A 3-year longitudinal study of cognitive impairment in patients with primary progressive multiple sclerosis: Speed matters

A 3-year longitudinal study of cognitive impairment in patients with primary progressive multiple sclerosis: Speed matters
of 8
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  A 3-year longitudinal study of cognitive impairment in patients with primary progressive multiple sclerosis: Speed matters Douglas R. Denney  a, ⁎ , Sharon G. Lynch  b , Brett A. Parmenter   a  a   Department of Psychology, University of Kansas, 1415 Jayhawk Blvd., Lawrence, KS 66045-7556, USA  b  Department of Neurology, University of Kansas Medical Center, Kansas City, KS, USA Received 1 June 2007; received in revised form 4 October 2007; accepted 5 October 2007Available online 31 October 2007 Abstract  Background:  The few controlled longitudinal studies of cognitive performance in MS patients all provide evidence of deterioration in at least a subset of the patients sampled. Only one of these studies has focused on primary progressive MS, and little attention has been paid to thespecific domains of cognitive functioning that change over time. The present study examined three principal cognitive domains in samples of  primary progressive MS patients and healthy controls followed over a period of 3 years.  Methods:  A battery of neuropsychological tests that included measures of strategic problem solving, verbal memory, and information processing speed was administered annually to 24 MS patients and 25 controls.  Results:  MS patients' performance on measures of processing speed showed significantly greater decline over the 3-year period than did that of controls. Similar results were not observed in the case of problem solving or verbal memory. There was no evidence of more dramaticdecline occurring in patients who were initially classified as cognitively impaired relative to those who were unimpaired at baseline.However, this failure may have been influenced by differential attrition from the sample; more impaired patients were less likely to completethe study. Conclusion:  Overall the results support the contention that information processing speed is the domain most sensitive to the impact of multiple sclerosis on cognitive functioning over time.© 2007 Elsevier B.V. All rights reserved.  Keywords:  Multiple sclerosis; Cognition; Neuropsychology; Information processing speed; Longitudinal study 1. Introduction Cognitive impairment in patients with multiple sclerosis(MS) is generally viewed as affecting a wide assortment of cognitive domains such as executive functions, verbal andspatial memory, visual – spatial processing, complex atten-tion, verbal fluency, and speeded information processing[1 – 5]. Previously we reported a study [6] comparing relapsing – remitting or primary progressive MS patients andhealthy controls on a battery of neuropsychological teststhat included measures of strategic problem solving, verbalmemory, and information processing speed. When thegroups were equated with respect to age, gender, educationlevel, fatigue and depression through covariance analysis,the only cognitive measures on which patients differed fromcontrols were those relating to the speed of information processing. Slowing in patients' performance was evident regardless of whether speed was an explicit feature of thetask at hand (e.g., the three trials of the Stroop Color  – WordInterference Test) or was recorded unobtrusively while the patient concentrated on arriving at the correct solution to a problem (e.g., initial planning times on the separate problems composing the Tower of London). An exploratoryfactor analysis of the cognitive measures yielded threefactors, but processing speed accounted for the greatest  proportion of the variance and was the only factor distinguishing patients from controls. Journal of the Neurological Sciences 267 (2008) 129 – ⁎  Corresponding author. Tel.: +1 785 864 9805; fax: +1 785 864 5696.  