Introduction

Dementia is a progressive clinical syndrome characterized by the impairment of multiple cognitive functions such as short- and long-term memory loss, thinking, judging, orientation, calculation, and learning ability as well as personality changes, impaired abilities, and motor skills. Patients with dementia experience emotional, social, and behavioral problems as well as cognitive problems. Dementia also negatively affects the activities of daily living (World Health Organization, 2018).

Over 55 million people live with dementia worldwide, and this number is expected to reach 78 million by 2030 (Gauthier et al., 2021). The prevalence of dementia in Turkey has been reported as comparable with that seen in developed countries (Gurvit et al., 2008). In parallel with the world, this number is expected to increase with the aging of the population in Turkey (The World Bank, 2019).

Dementia is a condition that causes many problems for both patients and their relatives/loved ones and caregivers. These patients may experience behavioral problems such as apathy, aberrant motor behavior, sleep deprivation, agitation, and delusions (Küçükgüçlü et al., 2017; Liu et al., 2017). Eating problems are one of the most common behavioral problems in patients with dementia. Factors such as impaired cognitive and motor skills, depression, environmental factors, social interactions, food culture, visual impairment, dental problems, and poor oral hygiene may cause eating difficulties in these patients (Cipriani et al., 2016; K. M. Lee & Song, 2015; Liu et al., 2020). Moreover, eating problems may increase the risk of malnutrition (Chang & Roberts, 2011). It is known that malnutrition adversely affects physical health, reduces quality of life, and increases morbidity and mortality (Saucedo Figueredo et al., 2016). However, few studies exploring the relationship between eating difficulties and nutritional status have been conducted on patients with dementia. Therefore, the aim of this study was to determine eating difficulties experienced by these patients and to evaluate the relationship between eating difficulties and nutritional status.

Methods Design and Sample

An observational, cross-sectional case–control study was conducted in a dementia outpatient clinic of a university hospital in Istanbul between February 2015 and August 2016. Patients with a confirmed diagnosis of dementia were included if they were still living at home and if they were able to stand independently on a scale to be weighed. Patients were excluded if they had any serious illnesses (e.g., cancer, stroke with sequelae, decompensated heart failure). The control group consisted of individuals not affected by dementia and without significant health problems (e.g., cancer, diabetes, myocardial infarction, stroke, heart failure). Most of the control group participants were the relatives of the patients from another clinic (internal medicine), whereas several were personal acquaintances of the authors. The sample size was calculated (α = 5% and power = 80%, Cohen effect size = 0.3) with a minimum of 44 in each group for a total of 88. Fifty participants were recruited for each group to account for possible attrition. Fifty patients who were approached declined to participate because their relatives did not agree (n = 25) or they could not stand on the scale (n = 25). The evaluations were completed with 50 patients. Fifty individuals with a similar mean age, gender ratio, and educational level as the patient group constituted the control group. The compliance of the study with ethical principles was evaluated and approved by the Clinical Research Ethics Committee of Istanbul Faculty of Medicine (15.01.2015/81). In addition, written informed consent was obtained from all of the participants or their relatives.

Data Collection

Patients and their relatives were approached while they were waiting for their appointment. All assessments (except for Mini Mental State Examination [MMSE]) were made by the first author with the individuals who met the inclusion criteria using a face-to-face interview method. Eating difficulties were evaluated using a questionnaire derived from the Minimal Eating Observation Form (MEOF), and nutritional status was evaluated using the Mini Nutritional Assessment (MNA). A patient information form, MMSE, Barthel Index of Activities of Daily Living (BI), Functional Activities Questionnaire (FAQ), and Charlson Comorbidity Index were also used to collect data. The MMSE was performed by the neurologist who examined the patient, and these data were obtained from the patient file at the end of the day.

Eating difficulties questionnaire

Eating difficulties were evaluated using a questionnaire based on questions from the MEOF (Westergren, 2019). The questionnaire consisted of problems related to eating activity (i.e., problems in swallowing, chewing, teeth, manipulating food on a plate, and using utensils; the need for assistive tools, inability to eat without help, loss of food awareness in mouth, and refusal of food). This form has a three-factor structure: deglutition, ingestion, and energy/appetite. The Cronbach's alpha value has been reported as .76, and interobserver agreement has been reported to be good (kappa value = .70; Westergren, 2019; Westergren et al., 2009).

Mini nutritional assessment

The MNA is an assessment developed to assess nutritional status that addresses nutritional habits/problems, some anthropometric measurements, and a physical assessment. This assessment is recommended by the European Society for Clinical Nutrition and Metabolism for use with older adults. In terms of scoring, a score of 24–30 indicates normal nutritional status, 17–23.5 indicates risk of malnutrition, and less than 17 indicates malnutrition (Rubenstein et al., 2001; Sarikaya et al., 2015).

Mini mental state examination

The MMSE was developed to evaluate five basic cognitive functions (orientation, memory, attention, calculation, and language). The maximum score that can be obtained from this test is 30, with lower scores indicating impairment in mental functions (Folstein et al., 1975; Gungen et al., 2002). The Turkish version of the MMSE has been validated in the Turkish cultural context and has been shown to have high discriminant validity and interrater reliability in the diagnosis of dementia. A cutoff score of 23–24 was found to have the highest sensitivity (0.91) and specificity (0.95). Interrater reliability analysis has shown high correlation (r = .99) and kappa value (.92; Gungen et al., 2002). A recent study confirmed these psychometric findings and also reported good internal consistency (Cronbach's α of .86; Çebi et al., 2020). The MMSE was assessed by the neurologist who examined the patient, and these data were obtained from the patient file at the end of the day.

Barthel index of activities of daily living

This index, developed to assess dependence in activities of daily living, is widely recommended as a measure for activities of daily living in older populations. Scoring ranges from 0 to 100, with 0 indicating complete dependence and 100 indicating complete independence. The cutoff value for independence is 60, with scores greater than 60 indicating the ability to function independently (Hopman-Rock et al., 2019). This index was validated in Turkish by Küçükdeveci et al. (2000). Internal consistency was good at .93 for stroke and .88 for spinal cord injury. The level of agreement between the two raters was sufficient with kappa levels greater than .5. The intraclass correlation coefficients were .99 and .77 for stroke and spinal cord injury, respectively. This index is widely used to survey older adult populations.

Functional activities questionnaire

The FAQ is a short questionnaire used to assess the performance of complex daily living activities. This questionnaire is usually conducted on the family caregivers of patients, with higher scores indicating greater dependence with regard to instrumental daily living activities (Pfeffer et al., 1982; Selekler et al., 2004). This scale was developed by Pfeffer et al., and a norm determination study on a Turkish sample in which participants were all 50 years or older was conducted by Selekler et al. (2004). In that study, the average and standard deviations of the participant scores obtained from the FAQ were determined based on age, gender, and educational level.

Statistical Analysis

Frequency and mean were used to describe the characteristics of the participants. The Kolmogorov–Smirnov test was used to evaluate the suitability of the data for normal distribution. Mann–Whitney U and chi-square tests (and the Fisher's exact test) were used to compare patient and control groups. Chi-square was used to evaluate the relationship between eating problems and nutritional status in the dementia group. Eating difficulties were evaluated using an independent samples t test for age, duration of illness, Barthel Index, and FAQ and using the Mann–Whitney U test for the Charlson Comorbidity Index, Clinical Dementia Rating, and MMSE. All of the analyses were conducted using IBM SPSS Statistics Version 21.0 (IBM Inc., Armonk, NY, USA).

Results

One hundred participants were enrolled in this study, including 50 patients and 50 patients in the control group. The sociodemographic characteristics of the participants in the patient and control groups (Table 1) show that the groups were similar in terms of sociodemographic characteristics such as age, gender, education, and income (p > .05).

Table 1. - Sociodemographic Characteristics of Patients and Controls Characteristic Dementia (n = 50) Control (n = 50) χ2/Z p n % n % Age (years; M ± SD and range) 71.8 ± 10.7 49–90 71.8 ± 10.8 48–90 −0.03 .978 Gender 0.16 .689  Male 27 54 25 50  Female 23 46 25 50 Education 0.00 1.000  < High school 33 66 33 66  ≥ High school 17 34 17 34 Perceived income 1.07 .300  High/moderate 34 68 29 58  Low 16 32 21 42 Marital status 0.00 1.000  Married 43 86 43 86  Divorced/widowed 7 14 7 14

Note. Mann–Whitney U test was used for age; chi-square test was used for other comparisons.

The mean duration of diagnosis was 32.3 (SD = 21.8) months. Most (84%) of the patients had mild dementia of which Alzheimer's was the most common (64%). The patients scored more poorly than their control peers on the MMSE (29.70 ± 0.58 vs. 19.48 ± 2.88, p < .001), BI (99.40 ± 1.64 vs. 67.00 ± 23.69, p < .001), FAQ (1.60 ± 2.48 vs. 27.00 ± 2.92, p < .001), and MNA (26.51 ± 2.46 vs. 19.80 ± 3.98, p < .001; Table 2).

