Underestimation of cognitive impairments by the Montreal Cognitive Assessment (MOCA) in an acute stroke unit population Edgar Chan, Sabah Khan, Rupert Oliver, Sumanjit K. Gill, David J. Werring, Lisa Cipolotti PII: DOI: Reference:
S0022-510X(14)00295-0 doi: 10.1016/j.jns.2014.05.005 JNS 13182
To appear in:
Journal of the Neurological Sciences
Received date: Revised date: Accepted date:
11 December 2013 29 April 2014 8 May 2014
Please cite this article as: Chan Edgar, Khan Sabah, Oliver Rupert, Gill Sumanjit K., Werring David J., Cipolotti Lisa, Underestimation of cognitive impairments by the Montreal Cognitive Assessment (MOCA) in an acute stroke unit population, Journal of the Neurological Sciences (2014), doi: 10.1016/j.jns.2014.05.005
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT UNDERESTIMATION OF COGNITIVE IMPAIRMENTS BY THE MONTREAL COGNITIVE ASSESSMENT (MOCA) IN AN ACUTE STROKE UNIT
IP
T
POPULATION
SC R
Edgar Chan1, PhD; Sabah Khan1, DPsych; Rupert Oliver1, MD; Sumanjit K Gill2, MD; David J Werring3, FRCP; Lisa Cipolotti1, PhD. 1
National Hospital for Neurology and Neurosurgery, London, UK, 2Watford General
MA
NU
Hospital, Watford, UK, 3UCL Institute of Neurology, Stroke Research Group, London, UK.
Correspondence to:
D
Dr Edgar Chan
Queen Square
CE P
London WC1N 3BG.
TE
Box 37, NHNN
Email [email protected].
AC
Tel +442034484793. Fax +442034484761
Cover title: Cognitive impairments in acute stroke Contents: Manuscript, Table 1 and 2, Figure 1 Keywords: stroke; MoCA; cognition; neuropsychology; cognitive assessment; vascular disease Word count: 2310
1
ACCEPTED MANUSCRIPT
Abstract
T
Background and Purpose - The Montreal Cognitive Assessment (MoCA) is an increasingly
IP
popular clinical screening tool for detecting cognitive impairment in stroke, but few studies
SC R
have directly compared performance on the MoCA with neuropsychological assessment. Our retrospective study examined the extent to which intact performance on the MoCA reflects intact cognition as determined by neuropsychological assessment.
NU
Methods –In this retrospective study, cognitive profiles were examined for 136 acute stroke
MA
patients admitted to the Acute Stroke Unit who had available MoCA and neuropsychological assessment data.
Results –22% of our patients were deemed cognitively intact on the MoCA. Of these, 78%
TE
D
were found to be impaired (≤5%ile) on neuropsychological assessment in one or more cognitive domains. The most common impairments were in general intelligence, information
CE P
processing speed and visual memory; three areas not assessed by the MoCA. In addition, a high proportion (up to 59%) of patients who scored the maximum points on one of the
AC
MoCA-specified domains were impaired on comparable neuropsychological assessment. Conclusions – Our findings suggest that although the MOCA may be a useful screening tool post-stroke in detecting gross impairments, neuropsychological assessment is still necessary for a comprehensive and reliable detection of domain-specific cognitive deficits, which can more reliably inform us for realistic goal setting and vocational advice vital for effective rehabilitation.
2
ACCEPTED MANUSCRIPT
T
Introduction
IP
Cognitive impairment is a common sequelae following stroke. It can have substantial
SC R
long-term functional consequences and is a significant burden on the health care system [1,2]. Early detection of domain-specific cognitive impairment post-stroke is invaluable for prognostic and rehabilitation purposes. Domain-specific cognitive impairments, in particular
NU
executive functioning and memory, have been shown to be good predictors for factors such
MA
as length of hospital stay, long-term impairment, and burden on community services [1-4]. The Montreal Cognitive Assessment (MoCA) is an increasingly popular cognitive
D
screening tool that has been shown to have better sensitivity in detecting post-stroke
TE
impairments compared with the Mini-Mental Status Examination (MMSE) [5-8]. In the UK, national guidelines [9,10] recommend that all patients should be assessed for cognitive
CE P
impairment within 6 weeks post-stroke using a validated tool, particularly highlighting the MoCA [11]. It is thought to be more cognitively demanding than the MMSE, containing
AC
more items assessing important domains such as executive functions and attention. However, like most screening tools, the MoCA was originally designed to measure cognitive deterioration in dementia using an overall cut-off score. As such, domain-specific impairments are often not considered. Moreover, cognitive domains which are often impaired after stroke, such as intellectual functioning, speed of information processing and non-verbal memory, are not assessed by the MoCA. Only few studies with stroke patients have compared the sensitivity and specificity of the MoCA with the “gold standard” of neuropsychological assessment, with variable results [12-14]. Some studies reported the MoCA had good sensitivity but only moderate specificity, possibly reflecting the fact that the MoCA total-score measures performance across multiple
3
ACCEPTED MANUSCRIPT domains with limited sensitivity for single-domain impairment [12]. One study found comparable sensitivity and specificity, but again that it was better for detecting amnestic
T
impairments than single-domain non-amnestic impairments [13]. Until now however, the
IP
domain-specific pattern of sensitivity for the MoCA has not been fully characterised. The aim
SC R
of our retrospective study was to compare intact performance on the MoCA, in regards to overall and domain-specific scores, with the “gold standard” of detailed neuropsychological
NU
assessment in a cohort of patients with acute stroke.