E-mail address: (D.R. Denney).0022-510X/$ - see front matter © 2007 Elsevier B.V. All rights reserved.doi:10.1016/j.jns.2007.10.007  The primary progressive patients in the preceding studywere enrolled in the PROMiSe Trial [7], a multi-center, 3-year, double blind, placebo controlled clinical trial evaluat-ing the effectiveness of glatiramer acetate (Copaxone®) intreating this particular form of MS. The 33 patients enrolledat our center were also asked to volunteer for a separateinvestigation of their cognitive functioning over the courseof the 3-year period. Likewise, most of the healthy controlsubjects in our aforementioned study were recruited for a 3-year longitudinal study wherein the battery of neuropsycho-logical tests would be re-administered annually.Longitudinal studies provide the best source of dataconcerning the natural history of patients' cognitivefunctioning. A number of such studies have been published[8]. Their findings are highly inconsistent, with some studiesdocumenting cognitive decline in at least a subset of the patients evaluated [9 – 18] and others offering little or noevidence of decline [19 – 29].The conflicting results may stem from numerous differ-ences between these studies in the composition of the patient sample, the measures included in the cognitive battery, thefrequency and duration over which assessments occur. Themost obvious distinction however is one of research design.The majority of these studies [12,13,17,21,23,24,26 – 28]were uncontrolled and focused on the relationship betweenchanges in patients' cognitive performance over time and brain imaging measures. Only six published reports of longitudinal studies [9 – 11,14 – 16] include a control groupthat was evaluated contemporaneously with the patient sample, and all of these studies provide evidence of cognitivedeterioration in at least a subset of the patients. The studiesare noteworthy from the standpoint of the patient samplesevaluated. Amato et al. [9,10] recruited only recentlydiagnosedpatientsandhavemanagedtofollowthesepatientsover the longest period (i.e., 10 years). Kujala et al. [16] presented evidence that patients who were already classifiedas cognitively impaired at baseline had more extensivecognitive decline over the 3 years of their study. Jennekens-Schinkel et al. [15] showed that such patients were also morelikely to drop out of longitudinal studies, thereby probablyattenuating the extent of cognitive deterioration evident over time. Huijbregts et al. [14] were the only investigators tocompare cognitive decline in patients with different subtypesof MS. Because they specifically examined changes in patients with primary progressive MS, the study byHuijbregts et al. [14] as well as an uncontrolled longitudinalstudy by Camp et al. [12] is the most relevant to the present work and will be considered in greater detail in theDiscussion section. 2. Methods 2.1. Subjects The 33 patients who were enrolled in the PROMiSe Trialat the University of Kansas Medical Center were also invitedto participate in the present study. Only one patient declinedthis invitation. In addition, one patient died and nine patientsceased participating in the PROMiSe Trial before the end of the full 3-year period. Two of these latter patients completedall but the last cognitive assessment, and their final year datawere estimated through imputation by multiple regression.The final sample consisted of 24 patients (12 males, 12females) with primary progressive MS. The patients rangedin age from 30 to 63 (  M  =50.3). Education level was rated ona 6-point scale (1 = did not finish high school, 2 = completedhigh school, 3 = completed some college credit, 4 =completed 4-year college degree, 5 = completed somegraduate work, and 6 = completed an advanced degree); patients' ratings ranged from 2 to 6 (  M  =3.5). Duration of illness ranged from 1 to 21 years (  M  =5.4), and initial scoreson the Expanded Disability Status Scale (EDSS) [30])ranged from 3.5 to 6.5 (Mdn=3.5,  M  =4.2).A healthy control group consisting of 31 subjects whowere free of chronic medical conditions was recruited for thisstudy through newspaper ads, posters, and contacts with personnel at the medical center. Over the course of the study,eight control subjects withdrew. Two of these subjectscompleted all but the last assessment, and their final year data were estimated through imputation. The final sampleconsisted of 25 controls (8 males, 17 females) ranging in agefrom 23 to 57 (  M  =44.4). Their ratings of education levelranged from 2 to 6 (  M  =4.2). 