Parkinsonism and Related Disorders 20 (2014) 627e631
Contents lists available at ScienceDirect
Parkinsonism and Related Disorders journal homepage: www.elsevier.com/locate/parkreldis
The utility of the Mattis Dementia Rating Scale in Parkinson’s disease mild cognitive impairment Eva Pirogovsky a, b, Dawn M. Schiehser a, b, Irene Litvan c, Kristalyn M. Obtera a, Mathes M. Burke a, Stephanie L. Lessig a, c, David D. Song a, c, Lin Liu d, J. Vincent Filoteo a, b, * a
Veterans Affairs, San Diego Health Care System, San Diego, CA, USA Department of Psychiatry, University of California San Diego, San Diego, CA, USA Department of Neurosciences, Movement Disorder Center, University of California, San Diego, CA, USA d Division of Biostatistics and Bioinformatics, University of California, San Diego, CA, USA b c
a r t i c l e i n f o
a b s t r a c t
Article history: Received 10 December 2013 Received in revised form 26 February 2014 Accepted 10 March 2014
Background: The Movement Disorders Society (MDS) recently proposed guidelines for diagnosis of mild cognitive impairment in Parkinson’s disease (PD-MCI) that includes two assessment levels: abbreviated (Level I) and comprehensive (Level II). The aim of this study was to determine the utility of the Mattis Dementia Rating Scale (MDRS), a recommended Level I test, for detecting Level II PD-MCI diagnosis. Methods: The study sample included 30 patients diagnosed with PD-MCI based on Level II MDS criteria and 68 PD patients with normal cognition (PD-NC). Receiver operator curve (ROC) analyses were generated to measure the sensitivity and specificity of various MDRS cutoff scores. To examine the utility of the MDRS as a screening tool, the optimal cutoff point was defined as the lowest value providing 80% sensitivity. For use of the MDRS as a diagnostic tool, the optimal cutoff point was defined as the highest value providing 80% specificity. Results: ROC analyses showed that the optimal MDRS cutoff score for screening purposes and diagnostic purposes were 140 and 137, respectively. However, an examination of sensitivity/specificity values for the screening cutoff scores suggested that a total score of 139 for screening purposes yielded a better balance between sensitivity (77%) and specificity (65%). Conclusions: In a clinical setting, in which detection of PD-MCI may be important, a total MDRS score of 139 can be used to detect PD-MCI. In research and other settings in which diagnostic certainty is more important, a score of 137 may be more useful. Published by Elsevier Ltd.
Keywords: Parkinson’s disease-mild cognitive impairment Mattis Dementia Rating Scale Optimal cuttoff scores
1. Introduction Parkinson’s disease mild cognitive impairment (PD-MCI) is common and a potential prodromal stage of PD dementia (PDD). Studies examining PD-MCI may lead to early interventions to delay onset of PDD and improve quality of life [1]. A major concern, however, has been a lack of uniformity in the criteria used to diagnose PD-MCI. The Movement Disorder Society (MDS) recently commissioned a task force to outline diagnostic criteria for PD-MCI, which provide a uniform framework that researchers can now use to study PD-MCI [2]. These criteria include two levels of assessment
* Corresponding author. UC San Diego/VA, 3350 La Jolla Village Drive San Diego, CA 92161-116A, USA. Tel.: þ1 858 552 8585x1122; fax: þ1 858 642 3023. E-mail address: vfi[email protected] (J.V. Filoteo). http://dx.doi.org/10.1016/j.parkreldis.2014.03.010 1353-8020/Published by Elsevier Ltd.
that differ based on the comprehensiveness of the cognitive assessment. Level I criteria is an abbreviated assessment that includes impairment on a scale of global cognition or impairment on a limited battery of neuropsychological tests, while Level II criteria involve a more extensive neuropsychological battery. An important area of research is to determine what Level I screening instruments of global cognition might be best at identifying PD-MCI based on Level II criteria. The Mattis Dementia Rating Scale (MDRS) [3] is an MDS Task Force recommended Level I screening instrument. The current study examined the utility of the MDRS to detect a diagnosis of PD-MCI that was based on Level II criteria. Although prior studies have examined the validity of the MDRS in identifying PD-MCI [4,5], no study to date has examined the ability of the MDRS to detect PD-MCI using these new criteria. The current study also examined the profile of impairment on the MDRS subscales in PD-MCI.
