INTRODUCTION
Parkinson’s disease (PD) is the second-most common neurodegenerative disorder in the United States of America (Kowal, Dall, Chakrabarti, Storm, & Jain, Reference Kowal, Dall, Chakrabarti, Storm and Jain2013) and is characterized by both motor and non-motor symptoms, including cognitive deficits. There is considerable heterogeneity in the expression of cognitive deficits among PD patients (Kehagia, Barker, & Robbins, Reference Kehagia, Barker and Robbins2010), but deficits in learning and memory, attention/working memory, and executive function are common (Kehagia, Barker, & Robbins, Reference Kehagia, Barker and Robbins2013; Kehagia et al., Reference Kehagia, Barker and Robbins2010). These cognitive deficits have been recognized even in the early stages of PD (Muslimović, Schmand, Speelman, & De Haan, Reference Muslimović, Schmand, Speelman and De Haan2007; Olchik, Ayres, Ghisi, Schuh, & Rieder, Reference Olchik, Ayres, Ghisi, Schuh and Rieder2016; Stefanova et al., Reference Stefanova, Žiropadja, Stojković, Stanković, Tomić, Ječmenica-Lukić and Kostić2015) and warrant a diagnosis of mild cognitive impairment (MCI) in approximately 27% of patients (Litvan et al., Reference Litvan, Goldman, Tröster, Schmand, Weintraub, Petersen and Emre2012). PD-MCI is often a precursor to dementia in PD (Hoogland et al., Reference Hoogland, Boel, de Bie, Schmand, Geskus, Dalrymple-Alford and Junque2019; Leroi, Pantula, McDonald, & Harbishettar, Reference Leroi, Pantula, McDonald and Harbishettar2012; Saredakis, Collins-Praino, Gutteridge, Stephan, & Keage, Reference Saredakis, Collins-Praino, Gutteridge, Stephan and Keage2019), rendering the early detection of PD-MCI and accompanying cognitive deficits critical for the purposes of monitoring and treatment (Barone et al., Reference Barone, Aarsland, Burn, Emre, Kulisevsky and Weintraub2011).
Subjective cognitive complaints – either by a patient, informant, or clinician – are often the first indication that a patient may have bonafide cognitive deficits (Erro et al., Reference Erro, Santangelo, Barone, Picillo, Amboni, Longo and Vitale2014; Jessen, Reference Jessen2014). However, there is substantial debate regarding the utility of subjective complaints in clinical practice as the concordance of subjective complaints and objective deficits are not well established and likely confounded by several factors (Copeland, Lieberman, Oravivattanakul, & Tröster, Reference Copeland, Lieberman, Oravivattanakul and Tröster2016; Naismith, Pereira, Shine, & Lewis, Reference Naismith, Pereira, Shine and Lewis2011), such as the relationship of the reporter to the patient, sample characteristics (e.g., cognitive status, mood symptoms), potential over- or under-estimation of symptoms, and assessment tools utilized (Rabin et al., Reference Rabin, Smart, Crane, Amariglio, Berman, Boada and Ellis2015). Evaluating the relationship of subjective and objective cognition in PD within the context of these potential confounds is important in order to establish the utility and validity of subjective cognitive assessment.
Subjective cognitive decline, defined as the perceived experience of change and deterioration of cognitive performance (Studart Neto & Nitrini, Reference Studart Neto and Nitrini2016), is a risk marker of dementia (Jessen & Rodriguez Francisca, Reference Jessen and Rodriguez (Née Then) Francisca2018) and future objectively established cognitive decline in non-demented PD (Erro et al., Reference Erro, Santangelo, Barone, Picillo, Amboni, Longo and Vitale2014; Jessen & Rodriguez Francisca, Reference Jessen and Rodriguez (Née Then) Francisca2018). Furthermore, subjective reports of cognitive decline, in addition to objectively measured deficits, are required for a diagnosis of PD-MCI (Litvan et al., Reference Litvan, Goldman, Tröster, Schmand, Weintraub, Petersen and Emre2012). Despite the critical importance of subjective reports of decline, self-reports can be problematic due to PD patients’ possible lack of insight into their deficits, especially within the context of cognitive impairment (Lehrner et al., Reference Lehrner, Kogler, Lamm, Moser, Klug, Pusswald and Auff2015; Seltzer, Vasterling, Mathias, & Brennan, Reference Seltzer, Vasterling, Mathias and Brennan2001). Inaccurate self-reports have been cited as a rationale for the exclusion of subjective cognitive complaints from prior PD-MCI criteria (Copeland et al., Reference Copeland, Lieberman, Oravivattanakul and Tröster2016). Reliable informants may provide valuable information about a patient’s cognitive decline (Goldman et al., Reference Goldman, Vernaleo, Camicioli, Dahodwala, Dobkin, Ellis and Simmonds2018), especially when the patient exhibits frontal behaviors (Zgaljardic, Borod, Foldi, & Mattis, Reference Zgaljardic, Borod, Foldi and Mattis2003), lacks awareness, or minimizes symptoms (Papay et al., Reference Papay, Mamikonyan, Siderowf, Duda, Lyons, Pahwa and Weintraub2011). As objective cognitive impairments increase, PD patients’ insight into their deficits may decrease (Lehrner et al., Reference Lehrner, Kogler, Lamm, Moser, Klug, Pusswald and Auff2015), which may render informant-based measures even more important as the disease progresses. Thus, informant-based subjective measures of cognitive decline have been recommended due to potential sensitivity to subtle changes in patients’ symptoms (Naismith et al., Reference Naismith, Pereira, Shine and Lewis2011).
