Cognitive screening instruments for the diagnosis of mild cognitive impairment

Transcription

1 MCI diagnosis Original research Cognitive screening instruments for the diagnosis of mild cognitive impairment Andrew J Larner MD, MRCP Mild cognitive impairment (MCI) may represent an early, and potentially treatable, phase of dementing disorders. Its correct clinical identification is therefore of paramount importance. Here, Dr Larner analyses data from a selection of short cognitive screening instruments (CSIs) employed in a secondary care setting to measure various parameters for diagnosis of MCI and to establish whether any of the instruments have superior diagnostic utility. Introduction Early diagnosis of cognitive disorders would seem to offer the best chance for their effective treatment, as in other neurological disorders and, indeed, in other spheres of medicine. To this end, the definition of disease biomarkers, which reflect the clinicobiological features of dementing diseases, and their incorporation in diagnostic criteria, 1 offers perhaps the best possibility of early diagnosis, possibly even in the presymptomatic stages of disease. For Alzheimer s disease (AD), this presymptomatic phase may last for decades. 2,3 Likewise, the early symptomatic phases of dementing disorders might also represent a significant opportunity for treatment, particularly if disease-modifying therapies can be discovered and brought to the clinical arena. Some consensus has grown up around the concept of mild cognitive impairment (MCI), not only as a prodromal phase of AD, but also of the cognitive impairment associated with other brain disorders, which can progress to dementia, including Parkinson s disease 4 and cerebrovascular disease (vascular cognitive impairment), 5 and possibly also frontotemporal dementia. 6 The category of MCI thus has the advantage of flexibility, accepting that it encompasses significant heterogeneity both with respect to aetiology and prognosis (conversion rate to dementia). 7 Other terminology may also sometimes be used to denote such patients, such as mild cognitive disorder (as in ICD10), cognitive impairment no dementia, 8 and mild cognitive dysfunction. 9 Although some criteria have sought to eschew MCI as a category relevant to AD, in favour of a more definitive diagnosis of early AD, other criteria persist with MCI. 10 The most recent AD diagnostic criteria are partly based around sophisticated neuroimaging and cerebrospinal neurochemistry biomarkers, tapping into the clinicobiological aspects of disease. 1 However, such investigations may not be readily available outside of specialist centres, and hence the simpler cognitive screening instruments (CSIs), which may be used to supplement clinical history taking, may still be the initial investigation in patients suspected of dementia or MCI or in patients with memory complaints. Hence it is pertinent to examine the screening utility of these CSIs for the early stages of disease when disease-modifying therapy might have the best chance of efficacy. Although, as their nomenclature indicates, these are screening and not diagnostic tests, nevertheless they are often subjected to test accuracy studies using essentially the same methodology as for diagnostic tests. 11 As proof-of-concept, a meta-analysis of studies of the Mini-Mental State Examination, perhaps the most commonly used CSI and a benchmark for other such instruments, has shown its diagnostic validity in MCI. 12 The aim of the study presented here was to collate information from several pragmatic studies of short CSIs (ie those taking approximately 5 10 minutes to administer), undertaken in a secondary care setting (a dedicated cognitive disorders clinic based in a regional neuroscience centre), to examine their screening utility for the diagnosis of MCI, versus either dementia or subjective memory complaint. Because of the time of administration criterion, certain iterations of the Addenbrooke s Cognitive Examination (ACE, ACE-R, ACE-III) which usually take around minutes to administer were not included, but the mini-addenbrooke s Cognitive Examination was. Materials and methods The datasets from several pragmatic prospective diagnostic test accuracy studies undertaken in the author s clinic were used, Progress in Neurology and Psychiatry March/April