628
E. Pirogovsky et al. / Parkinsonism and Related Disorders 20 (2014) 627e631
2. Methods 2.1. Participants The Veteran Affairs IRB approved the study and all participants gave verbal and written informed consent. The study included a convenience sample of 98 PD patients recruited from the Parkinson’s Disease Research Subject Database of the San Diego VA Health Care System/University of California at San Diego who met the following inclusion criteria: age of , optimal cuttoff scores 50 years; diagnosed by a board-certified neurologist specialized in movement disorders as having probable PD according to the UK Brain Bank criteria (Hughes, Daniel, Kilford, & Lees, 1992); without a history of neurologic conditions other than PD, major depression, severe mental illness (e.g., schizophrenia), substance abuse, or neurosurgery, and dementia based on formal MDS task force criteria [6]. 2.2. Procedures All participants received a comprehensive neuropsychological assessment that evaluated five cognitive domains: attention and working memory, language, memory, visuospatial functioning, and executive functioning (see Table 1). Participants were also administered the following measures as part of the assessment: Frontal Systems Behavior Scale (FrSBe) [7], the Informant Questionnaire on Cognitive Decline in the Elderly (IQCODE) [8], Geriatric Depression Scale (GDS) [9], PD Questionnaire 39-item version (PDQ-39) [10], and Lawton and Brody Activities of Daily Living scale (IADL questionnaire) [11]. For most participants, tests were administered over the course of two visits, with a mean of 8 days (standard deviation ¼ 8.9) between visits. The original version of the MDRS was always administered on the first visit. 2.3. PD-MCI classification PD-MCI was diagnosed following Level II MDS Task Force Criteria [2] which include: (1) cognitive decline reported by the patient or caregiver, (2) cognitive deficits not severe enough to significantly interfere with functional independence, and (3) cognitive deficits on formal neuropsychological testing. Cognitive complaint was assessed with the Executive Dysfunction subscale of the informant version of the FrSBe and the IQCODE (with the exception of functional decline questions; items 23 and 24). For patients who did not identify an informant (n ¼ 18), cognitive complaint was assessed with cognitive impairment items [12] from the GDS and cognition subscale of the PDQ-39. To exclude patients with PDD, functional independence was assessed with the IADL questionnaire; and items 23 and 24 on the IQCODE (items representing functional decline). For those patients without an informant, we used the Unified Parkinson’s Disease Rating Scale (UPDRS [13] Part I, item 1.1.), which assesses the effect of cognitive impairment on daily functioning. Significant functional impairment was determined using clinical decision based upon reviewing these questionnaires. As per MDS criteria, PD-MCI was diagnosed if there was impairment on at least two neuropsychological tests represented by either two impaired tests in one cognitive domain or one impaired test in two different cognitive domains. MDS criteria [2] suggest a cutoff for impairment on neuropsychological tests between 1 and 2 standard deviations below appropriate norms and the cutoff for impairment in the current study was set at z 1.33. This cutoff score was chosen because the published standard scores available for the neuropsychological measures in the included battery of tests are on varying scales (e.g., scaled score, T-score, z-score). A standard score that is equivalent to a z-score of 1.33 can be computed for each of the other scales (scaled score ¼ 6, Tscore ¼ 37); therefore, this z-score cutoff allows for scoring consistency in the administered tests that use different standard scores. PD patients were classified as PD-MCI or PD-NC by three neuropsychologists (Drs. Pirogovsky, Schiehser, and Filoteo). Two of the three neuropsychologists were randomized to classify each patient, so that each patient was given a diagnosis by two of the three neuropsychologists. After each rater independently classified the cognitive status of the PD patients, all three raters convened at a consensus conference to discuss any diagnostic discrepancy. For each discrepancy, the diagnostic category that had the majority of consensus ratings was entered as the final diagnosis.