Although informant-based measures of patients’ cognitive decline could provide insight into everyday difficulties that PD patients may experience, the concordance between these reports and objective testing has been debated (Erro et al., Reference Erro, Santangelo, Barone, Picillo, Amboni, Longo and Vitale2014). On one hand, Koerts et al. (Reference Koerts, van Beilen, Leenders, Brouwer, Tucha and Tucha2012) found that informant-based measures of cognition and objective cognition were not related in PD. However, this study only evaluated subjective and objective cognition with executive function tests (specifically, dysexecutive behavioral changes); thus, limited conclusions can be drawn for other cognitive domains. Other studies have found that informant-based measures are associated with patients’ objective cognitive performance. For example, Naismith et al. (Reference Naismith, Pereira, Shine and Lewis2011) found that informant reports of cognitive decline in memory/orientation were significantly correlated with lower objective performance on tests of psychomotor speed, learning/memory, language, and executive functioning in patients with PD-MCI. Likewise, Cooper et al. (Reference Cooper, Benge, Lantrip and Soileau2017) found that higher scores on the Everyday Cognition Scale, an informant-based measure of patients’ decline in cognitively mediated daily activities over the past 10 years, were significantly correlated with poorer overall cognition in PD patients (combined sample of 12 PD-normal controls, 24 probable PD-MCI, and 13 probable PD-dementia), even after controlling for disease duration and demographic factors (i.e., age and education; Cooper, Benge, Lantrip, & Soileau, Reference Cooper, Benge, Lantrip and Soileau2017). However, both studies did not control for mood symptoms or other demographic factors (i.e., gender). As mood and gender have been shown to impact self-reported cognitive decline (Jiménez-Huete et al., Reference Jiménez-Huete, Del Barrio, Riva, Campo, Toledano and Franch2017), it is important to account for these variables when assessing the relationship between informant-based measures and objective cognitive decline.
Importantly, none of the aforementioned studies (i.e., Cooper et al., Reference Cooper, Benge, Lantrip and Soileau2017; Koerts et al., Reference Koerts, van Beilen, Leenders, Brouwer, Tucha and Tucha2012; Naismith et al., Reference Naismith, Pereira, Shine and Lewis2011) examined relationships between informant-based measures of cognitive decline controlling for cognitive status. Understanding this relationship while accounting for a range of cognitive diagnoses is important given that cognitive deficits can diminish insight and subsequently impact the quality of self-reports (Lehrner et al., Reference Lehrner, Kogler, Lamm, Moser, Klug, Pusswald and Auff2015), particularly in those with MCI or dementia (Vogel et al., Reference Vogel, Stokholm, Gade, Andersen, Hejl and Waldemar2004). Thus, providers may rely on informant – particularly family member/close friend – reports of patient functioning and decline as a primary source of information (Potter et al., Reference Potter, Plassman, Burke, Kabeto, Langa, Llewellyn and Steffens2009). However, informant reports may be influenced by the informant’s perceived caregiver burden or own psychological symptoms (e.g., depression, anxiety; Kudlicka, Clare, & Hindle, Reference Kudlicka, Clare and Hindle2011; Morrell et al., Reference Morrell, Camic and Genis2019), which calls into question the accuracy of their reports. Moreover, other factors, such as patient mood and disease severity, may also impact informant reports of cognition (Morrell, Camic, & Genis, Reference Morrell, Camic and Genis2019). Therefore, evaluating the concordance of informant-based measures and objective cognitive functioning controlling for relevant factors is essential for the clinical care of PD patients.