2 Original research MCI diagnosis which examined the following short performance-based CSIs: the Mini-Mental State Examination (MMSE); 19 Mini-Mental Parkinson (MMP) 20 used as a general CSI; Montreal Cognitive Assessment (MoCA); 21 Test Your Memory () test; 22 Mini-Addenbrooke s Cognitive Examination (), 23 and the Six-item Cognitive Impairment Test (6CIT). 24 Where appropriate, data from pragmatic studies of informant-based screening instruments were also included, 25,26 specifically AD8, 27 and the short and screening versions of the Zarit Burden Interview (ZBI), 28,29 which has been examined as a possible informant-based CSI. 26 Study details (setting, sample size, prevalence of dementia and MCI in the case mix, sex ratio, age are shown in Table 1. For most of these instruments, lower scores indicate worse cognitive function (MMSE, MMP, MoCA,, ), but others are negatively scored such that lower values are better and higher are worse (6CIT, AD8, ZBI). These studies followed a standardised format used in this clinic, 30 of cross-sectional assessment of consecutive outpatient referrals with some or all of the following a. Performance-based cognitive screening instruments CSI Setting N Prevalence: Dementia, MCI F : M sex ratio (% female) Age range (years) Ref MMSE (1) CFC (dementia) 57 : 93 (38) (median 61) [14] MMSE (2) CFC (MCI) 86 : 115 (43) (median 62) [13] MMSE (3) CFC , : 81 (46) (median 60.5) [17] MMSE (4) CFC , : 64 (49) (median 64) [25] MMP CFC , : 115 (43) (median 62) [13] MoCA CFC , : 93 (38) (median 61) [14] CFC and Old Age Psychiatry memory clinic , : 130 (42) (mean 63) [15] CFC , : 71 (47) (median 60) [16] 6CIT CFC , : 124 (49) (median 59) [17] CSI Setting N Prevalence: Dementia, MCI F:M sex ratio (% female) Age range (years) Ref AD8 CFC , :106 (50) (median 64.5) [25] ZBI CFC , :31 (31) (median 65) [26] CSI, Cognitive Screening Instrument; CFC, Cognitive Function Clinic; MMSE, Mini-Mental State Examination; MMP, Mini-Mental Parkinson; MoCA, Montreal Cognitive Assessment;, Test Your Memory () test;, Mini- Addenbrooke s Cognitive Examination; 6CIT, Six-Item Cognitive Impairment Test; ZBI, Zarit Burden Interview. Table 1. Details of pragmatic studies of short cognitive screening instruments 22 Progress in Neurology and Psychiatry March/April

3 MCI diagnosis Original research elements: semi-structured patient history enquiring about cognitive symptoms and functional performance, with collateral history where possible; administration of CSIs; neuroradiological examination (CT all patients; interval MRI in some cases); and formal neuropsychological assessment in some cases. Standard diagnostic criteria for dementia (DSM-IV) and mild cognitive impairment (MCI) were used. Criterion diagnosis was by judgment of an experienced clinician based on diagnostic criteria but did not use CSI scores in order to avoid review bias. The following three forms of data analysis were performed, some based on previously published results, 11,30 others requiring re-analysis of individual study data. Firstly, significance testing based on the null hypothesis was undertaken comparing mean test scores for patients with dementia versus MCI, and for patients with MCI versus no cognitive impairment. Since all cases referred to the outpatient clinics in which these pragmatic studies were undertaken have at minimum subjective memory complaint (SMC), there was no normal control group, as befits pragmatic (or phase III ) diagnostic test accuracy studies. 11 A value of p<0.05 was taken as significant. Significance tests are, of course, post hoc, based on aggregate data once diagnoses have been made, and hence are of limited use in clinical practice where diagnosis is not known beforehand, and parameters such as sensitivity, specificity, predictive values and likelihood ratios are more typically used. 11,30 Hence the second analysis was to calculate these measures of discrimination for MCI diagnosis versus SMC, since this would most likely correspond to treatment threshold if any disease-modifying therapy were available. All these descriptors of outcome are dependent on the cut-off, cut point, threshold or dichotomisation point of the test scores selected by the investigator(s), which choice may have a major impact on these values. 