MA
Method
Three-hundred and sixty-four acute stroke patients admitted between January 2011 to
D
July 2013 to the Acute Stroke Unit of the National Hospital for Neurology and Neurosurgery
TE
(NHNN), London and assessed by the Neuropsychology Department, were included in a retrospective audit of available data. This unit accepts local stroke patients within 24 to 72
CE P
hours after admission to the University College Hospital Hyperacute Stroke Unit. All patients in the unit were assessed by a Clinical Neuropsychologist (E.C. or another member of the
AC
neuropsychology department, NHNN) as part of standard routine care. Inclusion criteria of the retrospective study were the availability of MoCA and neuropsychological data, and assessments had to be conducted within 3 months of the stroke event. Of those patients who had both MoCA and neuropsychological data available, six patients were excluded for having a history of psychiatric disorders and two patients were excluded for having a history of alcohol and drug abuse as determined from the hospital admissions/discharge report. The study was approved by the local clinical governance and ethics committees. One-hundred and thirty-six acute stroke patients were included. All patients were administered the MoCA followed by a neuropsychological assessment by the same clinician. Due to factors such as educational/occupational/cultural background/ability to cope with test
4
ACCEPTED MANUSCRIPT demands, the same tests and/or same cognitive domains were not administered/assessed in all patients. As such, 27 patients were administered only sub-sections of the MoCA. The
T
neuropsychological assessment comprised a standardized battery evaluating seven cognitive
IP
domains. However, as this was a retrospective study, patients were administered a subset of
SC R
tests deemed suitable by the Clinical Neuropsychologist at the time. Current general intellectual functioning was assessed using the Wechsler Adult Intelligence Scale—Third Edition (WAIS-III) [15]. Premorbid intellectual functioning was assessed using the National
NU
Adult Reading Test (NART) [16]. Verbal and visual memory functions were assessed with
MA
either the Recognition Memory Test [17]. or the Doors and People test [18] and naming skills were examined either with the Graded Naming Test [19] or the Oldfield Naming Test [20].
D
Perceptual functions were assessed using the Visual Object and Space Perception Battery
TE
[21]. Information processing speed was examined using one or more of the following tests: ‘O’ Cancellation, Digit Copy [22], Symbol Digit Modalities Test [23] or the Trail Making
CE P
Test (Part A) [24]. Executive functions were examined using one or more of the following tests: Stroop Test [25]; Trail Making Test Part B [23], Weigl Colour Form Sorting Task [26]
AC
or the Hayling and Brixton Test [27]. Testing lasted approximately 60 minutes in total and was generally conducted in one session unless if patients were too fatigued. Patients were classed as cognitively intact on the MOCA if they scored ≥ 25 out of 30 [12,13]. Derived scores on neuropsychological assessment were calculated based on published standardized normative data; performance in each cognitive domain was classed as impaired/unimpaired according to predefined criteria [28]. In general, scores at or below the fifth percentile on any one test were taken to indicate impairment in that respective domain. For intellectual functioning, a difference equal to or greater than 10 between performance on the WAIS-III Verbal or Performance IQ measures and the estimated premorbid functioning on the NART was taken as evidence of intellectual decline.
5
ACCEPTED MANUSCRIPT We conducted two main analyses: we first identified the patients classed as cognitively intact on the MoCA (i.e. MoCA-intact population) and investigated their
T
performance on detailed neuropsychological assessment; we then identified patients who
IP
scored the maximum points on each of the MoCA-specified cognitive domains, irrespective
SC R
of the overall score, and investigated their performance on neuropsychological assessments in the relevant domain and calculated the relevant negative predictive value (NPV).
NU
Results
MA
Thirty patients (22%) had MoCA-intact overall scores. Their demographic and clinical characteristics are shown in Table 1. There was no difference between the MoCA-
D
intact and MoCA-impaired group on relevant demographic variables. In contrast, the MoCA-
TE
intact group had significantly higher estimated premorbid intellectual functioning and higher
Insert Table 1 here.
AC
CE P
current verbal and non-verbal intelligence.
Twenty-three (77%) of the MoCA-intact population were impaired on at least one domain on detailed neuropsychological assessment. Of the 23 patients, sixteen patients (70%) were impaired in two or more cognitive domains while the remaining seven patients (30%) were impaired in one cognitive domain only. The distribution of neuropsychological impairment across cognitive domains in the MoCA-intact population is shown in Figure 1.
Insert Figure 1 here.
6
ACCEPTED MANUSCRIPT Of the MoCA-intact population, a high proportion of patients who were assessed on neuropsychological assessments had impairment in intellectual functioning (12 of 17
T
patients), speed of information processing (14 of 25 patients) and non-verbal memory (9 of
IP
25 patients); domains not assessed by the MoCA. Notably however, a high proportion of
SC R
patients also showed impairments in executive functioning (9 of 29 patients) and visuoperceptual/spatial domains (7 of 24 patients), which are assessed by the MOCA. Naming and verbal memory impairments were detected less frequently (3 of 30 and 3 of 28 patients
MA
NU
respectively).
For patients who scored the maximum points within MoCA-specified cognitive
D
domains irrespective of the overall score, their performance on neuropsychological
Insert Table 2 here.
AC
CE P
TE
assessment for each domain, and the NPV, are summarized in Table 2.
More than 50% of the patients were impaired on attention despite scoring the maximum points on this domain on the MoCA. A relatively high proportion of patients also showed impairments in memory (35%), visuospatial/executive function (30%), and naming (21%).
For completeness, we also examined the neuropsychological performance of the MoCA-impaired group and found that 103 out of the 106 patients (97%) were impaired on one or more cognitive domains. Examination of the three remaining patients revealed that they also showed cognitive deficits on neuropsychological assessment. Although they did not
7
ACCEPTED MANUSCRIPT score below the 5th percentile cut-off to be classed as “impaired”, all three patients were found to have mild to moderate executive dysfunction. Therefore, for our retrospective data
T
using a cut-off score of ≥ 25, the MoCA showed good sensitivity (0.82) but moderate
IP
specificity (0.7), very good positive predictive value (0.97) but very poor negative predictive
SC R
value (0.23).