There is currently no gold standard to assess for subjective cognitive decline in PD (Goldman et al., Reference Goldman, Vernaleo, Camicioli, Dahodwala, Dobkin, Ellis and Simmonds2018; Kjeldsen & Damholdt, Reference Kjeldsen and Damholdt2019). However, one of the most commonly utilized comprehensive measures of informant-based subjective cognitive decline in non-PD associated MCI and dementia is the Informant Questionnaire of Cognitive Decline in the Elderly (IQCODE; Ding et al., Reference Ding, Niu, Zhang, Liu, Zhou, Wei and Liu2018; Jorm, Reference Jorm2004), which assesses change (i.e., improved, worsened, and no change) on a wide range of cognitive abilities over a 10-year period. In non-PD samples without dementia (i.e., geriatric samples, elderly community samples), the IQCODE has been found to be significantly associated with worse cognitive performance in episodic memory/learning, language, attention/working memory, and executive functioning (Jorm, Reference Jorm2004; Jorm, Christensen, Korten, Jacomb, & Henderson, Reference Jorm, Christensen, Korten, Jacomb and Henderson2000; Jorm et al., Reference Jorm, Broe, Creasey, Sulway, Dent, Fairley and Tennant1996), and global cognition as indicated by a brief screening measure (i.e., Mini-Mental Status Exam; Jorm et al., Reference Jorm, Christensen, Korten, Jacomb and Henderson2000; Jorm, Scott, & Jacomb, Reference Jorm, Scott and Jacomb1989). The studies that do exist in PD have only examined the IQCODE as an indicator of subjective cognitive impairment for classifying PD-MCI versus PD-normal cognition (e.g., Pedersen, Larsen, Tysnes, & Alves, Reference Pedersen, Larsen, Tysnes and Alves2017; Pirogovsky-Turk et al., Reference Pirogovsky-Turk, Schiehser, Obtera, Burke, Lessig, Song and Filoteo2014). Thus, research examining whether the IQCODE is associated with a range of objective cognitive abilities in non-demented PD would be beneficial to both clinical research and practice.
In summary, the relationship between informant-based measures of decline and objective cognitive performance in PD is unclear. This is imperative research, as subjective cognitive decline is a risk marker for future, objective cognitive decline and dementia (Amariglio et al., Reference Amariglio, Becker, Carmasin, Wadsworth, Lorius, Sullivan and Rentz2012; Fernandez-Blazquez, Ávila-Villanueva, Maestú, & Medina, Reference Fernandez-Blazquez, Ávila-Villanueva, Maestú and Medina2016), and is an integral part of PD-MCI diagnosis (Litvan et al., Reference Litvan, Goldman, Tröster, Schmand, Weintraub, Petersen and Emre2012). Ascertaining these relationships in those with and without PD-MCI will increase knowledge about the utility of informant reports and improve the clinical identification of individuals who may require further evaluation and monitoring. Therefore, the purpose of this study was to evaluate the relationship between an informant measure of cognitive decline (IQCODE) and objective cognitive performance on a broad range of cognitive tests in non-demented PD patients. Based on previous studies of the IQCODE in non-PD samples (Jorm, Reference Jorm2004) as well as studies indicating the accuracy of informant reports in non-demented PD (Cooper et al., Reference Cooper, Benge, Lantrip and Soileau2017; Naismith et al., Reference Naismith, Pereira, Shine and Lewis2011), it was hypothesized that above and beyond mood, demographic, and disease characteristics, greater levels of informant-reported patient cognitive decline (IQCODE) would be significantly associated with poorer performances in overall cognition, learning, delayed recall, attention, executive function, and language, above and beyond cognitive status.