2.2. Measures The computerized cognitive battery has been describedelsewhere [6] and will only be summarized here. The Tower of London  (TOL) is a test of planning andstrategic problem solving consisting of 12 problems withthree trials permitted to solve each problem. In addition tocompiling a total point score for the problems, the computer measured the  “ initial planning time ”  for each problem, theinterval between the initial presentation of the problem andthe subject's first move  —  regardless of whether the subject succeeded in solving the problem during this initial attempt. The Wisconsin Card Sorting Test   (WCST). The WCST isa concept formation test consistingof cards that can be sortedaccording to color, shape, or number. The concept uponwhich the cards were to be sorted changed during the courseof the test. The computer displayed the cards, providedfeedback as to whether each sorting response by the subject was  “ right  ”  or   “ wrong, ”  and tallied the number of cardssorted, the total number of errors, and the percentage of  perseverative errors committed during the test. The Stroop Color   –  Word Interference Test   (Stroop). TheStroop consisted of three 60-second trials during which thesubject read color words (word reading), then named the color of a row of   X  's (color naming), and finally, named the color of the letters of color words (color  – word naming). The subject responded verbally to each stimulus as it appeared, and theexperimenter pressed the space bar to display the next  130  D.R. Denney et al. / Journal of the Neurological Sciences 267 (2008) 129  –  136   stimulus. The computer timed the trial and recorded thenumberofstimulicompletedduringthetrial.Inadditiontotheseparate scores for each trial, a score reflecting resistance tointerference by the incongruent stimulus information on thecolor  – word naming trial was derived using the formularecommended by Golden [31]. The Paired Associates Learning Test   (PALT). The PALTassessed both immediate recall and delayed recognitionmemory for verbal stimuli. A randomized sequence of eight related (e.g.,  “ APPLE – WORM ” ) and eight unrelated (e.g., “ GLASSES – BUS ” ) word pairs was displayed. Subjectswere then shown one word from each pair and asked to recallthe word with which it was paired. The computer recordedthe number of related and unrelated words correctly recalled,and the procedure was repeated a second time. After a delayduring which the Stroop was administered, subjects wereshown a randomized sequence of 32 words from the srcinalword pairs and 32 foils. The computer recorded the number of words correctly identified during this delayed recognitiontrial. Cognitive index . Scores on each of the 12 individualmeasures derived from the cognitive battery were convertedto standard scores using the mean and standard deviation for the full sample at baseline (i.e., Year B) on that measure.Each subject's standard scores on the 12 measures in a givenyear were then averaged, and the average transformed to ascale with a mean of 100 and a standard deviation set at 10.The resulting cognitive index provided a measure of thesubject's overall performance on the cognitive battery duringeach year of participation in the study, indexed to the baseline performance of the full sample of 49 subjects whoeventually completed the study. Self-report Questionnaires . At each annual testingsession, subjects' levels of fatigue and depression over thecourse of the preceding week were assessed using theFatigue Severity Scale (FSS) [32] and the Center for Epidemiologic Studies  —  Depression Scale (CES-D) [33]. 