Abstract
Background:
Dementia is a risk factor for undernutrition. However, it is not clear if specific cognitive deficits have a higher risk of undernutrition and how much appetite/nutritional problems and caregiver stress mediate this association.
Objective:
To evaluate the relationship between nutritional status and severity of global and function-specific scores of cognitive dysfunctions, and to which extent this association is mediated by appetite/nutritional problems and caregiver stress.
Methods:
Cross-sectional analysis of the ReGAl study data, including 761 older adults attending a Memory Clinic. Nutritional status was evaluated with Mini Nutritional Assessment (MNA). The relationship between scores at neuro-cognitive tests and risk of undernutrition was evaluated using logistic regression models adjusted for potential confounders. To allow comparison between different tests, all scores were standardized. Mediation analysis was used to evaluate how much appetite/nutritional problems and caregiver stress mediate this association.
Results:
Mean age was 77 years (SD: 9), 37.3% were women. Exploring different cognitive domains, a stronger association was documented for attentive matrices (OR:0.49, 95% CI: 0.34–0.72), the figure copy test (OR:0.63, 95% CI: 0.45–0.88), and the verbal judgement test (OR:0.61, 95% CI: 0.42–0.91). The proportion of the effect of cognition (MMSE) on nutritional status mediated by caregiver distress was 9.5% (95% CI: 0.002–0.27), the proportion mediated by appetite/nutritional problems was 11% (95% CI: –4.8–3.18).
Conclusion:
Risk of undernutrition is associated to cognitive decline; a stronger association was observed for attention, praxis, and reasoning. Caregiver distress is a mediator of this association. This information should be considered in the management plans of this population.
INTRODUCTION
The number of people currently living with dementia in Europe is almost 9.8 million and almost one in every 20 people over the age of 65 has Alzheimer’s disease (AD), the most common form of dementia [1, 2].
Dementia is a risk factor for undernutrition [3], and unintentional weight loss may be considered a hallmark of AD even at its early stages, leading to severe complications such as muscular atrophy and loss of independence [4, 5].
It is unknown whether the nutritional risk in AD is due to higher energy expenditure, reduced dietary intake, or a combination of both, and what is the influence of other factors (e.g., interaction with caregivers). Dietary intake is partly driven by olfactory and gustatory functioning and food preferences [6], and subjects with mild cognitive impairment or AD have difficulties in discriminating and identifying odors. It has been reported that this sensory impairment is directly associated to dementia severity [7–9]. Nonetheless, current available literature shows no differences between AD and cognitively healthy subjects with respect to energy intake or energy expenditure [10]; actually, food choice and dietary pattern of people with and without dementia seem comparable [11].
Other factors may play a key role in the nutritional status among people with dementia. Feeding difficulties and changes in eating and dietary habits may be secondary to apraxia, as cortical function decline decreases patient’s ability to obtain adequate nutrition, and malnutrition risk seems to be different according to different types of dementia [12], suggesting that specific cognitive domains impairment, implicated in feeding, might be associated with a higher risk of undernutrition. Very few studies are available on this topic, documenting that executive disfunction is associated with an impairment of oral intake [13], and that attention, language and visuospatial ability are associated with a higher risk of undernutrition in patients with dementia [14].
Among nursing home residents, the importance of staff-resident (dyadic) interactions during mealtime is acknowledged [15–18] and is considered an important modifiable factor to manage mealtime difficulties and promote intake; in particular, food intake seems to be promoted by dyadic verbal interactions [17]. Also, in the community setting, poor nutritional status may be the result of insufficient care due to a heavy burden for family caregivers who may be stressed, depressed, and socially isolated [19]. Actually, the risk of undernutrition of people with dementia, evaluated using the Mini-Nutritional Assessment (MNA), is associated with the MNA of their caregiver [20]; furthermore, the nutritional status of older people with dementia is associated with burden of family caregivers and can be worsened by physical and psychological difficulties experienced by the caregivers themselves [21, 22]. However, these studies evaluated the association between nutrition and caregiver burden among older adults with dementia, but there are no studies that analyze the role of caregiver distress as a mediator of the relationship between cognitive impairment and nutritional status.
Given this body of evidence, the risk of undernutrition may be associated with specific cognitive deficits; in particular, we hypothesized that an impairment of memory, attention, visual attention, executive and praxic abilities may affect the ability of the patient conceive and realize the idea of eating. Furthermore, we also hypothesized that this relationship may be affected by appetite/nutritional problems and caregiver stress. The aim of this study was to evaluate the relationship between nutritional status and severity of global and function-specific scores of cognitive dysfunctions, and to evaluate to which extent the effect of cognitive dysfunction on nutritional status is mediated by appetite/nutritional problems and caregiver stress.