NU
Discussion
MA
In our retrospective study, a striking 77% of patients classed as cognitively intact on the MoCA were found to be impaired across one or more cognitive domains on
D
neuropsychological assessment. The most common impairments were in intellectual
TE
functioning, speed of information processing, and visual memory; three domains not assessed by the MoCA. Impairments in these domains however, are frequently detected post-stroke
CE P
[1,3,4] and have high predictive value for functional outcome [29,30]. Although the MoCA may be useful in detecting gross cognitive impairments with a sensitivity of 0.82, it may have
AC
limitations in fully assessing relevant post-stroke impairments with a negative predictive value of 0.23.
Contrary to suggestions that poor specificity of the MoCA is related to poor detection of single-domain impairments [13,14], most of our patients were impaired on two or more cognitive domains. This may reflect the timing of assessments: whereas previous studies assessed patients at ≥1 year post-stroke [13,14], ours were assessed between 48h and 3 months post-stroke, when multiple-domain impairment may be more common [4]. We also found that intact performance within a MoCA-specified domain does not exclude impairment in that domain. More than half of our patients who scored the maximum points within the attention domain and one third of patients in the memory and
8
ACCEPTED MANUSCRIPT visuospatial/executive
domains
of
the
MoCA
were
impaired
on
comparable
neuropsychological assessment. Our findings corroborate with a recent study which found
T
poor correlation between performance on MoCA subtests and their neuropsychological
IP
counterpart [14]. The poor correspondence between performance on MoCA-specified
SC R
domains and comparable neuropsychological assessment is most likely due to the lack of breadth and depth of screening items, and the absence of normative data to take into account factors such as age and pre-morbid intellectual functioning. In addition, another contributing
NU
factor may be that the MoCA-specified domains do not accurately reflect the cognitive
MA
processes required to complete the relevant tasks. For example, the “attention” domain contains a digit-span backwards task and a subtraction task which are commonly regarded as
D
a test of working memory [15].
TE
Despite common apprehension regarding the feasibility of neuropsychological assessment during the acute phase post-stroke, our current study demonstrated that
CE P
neuropsychological assessment is possible and can be informative over and above a basic cognitive screen. A prospective study by van Zandwoort and colleagues [4] found that 74%
AC
of patients consecutively recruited on an acute stroke ward were able to complete a one and a half hour cognitive assessment battery 4-20 days post-stroke, with 77% of patients able to complete 82% of the assessment battery. Importantly, on follow-up, they found that testprofiles of patients was stable over time despite overall improvement, and early measure of intellectual functioning and other cognitive domains had significant predictive validity with respect to functional outcome. Our findings further corroborate that neuropsychological assessment is feasible; it does not necessarily need to be exhaustive but should nevertheless be comprehensive. As this was a retrospective study examining data that was collected as a part of routine clinical care, there are some limitations that need to be considered. First, the
9
ACCEPTED MANUSCRIPT proportion of patients that were included in the study was relatively small (37%). This is much lower than previous feasibility studies looking at completion rate of the MoCA and
T
neuropsychological assessment in stroke populations [4,8] and most likely reflects the post-
IP
hoc nature of the current study design rather than an anomaly with the patient sample.
SC R
Second, as there was no a priori neuropsychological battery, patients did not all complete the same set of tests and/or in the same order and it is difficult to determine retrospectively the reasons for missing data which may be clinically relevant. In addition, data regarding mood at
NU
time of assessment was not available for consideration. Finally, the clinician was not blind to
MA
the results of the MoCA prior to administering neuropsychological assessment which may have biased the selection of neuropsychological tests used.
D
In conclusion, our findings suggest that the MoCA does not fully assess cognitive
TE
domains that are commonly affected following stroke. Furthermore, intact performance on MoCA-specified domains does not preclude the presence of domain-specific cognitive
CE P
impairments. Our research advocates for the use of neuropsychological assessment tools for detecting cognitive impairment post-stroke. However, current resource limitations in local
AC
health services may not always allow for the provision of adequate neuropsychological services. Further research using a more rigid prospective design is needed to determine the potential clinical and functional importance and benefits of neuropsychological impairments not detected by the MoCA to help inform the development of more sensitive screening instruments for acute stroke.
Disclosures: None
10
ACCEPTED MANUSCRIPT References 1. Barker-Collo S & Feigin V. The impact of neuropsychological deficits on functional
IP
T
stroke outcomes. Neuropsychol Rev. 2006; 16: 53-64.
SC R
2. Galski, T., Bruno, R.L., Zorowitz, R. & Walker J. Predicting length of stay, functional outcome, and aftercare in the rehabilitation of stroke patients. The dominant role of higher-
NU
order cognition. Stroke. 1993; 24: 1794-1800.
MA
3. Tatemichi TK, Desmond DW, Stern Y, Paik M, Sano M & Bagiella E. Cognitive impairment after stroke: frequency, patterns, and relationship to functional abilities. Jour of
TE
D
Neur, Neurosurg, and Psych. 1994; 57: 202-207.
4. Van Zandvoort MJE, Kessels PC, Nys GMS, de Haan EHF & Kappelle LJ. Early
CE P
neuropsychological evaluation in patients with ischaemic stroke provides valid information.
AC
Clinical Neurology and Neurosurgery. 2005; 107: 385-392.
5. Blackburn DJ, Bafadhel L, Randall M & Harkness KA. Cognitive screening in the acute stroke setting. Age Ageing. 2013; 42:113-116.