METHOD
Participants and Procedures
Participants were 139 non-demented PD patients (38 PD-MCI; 101 PD with normal cognition [PD-NC]) and their caregivers/care partners derived from a parent study of cognition in PD conducted at the Veterans Administration San Diego Healthcare System. PD diagnosis was based on the United Kingdom Parkinson’s Disease Society Bank Criteria (Hughes, Ben-Shlomo, Daniel, & Lees, Reference Hughes, Ben-Shlomo, Daniel and Lees1992) and determined by a board-certified neurologist specializing in movement disorders. The Department of Veterans Affairs Institutional Review Board approved the study and all participants provided written consent. Exclusionary criteria included significant medical conditions (e.g., other neurological conditions, secondary causes of PD), prescribed medication with significant anticholinergic properties, or dementia based on the Diagnostic and Statistical Manual of Mental Disorders-IV-TR criteria (American Psychiatric Association, 2000) detailed in Emre et al. (Reference Emre, Aarsland, Brown, Burn, Duyckaerts, Mizuno and Dubois2007). As part of the overarching study, individuals were also excluded if they had a score of <124 on the Mattis Dementia Rating ScaleFootnote 1 (MDRS; Llebaria et al., Reference Llebaria, Pagonabarraga, Kulisevsky, García-Sánchez, Pascual-Sedano, Gironell and Martínez-Corral2008). Lastly, participants were excluded if they did not demonstrate adequate performance validity on neuropsychological testing, as indicated by Forced Choice scores ≤ 14 on the California Verbal Learning Test-II (CVLT-II; Delis, Kramer, Kaplan, & Ober, Reference Delis, Kramer, Kaplan and Ober2000). The original sample contained 142 PD patients, but 3 were excluded due to response bias. All patients were tested on their normal medication dosages (levodopa equivalent dosages; LED in Table 1).
Demographics and clinical characteristics for overall non-demented PD and informant samples
Bold font denotes p < .01.
Abbreviations: FTT = Finger Tapping Test; H&Y = Hoehn and Yahr Scale; LED = levodopa equivalent dosage; GDS = Geriatric Depression Scale; IQCODE = Informant Questionnaire on Cognitive Decline in the Elderly.
All values listed are means (standard deviations) unless otherwise indicated.
p-values were derived from one-way ANOVAs and χ 2 tests.
* Gender data is missing for five informants.
** Levodopa equivalents were calculated using the Parkinson’s measurement online calculator (Turner, Reference Turner2020) formula.
The cut-off score for impairment indicative of dementia on the IQCODE is a summed score of 87.9 (or an average score of 3.38; Jorm, Reference Jorm2004). Twenty-three PD patients (16.6%) exceeded this cut-off score.
Informants were spouses (81.6%), adult children (5.4%), siblings (4.1%), parents (1.4%), and/or friends (7.5%). Age, education, and gender variables were collected from both PD patients and corresponding informants. On average, informants knew PD participants for approximately 38.41 (SD = 15.62) years.
PD participants were classified into PD-MCI and PD-NC groups. PD-MCI was diagnosed based on the International Parkinson and Movement Disorder Society (MDS) task force criteria (Litvan et al., Reference Litvan, Goldman, Tröster, Schmand, Weintraub, Petersen and Emre2012), which requires deficits in one or more cognitive domains (i.e., visuospatial function, language, executive function, attention/working memory, and memory). Specific PD-MCI criteria were modeled after those of Pirogovsky-Turk et al. (Reference Pirogovsky-Turk, Schiehser, Obtera, Burke, Lessig, Song and Filoteo2014)Footnote 2 : ≤–1.33 SD on the Judgment Line of Orientation Test (Benton, Hamsher, Varney, & Spreen, Reference Benton, Hamsher, Varney and Spreen1983) and ≤6 scaled score on the Wechsler Memory Scale (WMS)-III Visual Reproduction Scale – Copy Total (Wechsler, Reference Wechsler1997) for visuospatial function; ≤–1.33 SD on the MDRS – Similarities (Mattis, Reference Mattis1988) and ≤6 on the Delis–Kaplan Executive Function System (D-KEFS) Verbal Fluency – Category Fluency Total Correct (Delis, Kaplan, & Kramer, Reference Delis, Kaplan and Kramer2001) for language; ≤37 standard score on the Wisconsin Card Sorting Test (WCST) – Perseverative Responses (Kongs, Reference Kongs2000) and ≤6 scaled score on the D-KEFS Color-Word Interference Test (CWIT) –Inhibition Condition (Delis, Kaplan, & Kramer, Reference Delis, Kaplan and Kramer2001) for executive functioning; ≤–1.33 SD on the Adaptive Digit Ordering Test (Werheid et al., Reference Werheid, Hoppe, Thöne, Müller, Müngersdorf and von Cramon2002) and ≤6 scaled score on the D-KEFS CWIT – Color Naming Condition (Delis, Kaplan, & Kramer, Reference Delis, Kaplan and Kramer2001) for attention; ≤–1.33 SD on the CVLT-II Long Delay Free Recall (Delis, Kramer, Kaplan, & Ober, Reference Delis, Kramer, Kaplan and Ober2000) and ≤6 scaled score on the WMS-III Logical Memory II Recall Total (Wechsler, Reference Wechsler1997) for memory. Standardized scores were derived from published manual and local normative data. Furthermore, intact independent functional abilities/activities of daily living as determined by informant responses on the medication management and handling finances items from the Lawton’s Instrumental Activities of Daily Living Scale (Lawton & Brody, Reference Lawton and Brody1969) were required (Litvan et al., Reference Litvan, Goldman, Tröster, Schmand, Weintraub, Petersen and Emre2012).