Table 1 Demographic and clinical characteristics of study participants.
Age (years) Education (years) Gender (% male and count M/F) Ethnicity (% Caucasian and count Caucasian/other) Geriatric Depression Scale Duration of PD (years) UPDRS Part III Modified Hoehn and Yahr stage Stage 0 Stage 1 Stage 1.5 Stage 2 Stage 2.5 Stage 3 Stage 4 Medication type L-dopa Dopamine Agonist MAO-B inhibitor (Selegiline or Rasagaline) COMT-inhibitor (Entacapone) Amantadine Anticholinergic Measures used in Diagnosis of PD-MCI Attention D-KEFS CWIT Color Naming DOTA Memory CVLT-II Long Delay Free Recall WMS-III Logical Memory II Executive Function WCST Perseverative responses D-KEFS CWIT Inhibition Language MDRS Similarities D-KEFS Category Fluency Visuospatial Function JLOT WMS-III Visual Reproduction Copy MDRS Total score Attention scale Initiation/Perseveration scale Construction scale Conceptualization scale Memory scale
PD-MCI
PD-NC
(n ¼ 30)
(n ¼ 68)
67.5 (7.7) 15.9 (2.8) 77 23/7 96 29/1
68.8 (7.5) 16.9 (2.5) 60 41/27 96 65/3
0.45 0.07 0.11 0.9
6.5 (6.1) 4.7 (3.8) 23.0 (11.9) % Count 1.5 1 13.2 9 2.9 2 54.4 37 7.4 5 17.6 12 2.9 2 % Count 63 19 66 20 43 13
6.1 (4.9) 6.1 (5.7) 25.0 (12.6) % Count 0 0 26.7 8 0 0 56.7 17 3.3 1 10.0 3 3.3 1 % Count 65 44 63 43 57 39
0.45 0.25 0.44 0.64
3 3 0
4 12 1
1 0.27 1
1 1 0
P-value
3 8 1
1 0.82 0.27
36.2 (6.8) 4.8 (1.6)
31.5 (5.6) 6.5 (1.9)
Contents lists available at ScienceDirect
Parkinsonism and Related Disorders journal homepage: www.elsevier.com/locate/parkreldis
The utility of the Mattis Dementia Rating Scale in Parkinson’s disease mild cognitive impairment Eva Pirogovsky a, b, Dawn M. Schiehser a, b, Irene Litvan c, Kristalyn M. Obtera a, Mathes M. Burke a, Stephanie L. Lessig a, c, David D. Song a, c, Lin Liu d, J. Vincent Filoteo a, b, * a
Veterans Affairs, San Diego Health Care System, San Diego, CA, USA Department of Psychiatry, University of California San Diego, San Diego, CA, USA Department of Neurosciences, Movement Disorder Center, University of California, San Diego, CA, USA d Division of Biostatistics and Bioinformatics, University of California, San Diego, CA, USA b c
a r t i c l e i n f o
a b s t r a c t
Article history: Received 10 December 2013 Received in revised form 26 February 2014 Accepted 10 March 2014
Background: The Movement Disorders Society (MDS) recently proposed guidelines for diagnosis of mild cognitive impairment in Parkinson’s disease (PD-MCI) that includes two assessment levels: abbreviated (Level I) and comprehensive (Level II). The aim of this study was to determine the utility of the Mattis Dementia Rating Scale (MDRS), a recommended Level I test, for detecting Level II PD-MCI diagnosis. Methods: The study sample included 30 patients diagnosed with PD-MCI based on Level II MDS criteria and 68 PD patients with normal cognition (PD-NC). Receiver operator curve (ROC) analyses were generated to measure the sensitivity and specificity of various MDRS cutoff scores. To examine the utility of the MDRS as a screening tool, the optimal cutoff point was defined as the lowest value providing 80% sensitivity. For use of the MDRS as a diagnostic tool, the optimal cutoff point was defined as the highest value providing 80% specificity. Results: ROC analyses showed that the optimal MDRS cutoff score for screening purposes and diagnostic purposes were 140 and 137, respectively. However, an examination of sensitivity/specificity values for the screening cutoff scores suggested that a total score of 139 for screening purposes yielded a better balance between sensitivity (77%) and specificity (65%). Conclusions: In a clinical setting, in which detection of PD-MCI may be important, a total MDRS score of 139 can be used to detect PD-MCI. In research and other settings in which diagnostic certainty is more important, a score of 137 may be more useful. Published by Elsevier Ltd.