18,31,32 Previously published cut-off scores for each CSI were used, where available, to minimise the risk of bias from defining optimal cut-off post hoc based on study data. 33 (NB two different cut-offs are cited for. 23 ) Single, global or unitary, indicators of test diagnostic performance have also been described, one of which is effect size, most frequently expressed as Cohen s d (the difference of the means of two groups divided by the weighted pooled standard deviations of the groups). This measure is independent of test cut-off chosen, and may be qualified (effect sizes of 0.2 to 0.3 are, 0.5 medium, and 0.8 large). 34 Hence the third analysis was calculation of effect size (Cohen s d) for the diagnosis of MCI versus SMC, some results from which have been previously reported. 35 This approach to test comparison was deemed preferable to weighted comparisons, 17,36,37 since the a. Performance-based cognitive screening instruments MMSE (1) MMSE (3) MMSE (4) MMP (0 32) MoCA (0 50) score: dementia 22.2+/ / <0.05 [14] 20.6+/ / <0.001 [17] 20.6+/ / <0.001 [25] 17.1+/ / <0.001 [13] 17.0+/ / <0.01 [14] 23.2+/ / <0.001 [15] 13.6+/ / <0.02 [16] 6CIT (28 0) 17.3+/ / <0.001 [17] score: dementia AD8 (8 0) 6.5+/ / <0.001 [25] ZBI short (48 0) ZBI screening (16 0) 16.5+/ / >0.5 [26] 6.9+/ / >0.5 [26] Table 2. Dementia versus MCI: significance testing (based on null hypothesis for test scores from pragmatic diagnostic test accuracy studies examining various short cognitive screening instruments; adapted from reference 11, p.83) Progress in Neurology and Psychiatry March/April

4 Original research MCI diagnosis latter are highly dependent on differences in test sensitivity which in turn are highly dependent on the test cut-off chosen. 11 Results As regards significance testing (see Table 2), for all the short performance-based CSIs examined (MMSE, MMP, MoCA,,, 6CIT) the null hypotheses that scores did not differ significantly between patients with dementia and patients with MCI was rejected (Table 2a). This was also true for AD8, an informant-based CSI (Table 2b), but not on either version of the ZBI, scores on which are recognised not to correlate with performance-based CSIs. 26 Likewise, examining the null hypotheses that scores did not differ a) Performance-based cognitive screening instruments score: no cognitive impairment (SMC) significantly between patients with MCI and patients with SMC (see Table 3) was rejected for all the short performance-based CSIs examined (MMSE, MMP, MoCA,,, 6CIT; Table 3a); only a trend was evident for AD8 for this differential diagnosis (Table 3b). Measures of discrimination were calculated for diagnosis of MCI versus SMC using specified test cut-offs for those short CSIs reaching or tending to statistical significance on null hypothesis testing (see Table 4). Evidently, of the performance-based CSIs examined, MMSE, MMP, MoCA, and 25/30 were highly sensitive for MCI diagnosis (Table 4a), as was the informant-based AD8 (Table 4b), with the risk of false positives; whereas 21/30 MMSE (2) 24.9+/ / <0.01 [13] MMSE (3) 24.8+/ / <0.001 [17] MMSE (4) 24.9+/ / <0.05 [25] MMP (0 32) 24.0+/ / <0.001 [13] MoCA 20.2+/ / <0.001 [14] (0 50) 37.5+/ / <0.01 [15] 17.1+/ / <0.001 [16] 6CIT (28 0) 7.6+/ / <0.001 [17] score: no cognitive impairment (SMC) AD8 (8 0) 5.2+/ / > p >0.05 ZBI short (48 0) 17.5+/ / >0.1 [26] ZBI screening (16 0) [25] 6.9+/ / >0.1 [26] Table 3. MCI versus no cognitive impairment (SMC): significance testing (based on null hypothesis for test scores from pragmatic diagnostic test accuracy studies examining various short cognitive screening instruments; adapted from reference 11, p.84) and 6CIT were more specific, with risk of false negatives. Cohen s d effect sizes for MCI diagnosis for short performancebased CSIs are shown in Table 5. Of note, the effect size was particularly impressive for two tests, MoCA and. Discussion The data reported here suggest that several short performance-based CSIs may be used to identify MCI cases with confidence, based on the findings of pragmatic diagnostic (or technically screening ) test accuracy studies. 