For the purpose of this study, subjective cognitive decline was determined by the following: T-score ≥65 on the Self-Rating Frontal Systems Behavior Scale (FrSBe) – Executive Dysfunction (after illness) subscale (Grace & Molloy, Reference Grace and Molloy2001), self-endorsement of item #10 “Do you feel you have more problems with memory than most?” on the self-report Geriatric Depression Scale (GDS), or self-endorsement of any of the following three screening questions: “Have you noticed changes with remembering things?”; Have you noticed changes with remembering people/names?”; “Have you noticed changes with getting around familiar places?”. If patients did not meet both of the aforementioned criteria (i.e., cognitive dysfunction AND subjective cognitive complaints), then they were classified as PD-NC.Footnote 3
Subjective Measures
Informants completed the IQCODE (Jorm & Jacomb, Reference Jorm and Jacomb1989; Jorm, Scott, & Jacomb, Reference Jorm, Scott and Jacomb1989), which contains 26 items related to the informant’s perception of a patient’s cognitive decline (i.e., changes compared to 10 years ago) in learning, delayed recall, language, attention, and executive functioning that is demonstrated via difficulties in everyday tasks (e.g., items that assess recognition of familiar people, remembering recent events, recalling conversations, adjusting to routine changes, and learning new material). For every statement, informants are required to choose a response on a scale from 1 (“much better”) to 5 (“much worse”). Scores are then summed, ranging from 26 (lowest) to 130 (highest). Higher scores reflect higher informant-rated subjective cognitive decline over a 10-year period.
Both PD participants and their informants completed the GDS (Yesavage et al., Reference Yesavage, Brink, Rose, Lum, Huang, Adey and Leirer1982), which is a 30-item, self-report measure of depressive symptoms experienced within the past week. PD participants and informants completed the GDS to reflect their own symptoms. Higher scores reflect increased depressive symptomatology.
For the purposes of diagnosis, PD patients completed the FrSBe Executive Dysfunction subscale – Self-Rating Form (Grace & Molloy, Reference Grace and Molloy2001) as a measure of subjective cognitive decline. Higher scores reflect more problematic behaviors in the areas of problem-solving, mental flexibility, organization, and planning.
Objective Cognitive Measures
PD participants completed a battery of cognitive tests. Scores from these cognitive tests were standardized into Z-scores from raw means and standard deviations of the overall PD sample. Z-scores were then combined by averaging the cognitive test scores to create the following composites: [1] Language: D-KEFS Verbal Fluency – Category Fluency Total Correct (Delis, Kaplan, & Kramer, Reference Delis, Kaplan and Kramer2001), MDRS – Similarities (Mattis, Reference Mattis1988); [2] Executive Function: WCST – Perseverative Responses (Kongs, Reference Kongs2000), D-KEFS CWIT – Inhibition/Switching Condition (Delis et al., Reference Delis, Kaplan and Kramer2001), D-KEFS CWIT – Inhibition Condition (Delis et al., Reference Delis, Kaplan and Kramer2001); [3] Attention/Working Memory: Adaptive Digit Ordering Test – Total (Werheid et al., Reference Werheid, Hoppe, Thöne, Müller, Müngersdorf and von Cramon2002), CVLT-II – Trial 1 (Delis et al., Reference Delis, Kramer, Kaplan and Ober2000), D-KEFS CWIT – Color Naming Condition (Delis et al., Reference Delis, Kaplan and Kramer2001); [4] Learning: CVLT-II – Trials 1 – 5 Total Score (Delis et al., Reference Delis, Kramer, Kaplan and Ober2000), WMS-III Logical Memory I – Recall Total Score (Wechsler, Reference Wechsler1997), WMS-III Visual Reproduction I – Recall Total Score (Wechsler, Reference Wechsler1997); [5] Delayed Recall: CVLT-II Long Delay Free Recall (Delis et al., Reference Delis, Kramer, Kaplan and Ober2000), WMS-III Logical Memory II – Recall Total Score (Wechsler, Reference Wechsler1997), WMS-III Visual Reproduction II – Recall Total Score (Wechsler, Reference Wechsler1997); and [6] Visuospatial Function: Judgment of Line Orientation Test – Total (Benton et al., Reference Benton, Hamsher, Varney and Spreen1983) and WMS-III Visual Reproduction – Copy Total (Wechsler, Reference Wechsler1997). A global cognition composite was also calculated by taking an average of the six aforementioned composites. Reliabilities were assessed for all cognitive composites. Cronbach’s alphas (α) ranged from .199 to .892 (Language α = .360, Executive Function α = .604, Attention α = .602, Learning α = .700, and Delayed Recall α = .655) and were within satisfactory limits (DeVellis, Reference DeVellis2017; Taber, Reference Taber2018), with the exception of Visuospatial Function (α = .199). Cronbach’s alpha for the global cognition composite was .892, and considered to be “very good” (DeVellis, Reference DeVellis2017).
Motor Measures
Motor functioning was evaluated via the modified Hoehn and Yahr Scale (Goetz et al., Reference Goetz, Poewe, Rascol, Sampaio, Stebbins, Counsell and Wenning2004), with higher scores reflecting worse disease severity, and the bilateral score from the Finger Tapping Test (FTT; Reitan & Wolfson, Reference Reitan and Wolfson1993), with higher scores reflecting better motor function.
Statistical Analyses
IBM SPSS Version 26.0 was used for all statistical analyses (IBM, 2019). Data were screened for normality. Skewness and kurtosis of all cognitive composites fell within acceptable limits (±3.29; Field, Reference Field2009), with the exception of the visuospatial function composite, in which there was one significant outlier. However, the deletion of this outlier did not change the results and thus, this participant was left in the sample. One-way analysis of variance (ANOVAs) and chi-square analysis were conducted to determine group differences between the PD patients and informants. Pearson (for continuous variables) or point-biserial (for categorical variables) correlations were conducted between the cognitive composites and IQCODE total with patient demographic (i.e., patient age, gender, and education) and disease variables (i.e., disease duration in years, LED, and motor function), as well as with patient’s and informant’s own GDS scores. To limit spurious results, p-values < .01 were considered significant.
All demographic variables (i.e., age, education, and gender), mood, and disease severity (i.e., modified Hoehn and Yahr score) were included as covariates in the analyses with cognition based on both theoretical and actual statistical significance with the criterion (Jones et al., Reference Jones, Kurniadi, Kuhn, Szymkowicz, Bunch and Rahmani2019; Maxwell, Delaney, & Kelley, 2017; Siciliano et al., Reference Siciliano, De Micco, Trojano, De Stefano, Baiano, Passaniti and Tessitore2017). Seven hierarchical linear regressions were conducted with these covariates and cognitive status (i.e., PD-NC vs. PD-MCI) entered in block 1, the IQCODE total score entered in block 2, and then each cognitive composite (i.e., global cognition, language, executive function, attention, learning, delayed recall, and visuospatial function) was entered as the criterion. Covariates were removed by backward elimination if they were not significantly associated with the criterion (i.e., p < .05); final models are reported in Table 3. P-values < .01 were considered significant for all analyses. Effect sizes were interpreted as the following: .10 for small, .30 for medium, and .50 and above as large (Cohen, Reference Cohen1988).
Bivariate Pearson correlations between the cognitive outcomes, IQCODE, and clinical characteristics in overall PD patient (n = 139) sample
*Note. All values are r (p-value).
Bold font denotes significant association at p < .01.
Models with IQCODE predicting objective cognition
Bold font denotes predictor variable p < .01.
Abbreviations: β = standardized beta; Patient GDS = Patient-reported Geriatric Depression Scale; Informant GDS = Informant Geriatric Depression Scale; IQCODE = Informant Questionnaire of Cognitive Decline in the Elderly.
RESULTS
Participants’ characteristics for the overall PD (n = 139) and informant (n = 139) samples are displayed in Table 1. Fifty of the PD patients were veterans (36% of the entire sample). PD participants and informants were equivalent in terms of education, but the PD participants were significantly older and had a greater proportion of men compared to the informants.
As detailed in Table 2, for the total sample, patient age, education, and gender were significantly correlated with all cognitive domains, except for age with visuospatial function, gender with attention and visuospatial function, and education with language, attention, learning, delayed recall, and visuospatial function. Hoehn and Yahr scores, FTT scores, and GDS scores (both patient and informant) were not significantly associated with any cognitive domains. Higher IQCODE scores were significantly associated with worse depressive symptoms for informants (i.e., higher GDS scores). The IQCODE was not associated with cognitive status (r = .053, p = .536).