2.3. Procedure This study was approved by the Human SubjectsCommittee of the University of Kansas Medical Center. As part of the PROMiSe Trial protocol, appointments in the MSClinic were conducted with the patients every 3 months.During the course of these appointments, patients wereadministered the EDSS and a battery of other neurologicaltests to monitor the status of their disease. Becausecomparable data were not collected for the control subjects,these measures (except for the patients' annual EDSSratings) are omitted. The cognitive battery and the fatigueand depression questionnaires were administered annually.The testing session usually lasted about 90 min. Thequestionnaires were administered first, followed by thecognitive measures in a fixed sequence consisting of theTOL, immediate recall portion of the PALT, Stroop, delayedrecognition portion of the PALT, and the WCST. Subjectswere tested either in the clinic or in their homes. The testingsite was chosen according to each subject's convenience andwas maintained consistently throughout the study. Hometesting for patients was conducted within 2 weeks of their annual clinic visit. 2.4. Design The study conformed to a 2 (Group)×4 (Year) mixedfactorial design, with repeated measures on the second factor.Wheneversignificantdifferencesbetweenthegroupsoccurredon extraneous variables that could potentially confound theresults of the cognitive measures, these variables were enteredas covariates. 3. Results 3.1. Comparisons between patients who withdrew early and  patients with full data The baseline data for the eight patients who started thestudy but withdrew before the end of the 3 years werecompared with that of the 24 patients with evaluable data at the end of the study. The two groups did not differ in gender,age,education,fatigue,ordepression.However,patientswhowithdrew early had a longer duration of MS (  M  =9.3 vs. 5.4, t  =1.86,  df   =30,  p =.07) and higher scores on the EDSS(  M  =5.2 vs. 4.2,  t  =2.31,  df   =30,  p = .03). They also hadlower scores on the cognitive index (  M  =90.6 vs. 97.5, t  =3.59,  df   =30,  p =.001), performing more poorly thancontrols on 8 of the 12 measures derived from the cognitive battery  —  including initial planning times on the TOL,color  – wordnamingand interferenceon theStroop;unrelatedrecall and delayed recognition scores on the PALT, andnumber of errors, percentage of perseverative errors, andtrials to solution on the WCST. In contrast to these findingsfor the MS patients, the six healthy control subjects whowithdrew did not differ from those who completed the studyin age, education level, depression, fatigue, or their initialscores on any of the cognitive measures. 3.2. Comparisons between patients and controls The 24 MS patients and 25 controls with evaluable datadid not differ with respect to gender. However, the patientswere older on average (  M  =50.3 vs. 44.4,  t  =2.36,  df   =47,  p =.02) and had lower ratings on education level (  M  =3.5vs. 4.2,  t  =2.02,  df   =47,  p =.05). Because cognitive perfor-mance could well be affected by both age and education,these variables were used as covariates in the analyses of thecognitive measures.Table 1 presents the scores for fatigue, depression, and allthe cognitive measures recorded annually over the course of the study. A 2 (Group)×4 (Year) analysis of covariance wasappliedtoeachofthesemeasures,withageandeducationlevelincluded as covariates. There were no significant main effects 131  D.R. Denney et al. / Journal of the Neurological Sciences 267 (2008) 129  –  136   for Year. Significant main effects for Group were found on theoverall cognitive index (  F  =15.4,  df   =1&45,  p b .001, η 2 =.26),initialplanningtimesontheTOL(  F  =17.8, df   =1&45,  p b .001, η 2 =.28), word reading (  F  =6.9,  df   =1&45,  p =.01,  η 2 =.13),color naming (  F  =13.2,  df   =1&45,  p =.001,  η 2 =.23), color  – word naming (  F  =15.8,  df   =1&45,  p b .001,  η 2 =.26), andresistance to interference on the Stroop (  F  =6.3,  df   =1&45,  p =.02,  η 2 =.12), and unrelated recall (  F  =9.9,  df   =1&44,  p =.003,  η 2 =.18) and delayed recognition on the PALT(  F  =6.5, df   =1&45,  p =.014, η 2 =.13).Onthreeofthemeasuresderived from the Stroop, the main effect for