METHODS
Study population
We used data from the Italian Geriatric Network on Alzheimer’s disease (Rete Geriatrica Alzheimer - ReGAl) project, a longitudinal multicentric study promoted by the Italian Society of Gerontology and Geriatrics (Società Italiana di Gerontologia e Geriatria, SIGG). The project has been described extensively elsewhere [23]. Briefly, it included all the older adults attending Memory Clinics in multiple centers in Italy, and recorded data routinely collected during the clinical workup, including demographic characteristics, medical history, pharmacological treatments, and clinical and neurological examination and a neuropsychological evaluation. Subjects with clinically severe psychiatric or systemic disease, intellectual disability, and severe sensory impairment (blindness, deafness) were excluded. The diagnosis of dementia was based on DSM-IV [24] and on Petersen’s criteria [25] for mild cognitive impairment. AD was diagnosed according to the NINDS-ADRDA criteria [26], vascular dementia on the NINDS-AIREN criteria [27], frontotemporal dementia on those described by Neary et al. [28], dementia with Lewy bodies on those by McKeith et al. [29], and Parkinson’s disease dementia according to Emre et al. [30]. The severity of dementia was rated by the Clinical Dementia Rating scale (CDR) [31]. For the present study, we included only patients attending the Memory Clinic of University of Perugia due to several missing of variables of interest in the other centers, and we only used baseline data. All the patients signed a written informed consent to participate to the study. The study was approved by the Ethic Committee of the University of Perugia (n. 2773/16).
Cognitive status
From the tests included in the database, for the present study we took into account the Mini-Mental State Examination (MMSE) [32], as test of general cognition, the immediate and delayed recall of the Rey’s Auditory Verbal Learning Test, as a test for episodic memory, the backward digit-span test and attentive matrices for attention, the Clock Drawing Test for executive abilities, the figure copy test for praxis, the Corsi block-tapping test for visuo-spatial working memory, the Token test for language, and the verbal judgment test for reasoning. For each test, details on administration procedures and Italian normative data for score adjustment for age and education as well as normality cut-off scores (95% of the lower tolerance limit of the normal population distribution) are available [33–35].
Current depressive symptoms were assessed by the 15-item version of the Geriatric Depression Scale [36]. Appetite and nutritional problems were definite on the basis of the relevant item of the Neuro-psychiatric Inventory (NPI) Questionnaire. Also caregiver stress was evaluated using the NPI [37].
Nutritional status
Nutritional status was assessed using the Mini Nutritional Assessment (MNA), that has been extensively validated in older people, including those with cognitive impairment [38]. The maximum score of the MNA is 30, scores between 17 and 23.5 points indicate nutritional risk and scores below 17 indicate overt undernutrition.
Statistical analysis
The sample was divided in four groups based on their CDR score: no dementia (0 points), very mild impairment (0.5 points), mild impairment (1 point) and moderate/severe impairment (≥2 points) [31]. The distribution of all continuous variables was visually inspected using histograms and quantile-quantile plots to check for substantial deviation from the normal distribution. Given the small number of patients with overt undernutrition, we analyzed them in the group of patients at risk of undernutrition. Group characteristics were compared using descriptive statistics (mean and standard deviations for numerical variables, proportion for categorical variables). We explored the distribution of MNA by dementia subtype and by severity of dementia using bar plots. Comparison between groups was performed using the chi-square test.
To study the association between cognitive performance test scores and risk of undernutrition, we excluded patients without dementia at the CDR. This association was evaluated using logistic regression models, both unadjusted and adjusted for age, sex, education, presence of caregiver, reported changes in appetite/eating, and caregiver stress. To allow comparison between different tests, all scores were standardized by centering (subtracting the mean) and rescaling (dividing by SD), thus the ORs express the changes in the dependent variable for a 1 SD change in the independent variable. To evaluate to which extent the effect of cognitive function on nutrition is mediated by appetite and nutritional problems or caregiver stress we used mediation analysis, that allows to break down the effect of an independent variable on a dependent variable into a “average direct effect” and an “average causal mediation effect”, that is the amount of effect that is mediated by a third variable considered in the model [39].
Analyses were performed using R version 3.5.0 (R Foundation for Statistical Computing, Vienna, Austria).
RESULTS
We included 761 participants with a mean age of 77 years (SD: 9); 37.3% were women and 95.3% have a study partner. Twenty-two percent of the participants was affected by mild cognitive impairment, 18% by AD dementia, 14% by mixed dementia, 12% by depression with cognitive impairment (Fig. 1). CDR classified as without dementia 244 patients, 210 as very mild impaired, 199 as mild impaired, and 92 as moderate or severe impaired. The mean MNA score in the four groups was 25.7 (SD: 2.0), 24.2 (SD: 2.5), 23.6 (SD: 2.2), and 21.9 (SD: 2.9), respectively (Table 1). Other characteristics of the groups are reported in Table 1.
Distribution of dementia subtypes.
General characteristics of the sample according to cognitive function
ADL, activities of daily living; GDS, Geriatric Depression Scale; IADL, instrumental activities of daily living.
Less than 10% of participants without cognitive impairment were at risk of undernutrition, while it arises up to 24% for mild cognitive impairment, to 44% for AD dementia, and 50% for mixed dementia (p < 0.001) (Fig. 2). Considering dementia severity, the proportion of patients with undernutrition increases with the severity of dementia, from <20% in participants without dementia to >60% in those with moderate-severe dementia (Fig. 3).