11 Such studies should have high external validity, and the results may be transferable to other similar clinics, informing day-to-day clinical practice. Of course, this approach is not without its shortcomings. Pragmatic studies, broadly inclusive with respect to patient age and comorbidity, lack the power of proof-of-concept studies with patient groups pre-selected by diagnosis, although the latter approach is of course alien to day-to-day clinical practice. Moreover, these pragmatic studies were cross-sectional, again the idiom of clinical practice, rather than longitudinal, with consequent risk of inaccurate criterion diagnosis, particularly when attempting to separate MCI and SMC. Furthermore, the use of purely quantitative aggregate measures may not necessarily be helpful clinically (see the Materials and methods section above): since MCI patients may be anticipated to do well on certain aspects of CSIs and poorly on others (eg memory recall in amnestic MCI), so the qualitative pattern of impairment on these tests may also be useful clinically, rather than simply using the overall test scores. Nevertheless, pragmatic studies, as analysed in this report, may give indications of potential test utility as a support for MCI diagnosis, or the requirement for 24 Progress in Neurology and Psychiatry March/April

5 MCI diagnosis Original research a. Performance-based cognitive screening instruments CSI (cut-off) [reference] MMSE ( 22/30) [13] MMP ( 20/32) [13] MoCA ( 26/30) [14] ( 42/50) [15] ( 25/30) [16] ( 21/30) [16] 6CIT ( 9/28) [17,18] Sensitivity ( ) 0.93 ( ) 0.79 ( ) ( ) 0.66 ( ) Specificity 0.28 ( ) 0.61 ( ) 0.60 ( ) 0.54 ( ) 0.43 ( ) 0.82 ( ) 0.70 ( ) Positive Predictive Value 0.23 ( ) 0.34 ( ) 0.44 ( ) 0.39 ( ) 0.49 ( ) 0.70 ( ) 0.53 ( ) Negative Predictive Value ( ) 0.96 ( ) 0.88 ( ) ( ) 0.80 ( ) Positive Likelihood Ratio 1.39 ( ) unimportant 2.37 ( ) 2.33 ( ) 1.74 ( ) unimportant 1.76 ( ) unimportant 4.26 ( ) 2.19 ( ) Negative Likelihood Ratio 0 large 0.14 ( ) moderate further patient investigation or longitudinal assessment. It was not the purpose of this analysis directly to compare the different CSIs, since the different studies from which the data were taken were not (and never could be) exactly comparable. Nevertheless, the studies being examined shared a similar setting, methodology (eg application of reference standard for diagnosis) and analysis, thereby minimising variability, although this could not be entirely excluded (eg dementia prevalence, based on the case mix seen in series 0.11 ( ) moderate CSI (cut-off) [reference] AD8 ( 2/8) [25] Sensitivity 0.98 ( ) Specificity 0.17 ( ) Positive predictive value 0.47 ( ) Negative predictive value 0.88 ( ) Positive likelihood ratio Negative likelihood ratio 1.17 ( ) unimportant 0.19 ( ) moderate 0.38 ( ) Table 4. Measures of discrimination (with 95% confidence intervals) for diagnosis of MCI versus SMC at specified cut-offs using those short cognitive screening instruments reaching or tending to statistical significance on null hypothesis testing (see Table 3) in pragmatic studies of unselected consecutive patients). Nevertheless, the suggestion that emerges is that MMSE, MMP, MoCA, and (cut off 25/30) may be acceptable as screeners for MCI in this clinical setting if high sensitivity is desired, as is often the case for clinical practice. However, the high false positive rate associated with high sensitivity would argue against the use of these instruments for use in clinical trials where high specificity and low false positive rate is more desirable. 