6. Dong Y, Sharma VK, Chan BP, Venketasubramanian N, Teoh HL, Seet RC et al. The Montreal Cognitive Assessment (MoCA) is superior to the Mini-Mental State Examination (MMSE) for the detection of vascular cognitive impairment after acute stroke. J Neurol Sci. 2010; 299:15-18.
11
ACCEPTED MANUSCRIPT 7. Pendlebury ST, Cuthbertson FC, Welch SJV, Mehta Z & Rothwell PM. Underestimation of cognitive impairment by mini-mental state examination versus the montreal cognitive
T
assessment in patients with transient ischemic attack and stroke: A population-based study.
SC R
IP
Stroke. 2010; 41: 1290-1293.
8. Cumming TB, Bernhardt J & Linden T. The montreal cognitive assessment: Short
NU
cognitive evaluation in a large stroke trial. Stroke. 2011; 42: 2642-2644.
MA
9. Intercollegiate Stroke Working Party. National clinical guideline for stroke, 4th edition.
D
London: Royal College of Physicians, 2012.
TE
10. National Institute for Health and Clinical Excellence (2013). Stroke rehabilitation: Long
CE P
term rehabilitation after stroke. London: NICE (www.guidance.nice.org.uk/CG162).
11. Nasreddine ZS, Phillips NA, Bédirian V, Charbonneau S, Whitehead V, Collin I, et al.
AC
The montreal cognitive assessment, MoCA: A brief screening tool for mild cognitive impairment. J Am Geriatr Soc. 2005; 53: 695-699.
12. Godefroy O, Fickl A, Roussel M, Auribault C, Bugnicourt JM, Lamy C, et al. Is the Montreal Cognitive Assessment superior to the Mini-Mental State Examination to detect poststroke cognitive impairment? A study with neuropsychological evaluation. Stroke. 2011; 42:1712-6.
13. Pendlebury ST, Mariz J, Bull L, Mehta Z & Rothwell PM. MoCA, ACE-R, and MMSE versus the National Institute of Neurological Disorders and Stroke-Canadian Stroke Network
12
ACCEPTED MANUSCRIPT Vascular Cognitive Impairment Harmonization Standards Neuropsychological Battery after
T
TIA and stroke. Stroke. 2012;43:464-9.
IP
14. Schweizer TA, Al-Khindi T & Macdonald RL. Mini-Mental State Examination versus
SC R
Montreal Cognitive Assessment: rapid assessment tools for cognitive and functional outcome after aneurysmal subarachnoid hemorrhage. J Neurol Sci. 2012;316:137-40.
NU
15. Wechsler, D. Wechsler Adult Intelligence Scale – 3rd Edition. San Antonio, TX: Harcourt
MA
Assessment; 1997.
D
16. Nelson, H.E. The National Adult Reading Test (NART): for the assessment of premorbid
TE
intelligence in patients with dementia. End ed. Windsor (UK): NFER-Nelson; 1991.
CE P
17. Warrington, E.K. Recognition Memory Test. Windsor (UK): NFER-Nelson, 1984.
AC
18. Baddeley, A.D., Emslie, H. & Nimmo-Smith, I. Doors and people: a test opf visual and verbal recall and recognition. Bury St. Edmunds, England: Thames Valley Test Co, 1994.
19. McKenna, P. & Warrington, E.K. Graded Naming Test. Windsor (UK): NFER-Nelson; 1983.
20. Oldfield, R.C. & Wingfield, A., Response latencies in naming objects. Q J Exp Psychol. 1965; 17: 273-281.
13
ACCEPTED MANUSCRIPT 21. Warrington, E.K. & James, M. Visual Object and Space Perception Battery. Bury St.
T
Edmunds (UK): Thames Valley Test Company; 1991.
SC R
psychomotor speed. Behav Neurol. 1992; 5: 113-116.
IP
22. Willison, J.R. & Warrington, E.K. Cognitive retardation in a patient with preservation of
MA
Western Psychological Services; 1982.
NU
23. Smith, A. Symbol Digit Modalities Test (SDMT) Manual (Revised). Los Angeles:
24. Reitan, R.M. (1958). Validity of the trail making test as an indicator of organic brain
TE
D
damage. Perceptual and Motor Skills. 1958; 8: 271-276.
25. Trennary, M.R. Crossen, B., DeBoe, J., & Leber, W.R. Stroop Neuropsychological
CE P
Screening Test (SNST). Odessa (FL): Psychological Assessment Resources: 1989.
AC
26. Weigl, E. On the psychology of so-called processes of abstraction. J Abnormal Soc Psychol. 1941; 36: 3-33.
27. Burgess, P.W. & Shallice, T. The Hayling and Brixton Tests. Technical report. Bury St. Edmunds (UK): Thames Valley Test Company; 1997.
28. Werring DJ, Frazer DW, Coward LJ, Losseff NA, Watt H. Cipolotti L et al. Cognitive dysfunction in patients with cerebral microbleeds on T2*-weighted gradient-echo MRI. Brain. 2004; 127: 2265-2275.
14
ACCEPTED MANUSCRIPT 29. Nys GMS, van Zandvoort MJE, de Kort PLM, van der Worp HB, Jansen BPW, Algra A et al.. The prognostic value of domain-specific cognitive abilities in acute first-ever stroke.
IP
T
Neurology. 2005; 64: 821-827.
SC R
30. Hurford, R and Charidimou, A and Fox, Z and Cipolotti, L and Werring, DJ. Domainspecific trends in cognitive impairment after acute ischaemic stroke. J Neurol. 2013; 260: 237
AC
CE P
TE
D
MA
NU
- 241.
15
ACCEPTED MANUSCRIPT Figures
AC
CE P
TE
D
MA
NU
SC R
IP
T
Figure 1. MOCA-intact population: Neuropsychological performance across domains.
16
ACCEPTED MANUSCRIPT Table 1 Demographic and clinical characteristics.
(n=136)
Intact
(n=106)
Impaired
16.0 (6.1)
p
S0022-510X(14)00295-0 doi: 10.1016/j.jns.2014.05.005 JNS 13182
To appear in:
Journal of the Neurological Sciences
Received date: Revised date: Accepted date:
11 December 2013 29 April 2014 8 May 2014
Please cite this article as: Chan Edgar, Khan Sabah, Oliver Rupert, Gill Sumanjit K., Werring David J., Cipolotti Lisa, Underestimation of cognitive impairments by the Montreal Cognitive Assessment (MOCA) in an acute stroke unit population, Journal of the Neurological Sciences (2014), doi: 10.1016/j.jns.2014.05.005
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT UNDERESTIMATION OF COGNITIVE IMPAIRMENTS BY THE MONTREAL COGNITIVE ASSESSMENT (MOCA) IN AN ACUTE STROKE UNIT
IP
T
POPULATION
SC R
Edgar Chan1, PhD; Sabah Khan1, DPsych; Rupert Oliver1, MD; Sumanjit K Gill2, MD; David J Werring3, FRCP; Lisa Cipolotti1, PhD. 1
National Hospital for Neurology and Neurosurgery, London, UK, 2Watford General
MA
NU
Hospital, Watford, UK, 3UCL Institute of Neurology, Stroke Research Group, London, UK.
Correspondence to:
D
Dr Edgar Chan
Queen Square
CE P
London WC1N 3BG.
TE
Box 37, NHNN
Email [email protected].
AC
Tel +442034484793. Fax +442034484761
Cover title: Cognitive impairments in acute stroke Contents: Manuscript, Table 1 and 2, Figure 1 Keywords: stroke; MoCA; cognition; neuropsychology; cognitive assessment; vascular disease Word count: 2310
1
ACCEPTED MANUSCRIPT
Abstract
T
Background and Purpose - The Montreal Cognitive Assessment (MoCA) is an increasingly
IP
popular clinical screening tool for detecting cognitive impairment in stroke, but few studies
SC R
have directly compared performance on the MoCA with neuropsychological assessment. Our retrospective study examined the extent to which intact performance on the MoCA reflects intact cognition as determined by neuropsychological assessment.
NU
Methods –In this retrospective study, cognitive profiles were examined for 136 acute stroke
MA
patients admitted to the Acute Stroke Unit who had available MoCA and neuropsychological assessment data.
Results –22% of our patients were deemed cognitively intact on the MoCA. Of these, 78%
TE
D
were found to be impaired (≤5%ile) on neuropsychological assessment in one or more cognitive domains. The most common impairments were in general intelligence, information
CE P
processing speed and visual memory; three areas not assessed by the MoCA. In addition, a high proportion (up to 59%) of patients who scored the maximum points on one of the
AC
MoCA-specified domains were impaired on comparable neuropsychological assessment. Conclusions – Our findings suggest that although the MOCA may be a useful screening tool post-stroke in detecting gross impairments, neuropsychological assessment is still necessary for a comprehensive and reliable detection of domain-specific cognitive deficits, which can more reliably inform us for realistic goal setting and vocational advice vital for effective rehabilitation.
2
ACCEPTED MANUSCRIPT
T
Introduction
IP
Cognitive impairment is a common sequelae following stroke. It can have substantial
SC R
long-term functional consequences and is a significant burden on the health care system [1,2]. Early detection of domain-specific cognitive impairment post-stroke is invaluable for prognostic and rehabilitation purposes. Domain-specific cognitive impairments, in particular
NU
executive functioning and memory, have been shown to be good predictors for factors such
MA
as length of hospital stay, long-term impairment, and burden on community services [1-4]. The Montreal Cognitive Assessment (MoCA) is an increasingly popular cognitive
D
screening tool that has been shown to have better sensitivity in detecting post-stroke
TE
impairments compared with the Mini-Mental Status Examination (MMSE) [5-8]. In the UK, national guidelines [9,10] recommend that all patients should be assessed for cognitive
CE P
impairment within 6 weeks post-stroke using a validated tool, particularly highlighting the MoCA [11]. It is thought to be more cognitively demanding than the MMSE, containing
AC
more items assessing important domains such as executive functions and attention. However, like most screening tools, the MoCA was originally designed to measure cognitive deterioration in dementia using an overall cut-off score. As such, domain-specific impairments are often not considered. Moreover, cognitive domains which are often impaired after stroke, such as intellectual functioning, speed of information processing and non-verbal memory, are not assessed by the MoCA. Only few studies with stroke patients have compared the sensitivity and specificity of the MoCA with the “gold standard” of neuropsychological assessment, with variable results [12-14]. Some studies reported the MoCA had good sensitivity but only moderate specificity, possibly reflecting the fact that the MoCA total-score measures performance across multiple
3
ACCEPTED MANUSCRIPT domains with limited sensitivity for single-domain impairment [12]. One study found comparable sensitivity and specificity, but again that it was better for detecting amnestic
T
impairments than single-domain non-amnestic impairments [13]. Until now however, the
IP
domain-specific pattern of sensitivity for the MoCA has not been fully characterised. The aim
SC R
of our retrospective study was to compare intact performance on the MoCA, in regards to overall and domain-specific scores, with the “gold standard” of detailed neuropsychological
NU
assessment in a cohort of patients with acute stroke.
MA
Method
Three-hundred and sixty-four acute stroke patients admitted between January 2011 to
D
July 2013 to the Acute Stroke Unit of the National Hospital for Neurology and Neurosurgery
TE
(NHNN), London and assessed by the Neuropsychology Department, were included in a retrospective audit of available data. This unit accepts local stroke patients within 24 to 72
CE P
hours after admission to the University College Hospital Hyperacute Stroke Unit. All patients in the unit were assessed by a Clinical Neuropsychologist (E.C. or another member of the
AC
neuropsychology department, NHNN) as part of standard routine care. Inclusion criteria of the retrospective study were the availability of MoCA and neuropsychological data, and assessments had to be conducted within 3 months of the stroke event. Of those patients who had both MoCA and neuropsychological data available, six patients were excluded for having a history of psychiatric disorders and two patients were excluded for having a history of alcohol and drug abuse as determined from the hospital admissions/discharge report. The study was approved by the local clinical governance and ethics committees. One-hundred and thirty-six acute stroke patients were included. All patients were administered the MoCA followed by a neuropsychological assessment by the same clinician. Due to factors such as educational/occupational/cultural background/ability to cope with test
4
ACCEPTED MANUSCRIPT demands, the same tests and/or same cognitive domains were not administered/assessed in all patients. As such, 27 patients were administered only sub-sections of the MoCA. The
T
neuropsychological assessment comprised a standardized battery evaluating seven cognitive
IP
domains. However, as this was a retrospective study, patients were administered a subset of
SC R
tests deemed suitable by the Clinical Neuropsychologist at the time. Current general intellectual functioning was assessed using the Wechsler Adult Intelligence Scale—Third Edition (WAIS-III) [15]. Premorbid intellectual functioning was assessed using the National
NU
Adult Reading Test (NART) [16]. Verbal and visual memory functions were assessed with
MA
either the Recognition Memory Test [17]. or the Doors and People test [18] and naming skills were examined either with the Graded Naming Test [19] or the Oldfield Naming Test [20].
D
Perceptual functions were assessed using the Visual Object and Space Perception Battery
TE
[21]. Information processing speed was examined using one or more of the following tests: ‘O’ Cancellation, Digit Copy [22], Symbol Digit Modalities Test [23] or the Trail Making
CE P
Test (Part A) [24]. Executive functions were examined using one or more of the following tests: Stroop Test [25]; Trail Making Test Part B [23], Weigl Colour Form Sorting Task [26]
AC
or the Hayling and Brixton Test [27]. Testing lasted approximately 60 minutes in total and was generally conducted in one session unless if patients were too fatigued. Patients were classed as cognitively intact on the MOCA if they scored ≥ 25 out of 30 [12,13]. Derived scores on neuropsychological assessment were calculated based on published standardized normative data; performance in each cognitive domain was classed as impaired/unimpaired according to predefined criteria [28]. In general, scores at or below the fifth percentile on any one test were taken to indicate impairment in that respective domain. For intellectual functioning, a difference equal to or greater than 10 between performance on the WAIS-III Verbal or Performance IQ measures and the estimated premorbid functioning on the NART was taken as evidence of intellectual decline.
5
ACCEPTED MANUSCRIPT We conducted two main analyses: we first identified the patients classed as cognitively intact on the MoCA (i.e. MoCA-intact population) and investigated their
T
performance on detailed neuropsychological assessment; we then identified patients who
IP
scored the maximum points on each of the MoCA-specified cognitive domains, irrespective
SC R
of the overall score, and investigated their performance on neuropsychological assessments in the relevant domain and calculated the relevant negative predictive value (NPV).
NU
Results
MA
Thirty patients (22%) had MoCA-intact overall scores. Their demographic and clinical characteristics are shown in Table 1. There was no difference between the MoCA-
D
intact and MoCA-impaired group on relevant demographic variables. In contrast, the MoCA-
TE
intact group had significantly higher estimated premorbid intellectual functioning and higher
Insert Table 1 here.
AC
CE P
current verbal and non-verbal intelligence.
Twenty-three (77%) of the MoCA-intact population were impaired on at least one domain on detailed neuropsychological assessment. Of the 23 patients, sixteen patients (70%) were impaired in two or more cognitive domains while the remaining seven patients (30%) were impaired in one cognitive domain only. The distribution of neuropsychological impairment across cognitive domains in the MoCA-intact population is shown in Figure 1.
Insert Figure 1 here.
6
ACCEPTED MANUSCRIPT Of the MoCA-intact population, a high proportion of patients who were assessed on neuropsychological assessments had impairment in intellectual functioning (12 of 17
T
patients), speed of information processing (14 of 25 patients) and non-verbal memory (9 of
IP
25 patients); domains not assessed by the MoCA. Notably however, a high proportion of
SC R
patients also showed impairments in executive functioning (9 of 29 patients) and visuoperceptual/spatial domains (7 of 24 patients), which are assessed by the MOCA. Naming and verbal memory impairments were detected less frequently (3 of 30 and 3 of 28 patients
MA
NU
respectively).
For patients who scored the maximum points within MoCA-specified cognitive
D
domains irrespective of the overall score, their performance on neuropsychological
Insert Table 2 here.
AC
CE P
TE
assessment for each domain, and the NPV, are summarized in Table 2.
More than 50% of the patients were impaired on attention despite scoring the maximum points on this domain on the MoCA. A relatively high proportion of patients also showed impairments in memory (35%), visuospatial/executive function (30%), and naming (21%).
For completeness, we also examined the neuropsychological performance of the MoCA-impaired group and found that 103 out of the 106 patients (97%) were impaired on one or more cognitive domains. Examination of the three remaining patients revealed that they also showed cognitive deficits on neuropsychological assessment. Although they did not
7
ACCEPTED MANUSCRIPT score below the 5th percentile cut-off to be classed as “impaired”, all three patients were found to have mild to moderate executive dysfunction. Therefore, for our retrospective data
T
using a cut-off score of ≥ 25, the MoCA showed good sensitivity (0.82) but moderate
IP
specificity (0.7), very good positive predictive value (0.97) but very poor negative predictive
SC R
value (0.23).
NU
Discussion
MA
In our retrospective study, a striking 77% of patients classed as cognitively intact on the MoCA were found to be impaired across one or more cognitive domains on
D
neuropsychological assessment. The most common impairments were in intellectual
TE
functioning, speed of information processing, and visual memory; three domains not assessed by the MoCA. Impairments in these domains however, are frequently detected post-stroke
CE P
[1,3,4] and have high predictive value for functional outcome [29,30]. Although the MoCA may be useful in detecting gross cognitive impairments with a sensitivity of 0.82, it may have
AC
limitations in fully assessing relevant post-stroke impairments with a negative predictive value of 0.23.
Contrary to suggestions that poor specificity of the MoCA is related to poor detection of single-domain impairments [13,14], most of our patients were impaired on two or more cognitive domains. This may reflect the timing of assessments: whereas previous studies assessed patients at ≥1 year post-stroke [13,14], ours were assessed between 48h and 3 months post-stroke, when multiple-domain impairment may be more common [4]. We also found that intact performance within a MoCA-specified domain does not exclude impairment in that domain. More than half of our patients who scored the maximum points within the attention domain and one third of patients in the memory and
8
ACCEPTED MANUSCRIPT visuospatial/executive
domains
of
the
MoCA
were
impaired
on
comparable
neuropsychological assessment. Our findings corroborate with a recent study which found
T
poor correlation between performance on MoCA subtests and their neuropsychological
IP
counterpart [14]. The poor correspondence between performance on MoCA-specified
SC R
domains and comparable neuropsychological assessment is most likely due to the lack of breadth and depth of screening items, and the absence of normative data to take into account factors such as age and pre-morbid intellectual functioning. In addition, another contributing
NU
factor may be that the MoCA-specified domains do not accurately reflect the cognitive
MA
processes required to complete the relevant tasks. For example, the “attention” domain contains a digit-span backwards task and a subtraction task which are commonly regarded as
D
a test of working memory [15].
TE
Despite common apprehension regarding the feasibility of neuropsychological assessment during the acute phase post-stroke, our current study demonstrated that
CE P
neuropsychological assessment is possible and can be informative over and above a basic cognitive screen. A prospective study by van Zandwoort and colleagues [4] found that 74%
AC
of patients consecutively recruited on an acute stroke ward were able to complete a one and a half hour cognitive assessment battery 4-20 days post-stroke, with 77% of patients able to complete 82% of the assessment battery. Importantly, on follow-up, they found that testprofiles of patients was stable over time despite overall improvement, and early measure of intellectual functioning and other cognitive domains had significant predictive validity with respect to functional outcome. Our findings further corroborate that neuropsychological assessment is feasible; it does not necessarily need to be exhaustive but should nevertheless be comprehensive. As this was a retrospective study examining data that was collected as a part of routine clinical care, there are some limitations that need to be considered. First, the
9
ACCEPTED MANUSCRIPT proportion of patients that were included in the study was relatively small (37%). This is much lower than previous feasibility studies looking at completion rate of the MoCA and
T
neuropsychological assessment in stroke populations [4,8] and most likely reflects the post-
IP
hoc nature of the current study design rather than an anomaly with the patient sample.
SC R
Second, as there was no a priori neuropsychological battery, patients did not all complete the same set of tests and/or in the same order and it is difficult to determine retrospectively the reasons for missing data which may be clinically relevant. In addition, data regarding mood at
NU
time of assessment was not available for consideration. Finally, the clinician was not blind to
MA
the results of the MoCA prior to administering neuropsychological assessment which may have biased the selection of neuropsychological tests used.
D
In conclusion, our findings suggest that the MoCA does not fully assess cognitive
TE
domains that are commonly affected following stroke. Furthermore, intact performance on MoCA-specified domains does not preclude the presence of domain-specific cognitive
CE P
impairments. Our research advocates for the use of neuropsychological assessment tools for detecting cognitive impairment post-stroke. However, current resource limitations in local
AC
health services may not always allow for the provision of adequate neuropsychological services. Further research using a more rigid prospective design is needed to determine the potential clinical and functional importance and benefits of neuropsychological impairments not detected by the MoCA to help inform the development of more sensitive screening instruments for acute stroke.
Disclosures: None
10
ACCEPTED MANUSCRIPT References 1. Barker-Collo S & Feigin V. The impact of neuropsychological deficits on functional
IP
T
stroke outcomes. Neuropsychol Rev. 2006; 16: 53-64.
SC R
2. Galski, T., Bruno, R.L., Zorowitz, R. & Walker J. Predicting length of stay, functional outcome, and aftercare in the rehabilitation of stroke patients. The dominant role of higher-
NU
order cognition. Stroke. 1993; 24: 1794-1800.
MA
3. Tatemichi TK, Desmond DW, Stern Y, Paik M, Sano M & Bagiella E. Cognitive impairment after stroke: frequency, patterns, and relationship to functional abilities. Jour of
TE
D
Neur, Neurosurg, and Psych. 1994; 57: 202-207.
4. Van Zandvoort MJE, Kessels PC, Nys GMS, de Haan EHF & Kappelle LJ. Early
CE P
neuropsychological evaluation in patients with ischaemic stroke provides valid information.
AC
Clinical Neurology and Neurosurgery. 2005; 107: 385-392.
5. Blackburn DJ, Bafadhel L, Randall M & Harkness KA. Cognitive screening in the acute stroke setting. Age Ageing. 2013; 42:113-116.
6. Dong Y, Sharma VK, Chan BP, Venketasubramanian N, Teoh HL, Seet RC et al. The Montreal Cognitive Assessment (MoCA) is superior to the Mini-Mental State Examination (MMSE) for the detection of vascular cognitive impairment after acute stroke. J Neurol Sci. 2010; 299:15-18.