Group wasqualified by a significant Group×Year interaction: wordreading (  F  =5.6,  df   =3&43,  p =.003,  η 2 =.28); color naming(  F  =7.7,  df   =3&43,  p b .001, η 2 =.35); and color  – word naming(  F  =5.7,  df   =3&43,  p =.002,  η 2 =.28). These interactionsresulted because the declines in patients' scores on thesemeasures were greater than those of controls. Education levelwas a significant covariate in the analysis of initial planningtimes on the TOL. Age was a significant covariate in theanalysis of the cognitive index, and the word reading, color naming, and color  – word naming scores derived from theStroop.Throughout the study, patients and controls also differed interms of fatigue and depression. The 2 (Group)×4 (Year)analyses performed on these scores revealed significant maineffects for Group on both fatigue (  F  =33.5,  df   =1&45,  p b .001, η 2 =.43) and depression (  F  =7.8,  df   =1&45,  p =.007,  η 2 =.15).Thus, all the analyses of the cognitive measures were repeatedusing fatigue and depression scores (averaged across the years)along with age and education as covariates. The outcome wasessentiallyunchanged;allthesignificantmaineffectsforGroupand the Group×Year interactions reported in the preceding paragraph continued to be significant when depression andfatigue were added as covariates. Likewise, the significant covariates remained confined to education and age; neither fatigue nor depression was significant in these analyses.In our earlier study using a larger sample of MS patients[6], a factor analysis performed on the scores derived fromthe cognitive battery yielded three factors: information processing speed, strategic problem solving, and verbalmemory. Using the factor loadings from this earlier study, weconverted the current scores on the cognitive battery intofactor scores and then subjected these factor scores to thesame 2 (Group)×4 (Year) analysis of covariance applied tothe individual cognitive scores, with age and educationserving as covariates. The results revealed a significant maineffect for Group (  F  =19.8,  df   =1&45,  p b .001,  η 2 =.31) anda significant Group×Year interaction (  F  =7.7,  df   =3&43,  p b .001,  η 2 =.35) for information processing speed. Therewas also a significant main effect for Group on verbalmemory (  F  =13.7,  df   =1&45,  p =.001,  η 2 =.23), but theGroup×Year interaction in this case did not attain signifi-cance (  F  =2.1,  df   =3&43,  p =.11,  η 2 =.13). There were nosignificant main effects or interactions for the problemsolving factor (all  p 's N .16). These results were unchangedwhen depression and fatigue were added to the analyses of covariance. The only significant covariate in any of theseanalyses was age, which emerged as a covariate in theanalysis of processing speed. The covariate adjusted meansfor the factors scores are presented graphically in Fig. 1.Because the overall cognitive index and the three factor scores adequately summarize the results stemming from thecognitive battery, only these scores are presented in theanalyses applied to subgroups of patients appearing in theremainder of the results. 3.3. Comparisons between initially impaired and unimpaired  patients Each of the MS patients' factor scores at baseline wereconverted into  z   scores, and these  z   scores were used, alongwiththecognitiveindexatbaseline,todividethepatientsintoinitially  “ impaired ”  and  “ unimpaired ”  subgroups. The nine patients (7 males, 2 females) in the impaired group scored below 95 of the cognitive index and also scored greater than Table 1Comparison of MS patients (P) and healthy controls (C) over timeMeasure Year B (  M  ±SD) 1 (  M  ±SD) 2 (  M  ±SD) 3 (  M  ±SD)Fatigue P 44.0±12.4 43.1±14.0 45.0±12.2 43.5±13.1C 26.0±10.3 26.2±11.6 26.3±13.9 27.0±13.0Depression P 12.8±8.3 11.5±7.2 12.8±9.9 10.8±7.6C 7.5±6.3 7.3±7.5 8.9±10.1 6.4±6.5Cognitive measuresTower of LondonPoint score P 31.6±2.9 33.5±2.5 32.9±2.7 33.5±2.6C 33.0±2.3 33.6±2.6 33.4±2.2 34.3±1.8Planning time P 22.6±6.7 23.9±7.5 22.3±10.4 21.3±9.5C 16.1±5.0 18.0±7.2 14.0±4.8 14.9±4.6Wisconsin Card Sorting Test