Bivariate correlations revealed that higher IQCODE score totals were significantly associated with lower Global Cognition (r = −.291, p = .001), Executive Function (r = −.235, p = .006), Learning (r = −.251, p = .003), Delayed Recall (r = −.247, p = .003), and Attention (r = −.310 p < .001) composite scores. The IQCODE was not significantly associated with Language (r = −.129, p = .131) or Visuospatial Function (r = −.093, p = .281).
As detailed in Table 3, when controlling for significant covariates in addition to cognitive status (i.e., PD-NC vs. PD-MCI), higher IQCODE total scores (i.e., greater levels of informant-rated patient subjective cognitive decline) were significantly associated with worse global cognition, executive function, learning, delayed recall, and attention. However, the IQCODE did not significantly predict language or visuospatial function, and the size of these effects was small. For all regressions, tolerance levels and variance inflation factors were also within acceptable limits (tolerance > .10 and variance inflation factors < 10.0; Fields, Reference Field2009).
DISCUSSION
The current study evaluated the relationship between informant-based measures of cognitive decline (as measured by the IQCODE) and cognitive functioning in non-demented patients with PD. The results supported our hypothesis that informant responses of cognitive decline were associated with PD patients’ objective performance on measures of global cognition, learning, delayed recall, executive function, and attention, above and beyond patient and informant mood, patient demographic factors, and disease severity in the overall PD sample. Consistent with the IQCODE’s purported broad measurement of cognitive abilities, our results indicate that informant-based responses of patient cognitive decline are sensitive to most areas of objective cognitive performance in a broad range of domains in non-demented PD patients. The current study also extended upon previous studies by demonstrating that these results held even when controlling for cognitive status (i.e., PD-MCI and PD-NC). Therefore, these results suggest that regardless of non-demented cognitive status (i.e., PD-MCI or normal cognition), informant responses on the IQCODE yield equally valuable clinical information about objective performance. This information could be particularly useful in certain circumstances, such as with patients who may not be able to provide self-ratings or when comprehensive neuropsychological assessment is not feasible or readily available. These results also support the use of such reports to improve diagnostic accuracy and provide early intervention.
Given that subjective reporting of cognitive decline is a harbinger for future objectively measured cognitive decline in PD (Hong et al., Reference Hong, Sunwoo, Chung, Ham, Lee, Sohn and Lee2014), our results underscore the importance of querying PD patients’ informants about the cognitive changes they observe in their partner. As informants typically assist PD patients with activities of daily living, provide emotional and social support, and offer advice on medical decisions (Goldman et al., Reference Goldman, Vernaleo, Camicioli, Dahodwala, Dobkin, Ellis and Simmonds2018), our results support the likelihood that informants possess critical knowledge to help distinguish between normal aging in PD versus concerning cognitive changes. Thus, if an informant observes decrements in activities such as following a story (i.e., attention), remembering recent events (i.e., delayed recall), learning new material (i.e., learning), adjusting to changes in routine (i.e., executive function), this may provide insight into a PD patient’s declining cognitive abilities.
Despite the association of the IQCODE with a broad range of patient cognitive abilities, informant-based measures were not significantly associated with objective language performance. Language is thought to involve access to mental lexicon ability, which involves the retrieval of grammatical representations and sound forms of words (Shao, Janse, Visser, & Meyer, Reference Shao, Janse, Visser and Meyer2014). Although the IQCODE may measure some aspects of language, informant-based responses on this measure do not appear to be sensitive to the objective semantic tasks that comprised the language composite in our current study. Future studies that utilize measures of comprehension may be more sensitive to phonemic and semantic domains and yield significant results. This contrasts with prior IQCODE studies in geriatric populations (with varying health conditions and associated MCI) that have found these overall informant-based reports to be related to language and semantic processing (Jorm, Reference Jorm2004). Given that PD (e.g., PD-MCI) patients may have higher verbal comprehension than those with MCI due to other etiologies (Pistacchi, Gioulis, Contin, Sanson, & Marsala, Reference Pistacchi, Gioulis, Contin, Sanson and Marsala2015), this may at least partially explain the lack of relationship with informant subjective measures.
