The rate of conversion of mild cognitive impairment to dementia: predictive role of depression

INTERNATIONAL JOURNAL OF GERIATRIC PSYCHIATRY

Int J Geriatr Psychiatry 2007; 22: 563–567.

Published online 29 November 2006 in Wiley InterScience

(www.interscience.wiley.com) DOI: 10.1002/gps.1716

The rate of conversion of mild cognitive impairmentto dementia: predictive role of depression

T. Gabryelewicz1*, M. Styczynska1, E. Luczywek1, A. Barczak1, A. Pfeffer1, W. Androsiuk2,M. Chodakowska-Zebrowska1, B. Wasiak1, B. Peplonska1 and M. Barcikowska1

1Department of Neurodegenerative Disorders, Medical Research Centre Polish Academy of Sciences, Warsaw, Poland2Department of Psychiatry, Medical University of Warsaw, Warsaw, Poland

SUMMARY

Background Mild cognitive impairment (MCI) is a condition referring to the persons with cognitive deficits measurable insome form or another, but not meeting criteria for dementia, and who have an increased risk of becoming demented.Objective To establish the rate of progression to dementia in MCI, to investigate the risk of conversion for amnesticvs multiple-domains subtypes, and to identify the predictors of progression.Methods MCI (n¼ 105) individuals enrolled in a longitudinal study received annual clinical and psychometric exam-inations for up to a mean of 3 years. The diagnosis of MCI according to Mayo Clinic Petersen’s Criteria was conducted by apanel of specialists.Results After 3 years of follow-up, 23 of 105 subjects with MCI were diagnosed with dementia. 40 showed cognitivedecline not dementia, 34 were stable and showed no cognitive decline or improvement, while eight showed cognitiveimprovement.Conclusions We conclude that conversion rate from MCI to DSM-IIIR dementia was 21.9% over a period of 3 years. Theoccurrence of depressive symptomsmay constitute a predictor for thosewho are more likely to progress to dementia. The riskof conversion to dementia was higher among the subjects with an evidence of impairment extending beyond memory thanwith those who suffered only from memory deficits, and the subjects who converted to dementia in this subtype hadsignificantly higher baseline plasma total homocysteine levels than non-converters. Copyright # 2006 John Wiley &Sons, Ltd.

key words—mild cognitive impairment; dementia; depression

INTRODUCTION

The significance of cognitive decline in the elderly hasbeen widely discussed in the literature. Variousapproaches have been proposed to describe thisdecline and its relationship with the developmentof dementia (Dawe et al., 1992). Unfortunately,the difficulty in disentangling ‘normal’, purelyage-associated cognitive changes from changes

*Correspondence to: Dr T. Gabryelewicz, Department of Neurode-generative Disorders, Medical Research Centre, Polish Academy ofSciences. 5 Pawinskiego St., 02-106 Warsaw, Poland.E-mail: [email protected]

Copyright # 2006 John Wiley & Sons, Ltd.

related to early degenerative or vascular diseaseshas resulted in a lack of agreement on terminology andspecific diagnostic criteria. This may explain manyconflicting results in the studies devoted to cognitivelyimpaired non-demented subjects. Although there are anumber of different definitions of mild cognitivedeficits, the general concept is of a subjective memoryimpairment, and/or another domain in the context ofcognitive impairment relative to age-matched controlsand yet showing no loss of function and no dementia.The first precise definition of MCI was proposed in1997 by the Mayo Clinic (Petersen et al., 1997). In2004 an international panel of experts recommendedthe newly revised and standardized Petersen’s criteria

Received 14 July 2006Accepted 11 October 2006

564 t. gabryelewicz ET AL.

for the diagnosis and classification of MCI andsubtypes in research and clinical settings (Winbladet al., 2004).Individuals with a MCI are at an increased risk of

developing dementia ranging from 1–25% per year. Ina systematic review of conversion studies publishedbetween 1991–2001 the mean annual conversion ratewas 10.24% although with substantial differencesbetween studies (Bruscoli and Lovestone, 2004).There is considerable heterogeneity in the rates ofconversion in these studies, with annual conversionrates ranging from 2–31%. The overall 10.24%conversion rate is five times the expected incidenceof dementia in people at this age. The field of MCIresearch is currently focusing on identifying the riskfactors of disease progression for the purpose of earlytherapeutic intervention, which may in turn delay oreven prevent the onset of dementia.The aim of this study was to establish the rate of

progression to dementia in MCI, to investigate the riskof conversion for amnestic vs multiple-domainssubtypes, and to identify the predictors of progression.