Total errors P 28.4±17.7 26.3±17.6 24.2±17.7 21.2±12.9C 24.2±21.9 18.4±11.2 18.0±11.9 14.2±5.7Perseverativeerrors (%)P 48.3±13.6 43.5±14.5 43.0±15.2 41.1±17.2C 41.4±13.0 38.6±13.3 37.0±16.1 35.3±16.0# of cards tosolutionP 101.7±22.6 100.9±23.2 97.9±23.8 98.1±21.7C 92.5±19.2 92.8±20.0 87.9±19.8 87.8±17.9Stroop Color  – Word Test Word reading P 78.0±8.0 78.6±11.3 72.1±11.9 69.4±11.2C 90.2±11.9 79.0±12.2 81.7±10.7 82.7±10.2Color naming P 63.7±7.5 64.0±8.2) 59.4±9.0 58.3±9.5C 73.1±9.2 68.9±8.9) 69.8±7.2 70.2±7.9Color  – WordnamingP 46.7±5.2 47.9±6.2 44.7±5.6 44.0±7.0C 54.5±7.8) 52.8±6.9 54.4±6.0 54.8±8.3Resistance tointerferenceP 11.7±3.7 12.8±4.3 12.2±4.7 12.4±4.9C 14.2±5.4 16.0±5.0 16.8±4.2 16.8±5.9Paired Associates Learning Test RecallRelateditemsP 14.4±2.3 15.2±1.5 15.0±1.4 15.3±1.0C 14.1±2.1 14.6±1.6 15.2±1.2 15.3±1.2UnrelateditemsP 8.0±4.4 8.1±4.5 7.6±4.1 8.4±4.5C 10.4±4.4 11.6±4.4 12.7±3.6 12.3±3.7DelayedrecognitionP 61.2±3.0 60.7±3.9 61.5±2.4 62.1±2.7C 61.8±2.7 62.6±1.4 63.3±1.3 63.3±1.3 Note: All means are unadjusted  —  computed prior to covariance analyses.132  D.R. Denney et al. / Journal of the Neurological Sciences 267 (2008) 129  –  136   one standard deviation below the mean on at least one of thethree factors. The 15 patients (5 males, 10 females) in theunimpaired group all scored above 93 on the cognitive index,and none of these patients scored one standard deviation below the mean on any of the factors. No significant differences were found between the two impairment groupsin terms of gender, age, education, or duration of MS.The data for the impaired and non-impaired groups on themeasures collected repeatedly in this study are presented inTable 2. A 2 (Impairment Group)×4 (Year) analysis of variance was applied to each of these measures. Nosignificant differences were found for fatigue or depression. Not surprisingly, the impaired group had significantly higher EDSS scores (  F  =7.4,  df   =1&21,  p =.013,  η 2 =.26) andlower scores on the cognitive index (  F  =14.1,  df   =1&22,  p =.001,  η 2 =.39) compared to the unimpaired group.Significant main effects for both Group and Year occurredin the analyses applied to the factor scores for processingspeed factor and verbal memory. Impaired patients hadsignificantly lower overall scores than the unimpaired patients on both factors (processing speed:  F  =9.6, df   =1&22,  p =.005,  η 2 =.30; verbal memory:  M  =66.2 vs.73.7,  F  =15.6,  df   =1&22,  p =.001,  η 2 =.41), and, over thecourse of the 3-year study, patients' scores declinedsignificantly on processing speed (  F  =10.0,  df   =3&20,  p b .001,  η 2 =.60), but rose significantly on verbal memory(  F  =4.9,  df   =3&20,  p =.01,  η 2 =.42). However, in all of these analyses, the interaction between Group and Year wasnonsignificant; change over time was essentially the samefor both initially impaired and unimpaired patients. 3.4. Comparisons between patients on Copaxone vs. placebo Following removal of the data blind for the PROMiSeTrial, we determined that 16 patients (8 males, 8 females)had been on Copaxone and 8 patients (4 males, 4 females) Table 2Comparisons between subgroups of patientsMeasure SubgroupImpaired(  N  =9)  M  ±SDUnimpaired(  N  =15)  M  ±SDEDSS 4.8±1.0 3.9±0.7 ⁎ Fatigue 8.4±8.5 41.2±11.8Depression 13.4±7.1 11.1±7.6Cognitive index 94.4±3.4 100.1±3.7 ⁎⁎⁎ Information processing speed 138.1±20.5 160.2±14.6 ⁎⁎ Verbal memory 66.2±4.3 73.7±4.6 ⁎⁎⁎ Strategic problem solving 106.8±24.5 121.8±34.1Measure Copaxone(  N  =16)  M  ±SDPlacebo(  N  =8)  M  ±SDEDSS 4.3±0.9 4.2±0.9Fatigue 45.4±8.7 41.0±15.1Depression 12.6±8.0 10.8±6.2Cognitive index 98.9±4.3 96.1±4.5Information processing speed 156.4±18.5 142.9±20.8Verbal memory 71.0±5.7 70.6±6.2Strategic problem solving 122.3±29.0 103.9±33.5 ⁎  p =.01;  ⁎⁎  p =.005;  ⁎⁎⁎  p =.001.Fig. 1. Factor scores on (a) information processing speed, (b), verbalmemory, and (c) strategic problem solving over time for MS patients andhealthy controls.133  D.R. Denney et al. / Journal of the Neurological Sciences 267 (2008) 129  –  136 
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks