: Data related to the current article are available from Dr Hillert, Karolinska Institutet. To access data from the Swedish MS Registry, a data transfer agreement needs to be completed between Karolinska Institutet and the institution requesting data access. This is in accordance with the data protection legislation in Europe (General Data Protection Regulation). Persons interested in obtaining access to the data should contact Jan Hillert (jan.hillert@ki.se).

References
1.
Hakim  EA, Bakheit  AMO, Bryant  TN,  et al.  The social impact of multiple sclerosis—a study of 305 patients and their relatives.  Disabil Rehabil. 2000;22(6):288-293. doi:10.1080/096382800296755PubMedGoogle ScholarCrossref
2.
Amato  MP, Portaccio  E, Goretti  B,  et al; TuSCIMS Study Group.  Relevance of cognitive deterioration in early relapsing-remitting MS: a 3-year follow-up study.  Mult Scler. 2010;16(12):1474-1482. doi:10.1177/1352458510380089PubMedGoogle ScholarCrossref
3.
Cortese  M, Riise  T, Bjørnevik  K,  et al.  Preclinical disease activity in multiple sclerosis: a prospective study of cognitive performance prior to first symptom.  Ann Neurol. 2016;80(4):616-624. doi:10.1002/ana.24769PubMedGoogle ScholarCrossref
4.
Waldman  A, Ness  J, Pohl  D,  et al.  Pediatric multiple sclerosis: clinical features and outcome.  Neurology. 2016;87(9)(suppl 2):S74-S81. doi:10.1212/WNL.0000000000003028PubMedGoogle ScholarCrossref
5.
Chabas  D, Green  AJ, Waubant  E.  Pediatric multiple sclerosis.  NeuroRx. 2006;3(2):264-275. doi:10.1016/j.nurx.2006.01.011PubMedGoogle ScholarCrossref
6.
Amato  MP, Krupp  LB, Charvet  LE, Penner  I, Till  C.  Pediatric multiple sclerosis: cognition and mood.  Neurology. 2016;87(9)(suppl 2):S82-S87. doi:10.1212/WNL.0000000000002883PubMedGoogle ScholarCrossref
7.
Smith  A.  Symbol Digit Modalities Test: Manual. Los Angeles: Western Psychological Services; 1982.
8.
Charvet  LE, Beekman  R, Amadiume  N, Belman  AL, Krupp  LB.  The Symbol Digit Modalities Test is an effective cognitive screen in pediatric onset multiple sclerosis (MS).  J Neurol Sci. 2014;341(1-2):79-84. doi:10.1016/j.jns.2014.04.006PubMedGoogle ScholarCrossref
9.
Parmenter  BA, Weinstock-Guttman  B, Garg  N, Munschauer  F, Benedict  RHB.  Screening for cognitive impairment in multiple sclerosis using the Symbol Digit Modalities Test.  Mult Scler. 2007;13(1):52-57. doi:10.1177/1352458506070750PubMedGoogle ScholarCrossref
10.
Piehl  F, Holmén  C, Hillert  J, Olsson  T.  Swedish natalizumab (Tysabri) multiple sclerosis surveillance study.  Neurol Sci. 2011;31(suppl 3):289-293. doi:10.1007/s10072-010-0345-yPubMedGoogle ScholarCrossref
11.
James  G, Witten  D, Hastie  T, Tibshirani  R.  An Introduction to Statistical Learning. Vol 112. New York, NY: Springer; 2013. doi:10.1007/978-1-4614-7138-7
12.
Berrigan  LI, Fisk  JD, Walker  LAS,  et al.  Reliability of regression-based normative data for the oral Symbol Digit Modalities Test: an evaluation of demographic influences, construct validity, and impairment classification rates in multiple sclerosis samples.  Clin Neuropsychol. 2014;28(2):281-299. doi:10.1080/13854046.2013.871337PubMedGoogle ScholarCrossref
13.
Benedict  RHB, DeLuca  J, Phillips  G, LaRocca  N, Hudson  LD, Rudick  R; Multiple Sclerosis Outcome Assessments Consortium.  Validity of the Symbol Digit Modalities Test as a cognition performance outcome measure for multiple sclerosis.  Mult Scler. 2017;23(5):721-733. doi:10.1177/1352458517690821PubMedGoogle ScholarCrossref
14.
Pfeifenbring  S, Bunyan  RF, Metz  I,  et al.  Extensive acute axonal damage in pediatric multiple sclerosis lesions.  Ann Neurol. 2015;77(4):655-667. doi:10.1002/ana.24364PubMedGoogle ScholarCrossref
15.
Ozakbas  S, Kaya  D, Idiman  E.  Early onset multiple sclerosis has worse prognosis than adult onset multiple sclerosis based on cognition and magnetic resonance imaging.  Autoimmune Dis. 2012;2012:563989.PubMedGoogle Scholar
16.
Fenu  G, Lorefice  L, Loi  L,  et al.  Adult brain volume in multiple sclerosis: the impact of paediatric onset.  Mult Scler Relat Disord. 2018;21:103-107. doi:10.1016/j.msard.2018.03.004PubMedGoogle ScholarCrossref
17.
Kail  R, Salthouse  TA.  Processing speed as a mental capacity.  Acta Psychol (Amst). 1994;86(2-3):199-225. doi:10.1016/0001-6918(94)90003-5PubMedGoogle ScholarCrossref
18.
Park  DC, Reuter-Lorenz  P.  The adaptive brain: aging and neurocognitive scaffolding.  Annu Rev Psychol. 2009;60:173-196. doi:10.1146/annurev.psych.59.103006.093656PubMedGoogle ScholarCrossref
19.
Eckert  MA.  Slowing down: age-related neurobiological predictors of processing speed.  Front Neurosci. 2011;5:25. doi:10.3389/fnins.2011.00025PubMedGoogle ScholarCrossref
20.
Akbar  N, Banwell  B, Sled  JG,  et al.  Brain activation patterns and cognitive processing speed in patients with pediatric-onset multiple sclerosis.  J Clin Exp Neuropsychol. 2016;38(4):393-403. doi:10.1080/13803395.2015.1119255PubMedGoogle ScholarCrossref
21.
MacAllister  WS, Christodoulou  C, Milazzo  M, Krupp  LB.  Longitudinal neuropsychological assessment in pediatric multiple sclerosis.  Dev Neuropsychol. 2007;32(2):625-644. doi:10.1080/87565640701375872PubMedGoogle ScholarCrossref
22.
Amato  MP, Goretti  B, Ghezzi  A,  et al; Multiple Sclerosis Study Group of the Italian Neurological Society.  Cognitive and psychosocial features of childhood and juvenile MS.  Neurology. 2008;70(20):1891-1897. doi:10.1212/01.wnl.0000312276.23177.faPubMedGoogle ScholarCrossref
23.
Amato  MP, Goretti  B, Ghezzi  A,  et al; Multiple Sclerosis Study Group of the Italian Neurological Society.  Cognitive and psychosocial features in childhood and juvenile MS: two-year follow-up.  Neurology. 2010;75(13):1134-1140. doi:10.1212/WNL.0b013e3181f4d821PubMedGoogle ScholarCrossref
24.
Amato  MP, Goretti  B, Ghezzi  A,  et al; MS Study Group of the Italian Neurological Society.  Neuropsychological features in childhood and juvenile multiple sclerosis: five-year follow-up.  Neurology. 2014;83(16):1432-1438. doi:10.1212/WNL.0000000000000885PubMedGoogle ScholarCrossref
25.
Till  C, Racine  N, Araujo  D,  et al.  Changes in cognitive performance over a 1-year period in children and adolescents with multiple sclerosis.  Neuropsychology. 2013;27(2):210-219. doi:10.1037/a0031665PubMedGoogle ScholarCrossref
26.
Charvet  LE, O’Donnell  EH, Belman  AL,  et al; US Network of Pediatric MS Centers.  Longitudinal evaluation of cognitive functioning in pediatric multiple sclerosis: report from the US Pediatric Multiple Sclerosis Network.  Mult Scler. 2014;20(11):1502-1510. doi:10.1177/1352458514527862PubMedGoogle ScholarCrossref
27.
Baruch  NF, O’Donnell  EH, Glanz  BI,  et al.  Cognitive and patient-reported outcomes in adults with pediatric-onset multiple sclerosis.  Mult Scler. 2016;22(3):354-361. doi:10.1177/1352458515588781PubMedGoogle ScholarCrossref
28.
Ruano  L, Branco  M, Portaccio  E,  et al.  Patients with paediatric-onset multiple sclerosis are at higher risk of cognitive impairment in adulthood: an Italian collaborative study.  Mult Scler. 2018;24(9):1234-1242. doi:10.1177/1352458517717341PubMedGoogle ScholarCrossref
29.
Kavaliunas  A, Danylaite Karrenbauer  V, Gyllensten  H,  et al.  Cognitive function is a major determinant of income among multiple sclerosis patients in Sweden acting independently from physical disability.  Mult Scler J. 2019;25(1):104-112. doi:10.1177/1352458517740212PubMedGoogle ScholarCrossref
30.
Strober  LB, Christodoulou  C, Benedict  RHB,  et al.  Unemployment in multiple sclerosis: the contribution of personality and disease.  Mult Scler. 2012;18(5):647-653. doi:10.1177/1352458511426735PubMedGoogle ScholarCrossref
31.
Christodoulou  C, Krupp  LB, Liang  Z,  et al.  Cognitive performance and MR markers of cerebral injury in cognitively impaired MS patients.  Neurology. 2003;60(11):1793-1798. doi:10.1212/01.WNL.0000072264.75989.B8PubMedGoogle ScholarCrossref
32.
Deloire  MS, Bonnet  MC, Salort  E,  et al.  How to detect cognitive dysfunction at early stages of multiple sclerosis?  Mult Scler. 2006;12(4):445-452. doi:10.1191/1352458506ms1289oaPubMedGoogle ScholarCrossref
33.
Wijnands  JMA, Kingwell  E, Zhu  F,  et al.  Health-care use before a first demyelinating event suggestive of a multiple sclerosis prodrome: a matched cohort study.  Lancet Neurol. 2017;16(6):445-451. doi:10.1016/S1474-4422(17)30076-5PubMedGoogle ScholarCrossref
×