11
ACCEPTED MANUSCRIPT 7. Pendlebury ST, Cuthbertson FC, Welch SJV, Mehta Z & Rothwell PM. Underestimation of cognitive impairment by mini-mental state examination versus the montreal cognitive
T
assessment in patients with transient ischemic attack and stroke: A population-based study.
SC R
IP
Stroke. 2010; 41: 1290-1293.
8. Cumming TB, Bernhardt J & Linden T. The montreal cognitive assessment: Short
NU
cognitive evaluation in a large stroke trial. Stroke. 2011; 42: 2642-2644.
MA
9. Intercollegiate Stroke Working Party. National clinical guideline for stroke, 4th edition.
D
London: Royal College of Physicians, 2012.
TE
10. National Institute for Health and Clinical Excellence (2013). Stroke rehabilitation: Long
CE P
term rehabilitation after stroke. London: NICE (www.guidance.nice.org.uk/CG162).
11. Nasreddine ZS, Phillips NA, Bédirian V, Charbonneau S, Whitehead V, Collin I, et al.
AC
The montreal cognitive assessment, MoCA: A brief screening tool for mild cognitive impairment. J Am Geriatr Soc. 2005; 53: 695-699.
12. Godefroy O, Fickl A, Roussel M, Auribault C, Bugnicourt JM, Lamy C, et al. Is the Montreal Cognitive Assessment superior to the Mini-Mental State Examination to detect poststroke cognitive impairment? A study with neuropsychological evaluation. Stroke. 2011; 42:1712-6.
13. Pendlebury ST, Mariz J, Bull L, Mehta Z & Rothwell PM. MoCA, ACE-R, and MMSE versus the National Institute of Neurological Disorders and Stroke-Canadian Stroke Network
12
ACCEPTED MANUSCRIPT Vascular Cognitive Impairment Harmonization Standards Neuropsychological Battery after
T
TIA and stroke. Stroke. 2012;43:464-9.
IP
14. Schweizer TA, Al-Khindi T & Macdonald RL. Mini-Mental State Examination versus
SC R
Montreal Cognitive Assessment: rapid assessment tools for cognitive and functional outcome after aneurysmal subarachnoid hemorrhage. J Neurol Sci. 2012;316:137-40.
NU
15. Wechsler, D. Wechsler Adult Intelligence Scale – 3rd Edition. San Antonio, TX: Harcourt
MA
Assessment; 1997.
D
16. Nelson, H.E. The National Adult Reading Test (NART): for the assessment of premorbid
TE
intelligence in patients with dementia. End ed. Windsor (UK): NFER-Nelson; 1991.
CE P
17. Warrington, E.K. Recognition Memory Test. Windsor (UK): NFER-Nelson, 1984.
AC
18. Baddeley, A.D., Emslie, H. & Nimmo-Smith, I. Doors and people: a test opf visual and verbal recall and recognition. Bury St. Edmunds, England: Thames Valley Test Co, 1994.
19. McKenna, P. & Warrington, E.K. Graded Naming Test. Windsor (UK): NFER-Nelson; 1983.
20. Oldfield, R.C. & Wingfield, A., Response latencies in naming objects. Q J Exp Psychol. 1965; 17: 273-281.
13
ACCEPTED MANUSCRIPT 21. Warrington, E.K. & James, M. Visual Object and Space Perception Battery. Bury St.
T
Edmunds (UK): Thames Valley Test Company; 1991.
SC R
psychomotor speed. Behav Neurol. 1992; 5: 113-116.
IP
22. Willison, J.R. & Warrington, E.K. Cognitive retardation in a patient with preservation of
MA
Western Psychological Services; 1982.
NU
23. Smith, A. Symbol Digit Modalities Test (SDMT) Manual (Revised). Los Angeles:
24. Reitan, R.M. (1958). Validity of the trail making test as an indicator of organic brain
TE
D
damage. Perceptual and Motor Skills. 1958; 8: 271-276.
25. Trennary, M.R. Crossen, B., DeBoe, J., & Leber, W.R. Stroop Neuropsychological
CE P
Screening Test (SNST). Odessa (FL): Psychological Assessment Resources: 1989.
AC
26. Weigl, E. On the psychology of so-called processes of abstraction. J Abnormal Soc Psychol. 1941; 36: 3-33.
27. Burgess, P.W. & Shallice, T. The Hayling and Brixton Tests. Technical report. Bury St. Edmunds (UK): Thames Valley Test Company; 1997.
28. Werring DJ, Frazer DW, Coward LJ, Losseff NA, Watt H. Cipolotti L et al. Cognitive dysfunction in patients with cerebral microbleeds on T2*-weighted gradient-echo MRI. Brain. 2004; 127: 2265-2275.
14
ACCEPTED MANUSCRIPT 29. Nys GMS, van Zandvoort MJE, de Kort PLM, van der Worp HB, Jansen BPW, Algra A et al.. The prognostic value of domain-specific cognitive abilities in acute first-ever stroke.
IP
T
Neurology. 2005; 64: 821-827.
SC R
30. Hurford, R and Charidimou, A and Fox, Z and Cipolotti, L and Werring, DJ. Domainspecific trends in cognitive impairment after acute ischaemic stroke. J Neurol. 2013; 260: 237
AC
CE P
TE
D
MA
NU
- 241.
15
ACCEPTED MANUSCRIPT Figures
AC
CE P
TE
D
MA
NU
SC R
IP
T
Figure 1. MOCA-intact population: Neuropsychological performance across domains.
16
ACCEPTED MANUSCRIPT Table 1 Demographic and clinical characteristics.
(n=136)
Intact
(n=106)
Impaired
16.0 (6.1)
p