Keywords: Parkinson’s disease-mild cognitive impairment Mattis Dementia Rating Scale Optimal cuttoff scores
1. Introduction Parkinson’s disease mild cognitive impairment (PD-MCI) is common and a potential prodromal stage of PD dementia (PDD). Studies examining PD-MCI may lead to early interventions to delay onset of PDD and improve quality of life [1]. A major concern, however, has been a lack of uniformity in the criteria used to diagnose PD-MCI. The Movement Disorder Society (MDS) recently commissioned a task force to outline diagnostic criteria for PD-MCI, which provide a uniform framework that researchers can now use to study PD-MCI [2]. These criteria include two levels of assessment
* Corresponding author. UC San Diego/VA, 3350 La Jolla Village Drive San Diego, CA 92161-116A, USA. Tel.: þ1 858 552 8585x1122; fax: þ1 858 642 3023. E-mail address: vfi[email protected] (J.V. Filoteo). http://dx.doi.org/10.1016/j.parkreldis.2014.03.010 1353-8020/Published by Elsevier Ltd.
that differ based on the comprehensiveness of the cognitive assessment. Level I criteria is an abbreviated assessment that includes impairment on a scale of global cognition or impairment on a limited battery of neuropsychological tests, while Level II criteria involve a more extensive neuropsychological battery. An important area of research is to determine what Level I screening instruments of global cognition might be best at identifying PD-MCI based on Level II criteria. The Mattis Dementia Rating Scale (MDRS) [3] is an MDS Task Force recommended Level I screening instrument. The current study examined the utility of the MDRS to detect a diagnosis of PD-MCI that was based on Level II criteria. Although prior studies have examined the validity of the MDRS in identifying PD-MCI [4,5], no study to date has examined the ability of the MDRS to detect PD-MCI using these new criteria. The current study also examined the profile of impairment on the MDRS subscales in PD-MCI.
628
E. Pirogovsky et al. / Parkinsonism and Related Disorders 20 (2014) 627e631
2. Methods 2.1. Participants The Veteran Affairs IRB approved the study and all participants gave verbal and written informed consent. The study included a convenience sample of 98 PD patients recruited from the Parkinson’s Disease Research Subject Database of the San Diego VA Health Care System/University of California at San Diego who met the following inclusion criteria: age of , optimal cuttoff scores 50 years; diagnosed by a board-certified neurologist specialized in movement disorders as having probable PD according to the UK Brain Bank criteria (Hughes, Daniel, Kilford, & Lees, 1992); without a history of neurologic conditions other than PD, major depression, severe mental illness (e.g., schizophrenia), substance abuse, or neurosurgery, and dementia based on formal MDS task force criteria [6]. 2.2. Procedures All participants received a comprehensive neuropsychological assessment that evaluated five cognitive domains: attention and working memory, language, memory, visuospatial functioning, and executive functioning (see Table 1). Participants were also administered the following measures as part of the assessment: Frontal Systems Behavior Scale (FrSBe) [7], the Informant Questionnaire on Cognitive Decline in the Elderly (IQCODE) [8], Geriatric Depression Scale (GDS) [9], PD Questionnaire 39-item version (PDQ-39) [10], and Lawton and Brody Activities of Daily Living scale (IADL questionnaire) [11]. For most participants, tests were administered over the course of two visits, with a mean of 8 days (standard deviation ¼ 8.9) between visits. The original version of the MDRS was always administered on the first visit. 