MATERIAL AND METHODS

Our sample included 105 MCI individuals who wereenrolled between May 2001–May 2002 in a longi-tudinal study at the Department of NeurodegenerativeDisorders of the Polish Academy of Sciences (DND/PAN) in Warsaw. The subjects were selectedconsecutively from attenders who had come to theDND/PAN for an evaluation of cognitive difficulties.They were either referred by their family doctors,family members, or by themselves. All subjects livedindependently in the community at the time of theirbaseline evaluation. Informed consent was obtainedfrom each participant or their relatives. The localEthics Committee for Medical Research approved thestudy.All patients were seen by psychiatrist and neurol-

ogist who obtained a medical history from the subjectas well as corroborating sources, performed theMini-Mental State Examination (MMSE; Folsteinet al., 1999), the Clock Drawing Test (Sunderlandet al., 1989), the Montgomery-Asberg DepressionRating – MADRS (Montgomery and Asberg, 1979),and a psychiatric and neurological examination.Persons with severe major depressive episode, definedaccording to DSM-IV criteria, which may havecontributed to the observed cognitive difficulties wereexcluded. The maximum MADRS score was 28,equivalent to mild and moderate depression. Individ-uals with other medical, psychiatric, or neurological


diagnosis that might interfere with performance on thepsychometric measures were excluded from the study.

Laboratory studies were performed, including thechemistry group, complete blood cell count, thesedimentation rate, vitamin B12 and folic acid levels,concentration of plasma total homocysteine (tHcy), andsensitive thyroid-stimulating hormone level. APOEgenotyping was performed blind to clinical diagnosis,using a slightly modified method of restrictiongenotyping described by Chapman et al. (1996). Allpatients received a head imaging study (computedtomography or magnetic resonance imaging).

Sessions of neuropsychological testing were com-pleted on all patients. The participants completed aspecial set of tests. Obtained results were comparedwith estimated evaluation of premorbid level ofcognitive functioning. Working memory wasmeasured by a Backward Digit Span from theWechsler Adult Intelligence Scale – WAIS (Wechsler,1987). Semantic memory was assessed by theSimilarities subtest of WAIS. Episodic memory wasassessed by the auditory learning of ten words withdelayed recall and recognition (Luria, 1973), and theComplex Rey Figure Reproduction after distraction(Osterrieth, 1944). Language abilities were measuredby verbal fluency and naming of 20 pictures from theBoston Naming Test (Kaplan et al., 1983). Visuo-construction and praxis were measured by theComplex Rey Figure Copy. Executive functions wereassessed in every conducted test as the ability to plan,efficiency and control. The final attentional test wasthe Trail-Making Test Part A & Part B (Reitan, 1958).The outcome for each patient at 3 years wasdetermined according to neuropsychologist evaluationbased on the comparison between the scores ofconducted tests, as well as observations and interviewswith subject and subject’s caregiver.

The diagnosis of MCI was made by a panel ofspecialists if the subject met the following criteria-similar to those initially proposed by Petersen et al.(1997): the presence of memory complaints, normalactivities of daily living, objective memory impair-ment or an impairment in another area of cognitivefunction, normal global cognitive function, notdemented. Stages of the severity of cognitivedisturbances was determined by the Clinical DementiaRating (CDR) conducted according to the publishedrules (Morris, 1993). The MCI participants were ratedat entry as CDR 0.5.

MCI was further classified into two subtypes:amnestic MCI (A-MCI), if there was prominentmemory impairment with relative preservation of func-tion in other cognitive domains; multiple-domains


DOI: 10.1002/gps

the rate of conversion of mci to dementia 565

MCI (MD-MCI) if there was impairment in memory,and in one or more non-memory cognitive domains.Diagnosis was made for dementia according toThe Diagnostic and Statistical Manual of MentalDisorders, 3rd edn, revised (American PsychiatricAssociation, 1987).

Participants were reassessed every 12 months. Theirperformance was reviewed by a panel of specialistsand the diagnosis was adjusted accordingly, ifnecessary.

Statistical analysis was performed with the use ofSAS 9.1 (SAS Institute Inc., Cary, NC, USA). Thedifferences between two groups were analyzed usingWilcoxon rank sum test (Mann–Whitney) for thequantitative variables, and Chi-Square test for thecategorical variables. The differences betweentree groups were analyzed using Kruskal-Wallis testor Chi-Square test. In all analyses p-values< 0.05were considered to be statistically significant.

RESULTS

Subjects with the assessment performed at least twotimes were included in the analysis reported here.During the 3-year follow-up, two subjects died, sevenrefused to continue to participate, two of whomrefused prior to the first follow-up assessment andwere not included in these analyses. At baseline, theconsecutively recruited participants were divided intotwo groups based on their cognitive status. Forty-twosubjects were classified as A-MCI and 63 as MD-MCI.

The multiple-domain group was statistically sig-nificantly older than the amnestic group, had a higherpercentage of APOE e4 carriers, higher mean plasmatHcy levels, and a higher mean MADRS score. Thecharacteristics of the sample at baseline are presentedin Table 1.

After 3 years of follow-up 42 subjects remained thesame or improved (eight), and 63 had progressive

Table 1. Baseline characteristics of the study sample by MCI subty

Total A

No. of patients 105 4Mean age, years (SD) 69.3 (7.2) 66.Female, % (n) 67.6 (71) 6Education �7 years, % (n) 81.9 (86) 90.Mean MMSE score (SD) 27.2 (1.8) 27.Mean MADRS score (SD) 9.8 (5.8) 8.APOE e4 carriers, % (n) 24.8 (26) 14.Homocysteine, umol/l (SD) 14.3 (4.2) 12.

*Wilcoxon Test; all other values – Chi-Square test.


cognitive disturbances, including 23 who converted todementia. Nineteen of the 23 converted dementiacases had possible AD (A-MCI 5, MD-MCI 14), twowere classified as VD (MD-MCI), one as mixeddementia – AD and VD (MD-MCI), and one as FTD(MD-MCI). An overall rate of conversion to dementiawas 21.9% at 3 years (annual rate 7.3% – calculated bydividing the observed conversion rate by the follow-uptime). The mean time for follow-up was 3.05 years,SD 0.675 (median, 3.2 years).All patients (n¼ 105) were categorized into three

groups based on their 3-year period of cognitive change:the stable group, the progressed group (no dementia)and the converted to dementia group, see Table 2.The mean MADRS score was statistically signifi-

cantly higher in the progressed and demented vs thestabile group. Relationship between baseline MADRSscore and two factors – MCI subtypes, and cognitivechange outcome was assessed using analysis of cova-riance (ANCOVA) with age as covariate. Type IIIsum of squares was taken into consideration.Homogeneity of regression slopes assumption wastested to confirm appropriateness of an ANCOVAanalysis. Analysis revealed significant relationshipbetween higher baseline MADRS score, progressionto dementia (F¼ 4.83, p¼ 0.010) and MD-MCIsubtype (F¼ 6.26, p¼ 0.014).Patients with MD-MCI were statistically signifi-

cantly more likely to convert to dementia than thosewith A-MCI. We also found that the subjects withMD-MCI who converted to dementia had statisticallyhigher baseline mean plasma tHcy levels then non-converters (16.53 umol/l vs 14.36 umol/l; U¼ 713,p¼ 0.037).

DISCUSSION

During a 3-year follow-up 23 incidence cases ofdementia developed while 44.8% of patients remained

pes

-MCI MD-MCI Statistics ( p-value)

2 63 —2 (7.85) 71.4 (5.9) p¼ 0.001*9 (29) 66.7 (42) p> 0.055 (38) 76.2 (48) p> 0.055 (1.85) 27.1 (1.7) p> 0.05*0 (5.3) 10.9 (5.9) p¼ 0.011*3 (6) 31.75 (20) p¼ 0.0428 (3.7) 15.2 (4.3) p¼ 0.001


DOI: 10.1002/gps

Table 2. Baseline characteristics of the study sample by outcome

Characteristics Total Stable Progressed (no dementia) Demented Statistics ( p-value)

No. of patients, % (n) 105 40 (42) 38.1 (40) 21.9 (23) —Mean age, years (SD) 69.3 (7.2) 68.6 (7) 68.7 (6.6) 71.7 (8.2) p> 0.05*Female, % (n) 67.6 (71) 66.7 (28) 72.5 (29) 60.9 (14) p> 0.05Education � 7 years, % (n) 81.9 (86) 83.3 (35) 85 (34) 73.9 (17) p> 0.05Mean MMSE score (SD) 27.2 (1.8) 27.4 (1.5) 27.4 (1.9) 26.6 (2.1) p> 0.05*Mean MADRS score (SD) 9.8 (5.8) 7.4 (5.4) 11.3 (5.6) 11.5 (6.9) p¼ 0.001*APOE e4 carriers, % (n) 24.8 (26) 16.7 (7) 25 (10) 39.1 (9) p> 0.05Homocysteine, umol/l (SD) 14.3 (4.2) 14.0 (4.4) 13.7 (4.2) 15.3 (3.8) p> 0.05MD-MCI, % (n) 60 (63) 42.9 (18) 67.5 (27) 78.3 (18) p¼ 0.009

*Kruskal–Wallis test; all other values – Chi-Square test.

KEY POINTS

� MCI was a significant predictor of dementiawithin a 3-year period, with the 21.9% conver-sion rate, and the occurrence of depressivesymptoms may predict which patients are morelikely to progress to dementia.

566 t. gabryelewicz ET AL.

the same or improved (7.6%), and 55.2% hadprogressive cognitive disturbances, including thosewho converted to dementia. The status of theseindividuals after 3 years of annual follow-upemphasizes the heterogeneity that existed amongthem. Our findings are similar to those presented insome other studies. In a 3-year longitudinal study byRitchie et al. (2001) the annual conversion rate was5.6%. Daly et al. (2000) found dementia in 23 (19%)of 123 subjects with CDR rating, 0.5 (annualconversion rate 6.3%). Grober et al. (2000) estimateda 6.2% MCI annual conversion rate. In a study byTiernay et al. (1996) the studied annual conversionrate was 10.9% over 2 years, and in Petersen et al.(1999) the studied conversion rate was 12% per yearfor 4 years. In a study by Ravaglia et al. (2006) 29% ofMCI subjects converted to dementia over 3 years offollow-up.One of the most important aims of the clinical

evaluation of patients with MCI is to predict who willdevelop clear dementia, and to identify the predictorsof a fast progression.In our study, MD-MCI subjects were more likely to

convert to dementia, than those with A-MCI. Similarresults, revealing that patients with evidence ofimpairment extending beyond memory are morelikely to have dementia than those with only memorydeficits were presented by Bozoki et al. (2001).Our multiple-domain group had a higher percentage

of APOE e4 carriers. A few studies have suggestedthat MCI sufferers carrying the APOE e4 allele aremore likely to progress to AD than non-carriers(Petersen et al., 1995). We also found significantlyhigher mean plasma tHcy levels among MD-MCI vsA-MCI patients, and higher baseline mean plasmatHcy levels amongMD-MCI converters. A few studiesshowed that homocysteine levels correlate inversely


with cognitive performance (Duthie et al., 2002), andan increase in plasma homocysteine is an independentrisk factor for the development of AD (Seshadri et al.,2002).

Early depressive symptoms among subjects withMCI and the lack of improvement in cognition scores,following the treatment of depression in MCI patientsmay be associated with an increased risk of meetingdiagnostic criteria for dementia during the follow-up (Liet al., 2001; Devanand et al., 2003). We found that atbaseline, the prevalence of depressive symptomsmeasured by MADRS was significantly lower amongstable vs progressive and demented patients, andanalysis revealed significant relationship betweenhigher baseline MADRS score, progression to dementiaand MD-MCI subtype. Our findings show that theoccurrence of depressive symptoms may be of help inpredicting progression and conversion to dementia.

We conclude that the currently used diagnosiscriteria for MCI characterized a heterogeneouspopulation of patients with an increased risk ofdeveloping dementia. In our study the risk ofconversion to dementia was higher among theindividuals with MD-MCI, and the subjects whoconverted to dementia in MD-MCI subtype hadstatistically significant higher baseline mean plasmatHcy levels than non-converters.


DOI: 10.1002/gps

the rate of conversion of mci to dementia 567

ACKNOWLEDGEMENTS

The study was supported by a grant 2PO5B12227 forthe Study of MCI, from KBN-State Committee ForScientific Research.

REFERENCES

American Psychiatric Association. 1987. Diagnostic and StatisticalManual of Mental Disorders, 3rd edn, revised. American Psy-chiatric Association: Washington, DC.

Bozoki A, Giordani B, Heidebrink J, et al. 2001. Mild cognitiveimpairments predict dementia in nondemeneted elderly patientswith memory loss. Arch Neurol 58: 411–416.

Bruscoli M, Lovestone S. 2004. Is MCI really just early dementia?A systematic review of conversion studies. Int Psychogeriatr 16:129–140.

Chapman J, Estupinan J, Asherow A, Goldfarb L. 1996. A simpleand efficient method for apolipoprotein E genotype determi-nation. Neurology 46: 1484–1485.

Daly E, Zaitchik D, Copeland M, et al. 2000. Predicting conversionto Alzheimer disease using standardized clinical information.Arch Neurol 57: 675–680.

Dawe B, Procter A, Philpot M. 1992. Concepts of mild memoryimpairment in the elderly and their relations to dementia: areview. Int J Geriatr Psychiatry 7: 473–479.

Devanand DP, Pelton GH, Marston K, et al. 2003. Sertralinetreatment of elderly patients with depression and cognitiveimpairment. Int J Geriatr Psychiatry 18: 123–130.

Duthie SJ, Whalley LJ, Collins AR, et al. 2002. Homocysteine, Bvitamin status, and cognitive function in the elderly. Am J ClinNutr 75: 908–913.

Folstein MF, Folstein SE, McHugh PR. 1975. ‘Mini-Mental State’:a practical method for grading the cognitive state of patients forthe clinician. J Psychiatr Res 12: 189–198.

Grober E, Lipton RB, Hall C, Crystal H. 2000. Memory impairmenton free and cued selective reminding predicts dementia. Neurol-ogy 54: 827–832.

Kaplan EF, Goodglass H, Weintraub S. 1983. Naming The BostonTest. Lea & Febiger: Philadelphia, PA.


Li YS, Meyer JS, Thorby J. 2001. Longitudinal follow-up ofdepressive symptoms among normal versus cognitively impairedelderly. Int J Geriatr Psychiatry 16: 718–727.

Luria AR. 1973. The Working Brain: An Introduction to Neurop-sychology. Basic Books: New York.

Montgomery SA, Asberg M. 1979. A new depression Scale designedto be sensitive to change. Br J Psychiatry 134: 382–389.

Morris JC. 1993. The Clinical Dementia Rating (CDR): currentversion and scoring rules. Neurology 43: 2412–2414.

Osterrieth PA. 1944. Le test de copie d’une figure complexe.Archives de Psychologie 30: 206–356.

Petersen RC, Smith GE, Ivnik RJ, et al. 1995. Apolipoprotein Estatus as a predictor of the development of Alzheimer’sdisease in memory-impaired individuals. JAMA 273: 1274–1278.

Petersen RC, Smith GE, Waring SC, et al. 1997. Ageing, memory,and Mild Cognitive Impairment. Int Psychogeriatr 9: 65–70.

Petersen RC, Smith GE, Waring SC, et al. 1999. Mild cognitiveimpairment: clinical characterization and outcome. Arch Neur 56:303–308.

Ravaglia G, Forti P, Maioli F, et al. 2006. Conversion of mildcognitive impairment to dementia: predictive role of mild cog-nitive impairment subtypes and vascular risk factors. DementGeriatr Cogn Disord 21: 51–58.

Reitan RM. 1958. Validity of trail making test as an indication oforganic brain damage. Percept Mot Skills 8: 271–276.

Ritchie K, Artero S, Touchon J. 2001. Classification criteria for mildcognitive impairment. A population-based validation study. Neu-rology 56: 37–42.

Seshadri S, Beiser A, Selhub J, et al. 2002. Plasma homocysteine asa risk factor for dementia and Alzheimer’s disease. N Engl J Med346: 476–483.

Sunderland T, Hill JL, Mellow AM, et al. 1989. Clock Drawing inAlzheimer’s disease. A novel measure of dementia severity. J AmGeriatr Soc 37: 725–729.

Tierney MC, Szalai JP, Snow WG, et al. 1996. Prediction ofprobable Alzheimer’s disease in memory-impaired patients: aprospective longitudinal study. Neurology 46: 661–665.

Wechsler D. 1981. WAIS-R manual: Wechsler Adult IntelligenceScale – Revised. The Psychological Corporation: New York.

Winblad B, Palmer K, Kipivelto M, et al. 2004. Mild cognitiveimpairment – beyond controversies, towards a consensus: reportof the International Working Group on Mild Cognitive Impair-ment. J Intern Med 256: 240–246.


DOI: 10.1002/gps

Documents

The rate of conversion of mild cognitive impairment to dementia: predictive role of depression