1 With, in addition, their large effect size (Cohen s d), MoCA and 0 large 0.28 ( ) may be of particular value for MCI diagnosis. Since MoCA was designed specifically to identify MCI cases, 21 this finding is perhaps not unexpected. There are, of course, other CSIs that might be used for MCI diagnosis, which were not encompassed in this study, such as the Quick mild cognitive impairment (Qmci) screen, 38 DemTect, 39,40 MCI Screen, 41 ABCS 135, 42 and even the Clock Drawing Test. 43 An important issue for all these possible screening instruments for MCI will be to examine how well they correlate with biomarkers of dementia disorders, such as AD. 44 Another avenue for future study would be combining performancebased and informant-based CSIs. This has been attempted for diagnosis of dementia (MMSE and AD8; 6CIT and AD8), 25 with improved sensitivity for MMSE and AD8 in the parallel ( Or rule) paradigm. In conclusion, a number of short performance-based CSIs may be helpful in clinical practice as a first step to screen for the diagnosis of MCI. Dependent upon their 0.49 ( ) Progress in Neurology and Psychiatry March/April

6 Original research MCI diagnosis CSI Cohen s d Effect size MMSE MMP MoCA 6CIT AD Large Large Large Table 5. MCI versus no cognitive impairment (SMC): Cohen s d effect size (adapted from references 11,35) results, more detailed assessment including sophisticated biomarker studies and formal cognitive assessment may be indicated to establish or refute the diagnosis of MCI. Dr Larner is Consultant Neurologist at the Cognitive Function Clinic, Walton Centre for Neurology and Neurosurgery, Liverpool. Conflict of interest The author declares no conflict of interest. References 1. Dubois B, Feldman HH, Jacova C, et al. Advancing research diagnostic criteria for Alzheimer s disease: the IWG-2 criteria. Lancet Neurol 2014;13: [Erratum Lancet Neurol 2014;13:757]. 2. Jack CR Jr, Knopman DS, Jagust WJ, et al. Tracking pathophysiological processes in Alzheimer s disease: an updated hypothetical model of dynamic biomarkers. Lancet Neurol 2013;12: Yau WY, Tudorascu DL, McDade EM, et al. Longitudinal assessment of neuroimaging and clinical markers in autosomal dominant Alzheimer s disease: a prospective cohort study. Lancet Neurol 2015;14: Litvan I, Goldman JG, Troster AI, et al. Diagnostic criteria for mild cognitive impairment in Parkinson s disease: Movement Disorder Society Task Force guidelines. Mov Disord 2012; 27: Gorelick PB, Scuteri A, Black SE, et al. Vascular contributions to cognitive impairment and dementia: a statement for healthcare professionals from the American Heart Association / American Stroke Association. Stroke 2011;42: De Mendonca A, Ribeiro F, Guerreiro M, et al. Frontotemporal mild cognitive impairment. J Alzheimers Dis 2004;6: Mitchell AJ, Shiri-Feshki M. Rate of progression of mild cognitive impairment to dementia meta-analysis of 41 robust inception cohort studies. Acta Psychiatr Scand 2009;119: Graham JE, Rockwood K, Beattie BL, et al. Prevalence and severity of cognitive impairment with and without dementia in an elderly population. Lancet 1997;349: Kozora E, Filley CM. Cognitive dysfunction and white matter abnormalities in systemic lupus erythematosus. J Int Neuropsychol Soc 2011;17: Albert MS, DeKosky ST, Dickson D, et al. The diagnosis of mild cognitive impairment due to Alzheimer s disease: recommendations from the National Institute on Aging-Alzheimer s Association workgroups on diagnostic guidelines for Alzheimer s disease. Alzheimers Dement 2011;7: Larner AJ. Diagnostic test accuracy studies in dementia. A pragmatic approach. London: Springer, Mitchell AJ. The Mini-Mental State Examination (MMSE): an update on its diagnostic validity for cognitive disorders. In: Cognitive screening instruments. A practical approach. Larner AJ, ed. London: Springer, 2013; Larner AJ. Mini-Mental Parkinson (MMP) as a dementia screening test: comparison with the Mini-Mental State Examination (MMSE). Curr Aging Sci 2012;5: Larner AJ. Screening utility of the Montreal Cognitive Assessment (MoCA): in place of or as well as the MMSE? Int Psychogeriatr 2012;24: Hancock P, Larner AJ. Test Your Memory test: diagnostic utility in a memory clinic population. Int J Geriatr Psychiatry 2011;26: Larner AJ. Mini-Addenbrooke s Cognitive Examination: a pragmatic diagnostic accuracy study. Int J Geriatr Psychiatry 2015;30: Abdel-Aziz K, Larner AJ. Six-item cognitive impairment test (6CIT): pragmatic diagnostic accuracy study for dementia and MCI. Int Psychogeriatr 2015;27: Larner AJ. Implications of changing the Sixitem Cognitive Impairment Test cutoff. Int J Geriatr Psychiatry 2015;30: Folstein MF, Folstein SE, McHugh PR. Mini- Mental State. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975;12: Mahieux F, Michelet D, Manifacier MJ, et al. Mini-Mental Parkinson: first validation study of a new bedside test constructed for Parkinson s disease. Behav Neurol 1995;8: Nasreddine ZS, Phillips NA, Bédirian V, et al. The Montreal Cognitive Assessment, MoCA: a brief screening tool for mild cognitive impairment. J Am Geriatr Soc 2005;53: Brown J, Pengas G, Dawson K, et al. Self administered cognitive screening test () for detection of Alzheimer s disease: cross sectional study. BMJ 2009;338:b Hsieh S, McGrory S, Leslie F, et al. The Mini- Addenbrooke s Cognitive Examination: a new assessment tool for dementia. Dement Geriatr Cogn Disord 2015;39: Brooke P, Bullock R. Validation of a 6 item cognitive impairment test with a view to primary care usage. Int J Geriatr Psychiatry 1999; 14: Larner AJ. AD8 informant questionnaire for cognitive impairment: pragmatic diagnostic test accuracy study. J Geriatr Psychiatry Neurol 2015;28: Stagg B, Larner AJ. Zarit Burden Interview: pragmatic study in a dedicated Cognitive Function Clinic. Prog Neurol Psychiatry 2015; 19(4): Galvin JE, Roe CM, Powlishta KK, et al. The AD8. A brief informant interview to detect dementia. Neurology 2005;65: Zarit SH, Reever KE, Bach-Peterson J. Relatives of the impaired elderly: correlates of feelings of burden. Gerontologist 1980;20: Bédard M, Molloy DW, Squire L, et al. The Zarit Burden Interview: a new short version and screening version. Gerontologist 2001;41: Larner AJ. Dementia in clinical practice: a neurological perspective. Pragmatic studies in the cognitive function clinic, 2 nd ed. London: Springer, Larner AJ. Optimizing the cutoffs of cognitive screening instruments in pragmatic diagnostic accuracy studies: maximising accuracy or Youden index? Dement Geriatr Cogn Disord 2015;39: Larner AJ. The Q* index: a useful global measure of dementia screening test accu racy? Dement Geriatr Cogn Disord Extra 2015;5: Davis DH, Creavin ST, Noel-Storr A, et al. Neuropsychological tests for the diagnosis of Alzheimer s disease dementia and other de m en tias: a generic protocol for cross-sectional and delayed-verification studies. Cochrane Database Syst Rev 2013;3:CD Cohen J. Statistical power analysis for the behavioural sciences, 2 nd ed. Hillsdale, New Jersey, USA: Lawrence Erlbaum, Larner AJ. Effect size (Cohen s d) of cognitive screening instruments examined in pragmatic diagnostic accuracy studies. Dement Geriatr Cogn Disord Extra 2014;4: Larner AJ. Comparing diagnostic accuracy of cognitive screening instruments: a weighted comparison approach. Dement Geriatr Cogn Disord Extra 2013;3: Larner AJ, Hancock P. ACE-R or MMSE? A weighted comparison. Int J Geriatr Psychiatry 2014;29: O Caoimh R, Gao Y, McGlade C, et al. Comparison of the quick mild cognitive impairment (Qmci) screen and the SMMSE in screening for mild cognitive impairment. Age Ageing 2012;41: Kalbe E, Kessler J, Calabrese P, et al. Dem- Tect: a new, sensitive cognitive screening test to support the diagnosis of mild cognitive impairment and early dementia. Int J Geriatr Psychiatry 2004;19: Larner AJ. DemTect: 1-year experience of a neuropsychological screening test for dementia. Age Ageing 2007;36: Shankle WR, Romney AK, Hara J, et al. Methods to improve the detection of mild cognitive impairment. Proc Natl Acad Sci USA 2005;102: Standish TI, Molloy DW, Cunje A, et al. Do the ABCS 135 short cognitive screen and its subtests discriminate between normal cognition, mild cognitive impairment and dementia? Int J Geriatr Psychiatry 2007;22: Umidi S, Trimarchi PD, Corsi M, et al. Clock drawing test (CDT) in the screening of mild cognitive impairment (MCI). Arch Gerontol Geriatr 2009;49(Suppl1): Galvin JE, Fagan AM, Holtzman DM, et al. Relationship of dementia screening tests with biomarkers of Alzheimer s disease. Brain 2010;133: Progress in Neurology and Psychiatry March/April