Not unexpectedly, informant reports of cognitive decline also did not correlate with visuospatial functioning. The IQCODE does not purport to measure this ability, and thus, it may be limited in this regard. Considering that the IQCODE was developed for geriatric patients with memory problems, and that the majority of items appear to measure memory skills, the IQCODE itself may not be sensitive to language and visuospatial function in non-demented PD. These findings caution against reliance on the IQCODE for gauging these abilities, particularly in cases of diagnostic assessment in PD (e.g., PD-MCI criteria) when language or visuospatial abilities are a concern. Alternatively, it is possible that neuropsychological assessments might detect subtle changes in cognition (i.e., language) that are not yet obvious to informants. As such, it may be that informants have difficulty characterizing cognition related to these areas. It may become easier to subjectively characterize these cognitive impairments as the disease progresses (i.e., PD-dementia). Therefore, the use of alternative measures for these domains is recommended to supplement the IQCODE in the assessment of cognitive decline in non-demented PD. Nevertheless, our overall findings support the general utility of the IQCODE in detecting a wide range of neurocognitive difficulties in PD, which could help identify those who may benefit from further evaluation and comprehensive assessment.
It is important to note that the IQCODE was not significantly associated with patient demographics and disease characteristics in the overall non-demented PD sample. These results are consistent with previous literature, which has demonstrated either weak or nonsignificant relationships between the IQCODE and socioeconomic status, occupational status, and gender in samples of elderly individuals with varying health conditions (Jorm & Korten, Reference Jorm and Korten1988; Jorm, Reference Jorm2004) including dementia (Jorm, Reference Jorm2004; Jorm et al., Reference Jorm, Scott and Jacomb1989) and support adequate discriminant validity of the IQCODE. Depression scores, as measured by the informant and patient GDS, were significantly correlated with the IQCODE. This is consistent with previous findings that higher IQCODE scores are related to higher informant depressive symptoms in older adults across a wide variety of health conditions (Jorm, Reference Jorm2004). While informant reporting can be impacted by their own mood as well as patients’ mood (Jiménez-Huete et al., Reference Jiménez-Huete, Del Barrio, Riva, Campo, Toledano and Franch2017; Jorm, Reference Jorm2004), our results clearly indicated that informant-based measures of cognitive decline related to objective cognitive performance above and beyond the influence of depressive symptoms. These findings support the overall accuracy of informant-based measures of cognitive decrements in their PD partners regardless of their partner’s or their own mood symptoms.
Limitations of this study include a PD sample that was mostly male, of Caucasian/White descent, and highly educated, as well as informants that were mostly female, of Caucasian/White descent, and similarly highly educated. Although the gender breakdown of the PD sample is consistent with the epidemiological studies of PD (Willis, Evanoff, Lian, Criswell, & Racette, Reference Willis, Evanoff, Lian, Criswell and Racette2010), these results may not generalize to non-demented female PD patients or those from more diverse cultural backgrounds. Likewise, our informants were mostly female spouses, which could limit the generalizability to male informants and informants who are not spouses. Furthermore, our informants had known their PD partners for much longer than 10 years, which could limit the generalizability of these results to informants with more limited interactions or knowledge regarding their partners. As such, future research on diverse samples with a broad range of educational, ethnic, and cultural backgrounds is warranted. Also, this study only examined the relationship between the IQCODE and objective cognition in non-demented PD samples (i.e., those who were diagnosed with PD-MCI or PD-NC); thus, findings cannot be generalized to those with PD dementia. Future research examining the longitudinal relationship between the IQCODE and objective measures of cognition in non-demented PD is warranted. Lastly, future research should explore cut-off scores in order to increase the clinical application of the IQCODE in PD.
In summary, this is the first study to assess the relationship between the IQCODE and objective concurrent cognitive performance in non-demented PD. Findings provide evidence of concordance between informant-based measures of patient cognitive decline and objective patient performance on measures of overall cognition, executive function, learning, delayed recall, and attention in non-demented PD regardless of patient cognitive status (i.e., PD-MCI and PD-NC). These results provide support for the utility of the IQCODE in assessing concurrent cognitive functioning in non-demented PD on a wide array of cognitive domains. The IQCODE could also aid in the identification of patients who may need comprehensive objective cognitive testing, monitoring, and treatment.
ACKNOWLEDGMENTS
This project was supported by VA RR&D Merit Award to Dawn M. Schiehser, Ph.D. [RX001691-02] and VA CSR&D Merit Award to J. Vincent Filoteo, Ph.D. [CX000813], by the Department of Veterans Affairs, Veteran Health Administration. Additional thanks to the University of California, San Diego’s Strategic Enhancement of Excellence through Diversity (SEED) Fellowship to Marina Zaher Nakhla. Kelsey Holiday received salary support from the National Institute of Neurological Disorders and Stroke of the National Institutes of Health [F31NS108573].
We would like to thank Celina Pluim, Nicole Whiteley, and Angelie Cabrera Tuazon for their assistance with data collection and data entry. We also appreciate all participants for their contributions to this study.
CONFLICT OF INTEREST
The authors have nothing to disclose.