Distribution of risk of malnutrition according to dementia subtypes.
Distribution of risk of malnutrition according to dementia severity.
MMSE score was inversely associated with risk of undernutrition (standardized OR 0.64, 95% CI 0.53–0.77), as also Clock Drawing Test (standardized OR 0.68, 95% CI 0.56–0.82), backward digit span (standardized OR 0.74, 95% CI 0.6–0.91), Auditory verbal learning test (AVLT) delayed (standardized OR 0.76, 95% CI 0.6–0.96), attentive matrices (standardized OR 0.54, 95% CI 0.43–0.68), figure copy test (standardized OR 0.64, 95% CI 0.52–0.79), Token test (standardized OR 0.94, 95% CI 0.91–0.98), and verbal judgement test (standardized OR 0.98, 95% CI 0.96–0.99). Adjustment for potential confounders, however, reduced the estimated effect of cognitive function on nutritional status, except for attentive matrices (standardized OR 0.49, 95% CI 0.34–0.72), figure copy test (standardized OR 0.63, 95% CI 0.45–0.88), and verbal judgement test (standardized OR 0.61, 95% CI 0.42–0.91) (Table 3).
Mediation analysis showed that the proportion of the effect of cognition (expressed by the MMSE) on nutritional status mediated by caregiver distress was 9.5% (95% CI: 0.002, 0.27), independently of appetite/nutritional problems. On the other hand, the proportion of the effect mediated by appetite/nutritional problems, independently of caregiver distress, was 11% (95% CI: –4.8, 3.18).
DISCUSSION
Our data confirm an association between cognitive impairment and nutritional risk that is independent of age, sex, and other potential confounders. Among specific cognitive domains considered, attention, praxis, and reasoning seem to have the greatest potential impact on nutritional status, and their effect is partly mediated by caregiver distress. Among the very few available studies evaluating the association between nutritional status and cognitive performance in different domains, Doorduijn et al. showed that a poor performance in the domains of attention, language and visuospatial ability and executive functioning was associated with lower modified MNA scores [14].
These partially discording results may be explained by different sample characteristics. Indeed, Doorduijn et al enrolled patients from no cognitive impairment to dementia, younger, dosed AD biomarkers, and used a different set of neuro-psychological tests. To the best of our knowledge, no other studies are available on this topic on patients with dementia.
It is conceivable that nutritional problems arise in different disease stages as a consequence of pathophysiological changes, dementia-specific nutritional problems, and the continuous decline of competences contributes to the reduced dietary intake and progression of the disease itself as a vicious circle [40].
Association between cognitive tests score and risk of undernutrition
+Variables are scaled so that coefficients express changes in MNA by changes in cognitive function tests using SD as unit. *Adjusted for age, sex, education, caregiver, nutritional/appetite problems, caregiver distress score.
While the proportion of undernutrition risk was high in our sample of people with cognitive impairment, we found few patients with overt undernutrition. This is in contrast to some studies available in the recent literature [3, 42], probably because we took in consideration only ambulatory patients, and with a relatively young age (mean age 77 years): there is strong evidence that the prevalence of undernutrition varies across settings [43, 44], being more prevalent among nursing home residents and in acute care hospital or rehabilitation settings [45–48].
We have found that a sizable part (about 10%) of the effect of cognitive impairment on nutritional status is mediated by caregiver distress. To our knowledge, this is the first study reporting such a relationship through a mediation analysis rather than as a pure association [21, 22]. Within the dementia literature several studies examined the experience of being a caregiver of people with dementia, but less research explored the effect of caregiver burden on nutritional status. There is growing evidence that stress increases the vulnerability of chronic dementia caregiver to disease and diminishes their ability to provide optimal care [49]. Indeed, caregiver burnout is associated with poor outcomes for the care recipient, including early institutionalization, depression, poor quality of life, and mortality [50, 51]. The importance of this finding lies in the fact that caregiver distress is potentially modifiable, and this is relevant for intervention aimed at the prevention and care of undernutrition in people with dementia, that should include the evaluation and management of caregiver distress.
This study has some limitations. First, it is a mono-centric study and therefore our results may not be generalizable to other population, although the general characteristics of our sample are similar to that of most samples of older people. Second, our results are relevant only with respect to undernutrition risk, as we found only a few patients with overt undernutrition: further studies are needed to confirm our findings also with respect to this outcome. Third, the mediation analysis could be affected by the cross-sectional design of the study. Finally, given the design and setting of the study, a power analysis was not conducted.
In conclusion, we showed that the risk of undernutrition is clearly associated to cognitive decline and caregiver distress is one of the mediators of undernutrition in people with dementia. This information should be considered in the management plans of people with dementia, especially those with more severe impairment.
DISCLOSURE STATEMENT
Authors’ disclosures available online (https://www.j-alz.com/manuscript-disclosures/21-5732r2).