2.3. PD-MCI classification PD-MCI was diagnosed following Level II MDS Task Force Criteria [2] which include: (1) cognitive decline reported by the patient or caregiver, (2) cognitive deficits not severe enough to significantly interfere with functional independence, and (3) cognitive deficits on formal neuropsychological testing. Cognitive complaint was assessed with the Executive Dysfunction subscale of the informant version of the FrSBe and the IQCODE (with the exception of functional decline questions; items 23 and 24). For patients who did not identify an informant (n ¼ 18), cognitive complaint was assessed with cognitive impairment items [12] from the GDS and cognition subscale of the PDQ-39. To exclude patients with PDD, functional independence was assessed with the IADL questionnaire; and items 23 and 24 on the IQCODE (items representing functional decline). For those patients without an informant, we used the Unified Parkinson’s Disease Rating Scale (UPDRS [13] Part I, item 1.1.), which assesses the effect of cognitive impairment on daily functioning. Significant functional impairment was determined using clinical decision based upon reviewing these questionnaires. As per MDS criteria, PD-MCI was diagnosed if there was impairment on at least two neuropsychological tests represented by either two impaired tests in one cognitive domain or one impaired test in two different cognitive domains. MDS criteria [2] suggest a cutoff for impairment on neuropsychological tests between 1 and 2 standard deviations below appropriate norms and the cutoff for impairment in the current study was set at z 1.33. This cutoff score was chosen because the published standard scores available for the neuropsychological measures in the included battery of tests are on varying scales (e.g., scaled score, T-score, z-score). A standard score that is equivalent to a z-score of 1.33 can be computed for each of the other scales (scaled score ¼ 6, Tscore ¼ 37); therefore, this z-score cutoff allows for scoring consistency in the administered tests that use different standard scores. PD patients were classified as PD-MCI or PD-NC by three neuropsychologists (Drs. Pirogovsky, Schiehser, and Filoteo). Two of the three neuropsychologists were randomized to classify each patient, so that each patient was given a diagnosis by two of the three neuropsychologists. After each rater independently classified the cognitive status of the PD patients, all three raters convened at a consensus conference to discuss any diagnostic discrepancy. For each discrepancy, the diagnostic category that had the majority of consensus ratings was entered as the final diagnosis.
Table 1 Demographic and clinical characteristics of study participants.
Age (years) Education (years) Gender (% male and count M/F) Ethnicity (% Caucasian and count Caucasian/other) Geriatric Depression Scale Duration of PD (years) UPDRS Part III Modified Hoehn and Yahr stage Stage 0 Stage 1 Stage 1.5 Stage 2 Stage 2.5 Stage 3 Stage 4 Medication type L-dopa Dopamine Agonist MAO-B inhibitor (Selegiline or Rasagaline) COMT-inhibitor (Entacapone) Amantadine Anticholinergic Measures used in Diagnosis of PD-MCI Attention D-KEFS CWIT Color Naming DOTA Memory CVLT-II Long Delay Free Recall WMS-III Logical Memory II Executive Function WCST Perseverative responses D-KEFS CWIT Inhibition Language MDRS Similarities D-KEFS Category Fluency Visuospatial Function JLOT WMS-III Visual Reproduction Copy MDRS Total score Attention scale Initiation/Perseveration scale Construction scale Conceptualization scale Memory scale
PD-MCI
PD-NC
(n ¼ 30)
(n ¼ 68)
67.5 (7.7) 15.9 (2.8) 77 23/7 96 29/1
68.8 (7.5) 16.9 (2.5) 60 41/27 96 65/3
0.45 0.07 0.11 0.9
6.5 (6.1) 4.7 (3.8) 23.0 (11.9) % Count 1.5 1 13.2 9 2.9 2 54.4 37 7.4 5 17.6 12 2.9 2 % Count 63 19 66 20 43 13
6.1 (4.9) 6.1 (5.7) 25.0 (12.6) % Count 0 0 26.7 8 0 0 56.7 17 3.3 1 10.0 3 3.3 1 % Count 65 44 63 43 57 39
0.45 0.25 0.44 0.64
3 3 0
4 12 1
1 0.27 1
1 1 0
P-value
3 8 1
1 0.82 0.27
36.2 (6.8) 4.8 (1.6)
31.5 (5.6) 6.5 (1.9)