Table 2. - Clinical Characteristics of Patients and Controls Characteristic Dementia (n = 50) Control (n = 50) χ2/Z p n % n % Duration of illness (months; M ± SD and range) 32.30 ± 21.80 0–75 Type of dementia  Alzheimer's 32 64  Vasculary 12 24  Lewy body 2 4  Frontotemporal 4 8 Clinical Dementia Rating  Very mild 2 4  Mild 42 84  Moderate 6 12 Charlson Comorbidity Index (M ± SD and range) 0.80 ± 0.90 1–4 0.10 ± 0.30 0–1 −5.27 < .001 No. of medications (M ± SD and range) 4.40 ± 1.50 2–10 3.60 ± 0.90 2–5 −1.91 .056 MMSE (M ± SD and range) 19.48 ± 2.88 13–25 29.70 ± 0.58 28–30 −8.89 < .001 Barthel Index: total (M ± SD and range) 67 ± 23.69 15–100 99.40 ± 1.64 95–100 −8.37 < .001 Barthel: feeding  Independent 33 66 50 100 18.14 < .001  With assistance/dependent 17 34 – – FAQ (M ± SD and range) 27.00 ± 2.92 18–30 1.60 ± 2.48 0–7 −8.80 < .001 MNA (M ± SD and range) 19.80 ± 3.98 8.5–25.5 26.51 ± 2.46 15.5–30.0 −7.79 < .001

Note. Chi-square test was used for Barthel “feeding”; Mann–Whitney U test was used for other comparisons. MMSE = Mini Mental State Examination; FAQ = Functional Activities Questionnaire; MNA = Mini Nutritional Assessment.

All of the patients in this study were fed orally. The eating difficulties of the groups are described in Table 3. Although no between-group differences were found in terms of swallowing problems, chewing problems, tooth loss, or food awareness in the mouth, patients with dementia reported greater difficulties in terms of self-feeding skills. Whereas problems related to manipulating food on the plate and utensil use were not seen in the control group, these problems were found in 30% of the dementia group (p < .001). Thirty percent of the patient group could not eat without assistance, whereas no one in the control group required assistance to eat (p < .001). In addition, whereas 56% of the patients reported refusal to eat, this ratio was 2% in the control group (p < .001; Table 3).

Table 3. - Comparison of Eating Difficulties Between Patients and Controls Problems in… Dementia
(n = 50) Control
(n = 50) χ2 p n % n % Swallowing a – .242  Yes 3 6 – –  No 47 94 50 100 Chewing a – .495  Yes 2 4 – –  No 48 96 50 100 Tooth loss 0.64 .422  No 21 42 25 50  Yes 29 58 25 50 Manipulating food on plate b 15.37 < .001  Yes 15 30 – –  No 35 70 50 100 Using utensils b 15.37 < .001  Yes 15 30 – –  No 35 70 50 100 Need for assistive tools b 11.46 .001  Yes 12 24 – –  No 38 76 50 100 Ability to eat without help b 15.37 < .001  Yes 35 70 50 100  No 15 30 – – Food awareness in mouth a – 1.000  Yes 49 98 50 100  No 1 2 – – Refusal of food b 32.83 < .001  Yes 28 56 1 2  No 22 44 49 98

a Fisher's exact test. b Continuity correction.

Next, eating difficulties in the patient group were evaluated to identify possible relationships with sociodemographic and clinical characteristics. Patients with tooth loss tended to be older and to have longer disease durations (p < .05). Compared with the group with no problem in self-feeding, the poor self-feeding skills group tended to be older, have a longer disease duration, and have poorer MMSE scores, although none of these relationships reached statistical significance. Moreover, patients with poor functional states (BI and FAQ) had problems with all of their eating skills and exhibited a high rate of refusal to eat (Table 4). When the relationships among the scores of the scales used in this study were analyzed, a significantly strong correlation was found between BI and FAQ (r = −.76, p < .001) and moderate correlations between MMSE and BI (r = .51, p < .001) and FAQ (r = −.61, p < .001).

Table 4. - Eating Difficulties Based on Sociodemographic and Clinical Characteristics in the Dementia Group Problems in… Age Duration of Illness Charlson Index MMSE Barthel Index FAQ Tooth loss  Yes (n = 29) 77.93 6.95 38.58 18.90 0.86 0.95 19.00 0.79 58.27 23.76 27.89 2.76  No (n = 21) 63.47 9.27 23.80 23.12 0.90 0.88 20.21 0.49 79.04 17.93 25.76 2.73   t/Z −6.30 −2.48 −0.24 −1.33 3.37 −2.70   p < .001 .017 .807 .182 .002 .009 Manipulating food on plate  Yes (n = 15) 76.26 9.90 41.26 18.61 0.86 0.19 18.20 0.97 40.33 14.93 29.53 0.91  No (n = 35) 69.97 10.61 28.57 22.24 0.88 0.16 20.19 0.52 78.42 16.39 25.91 2.82   t/Z −1.96 −1.94 −0.33 −2.30 7.72 −6.80   p .056 .059 .740 .016 < .001 < .001 Using utensils  Yes (n = 15) 76.26 9.90 41.26 18.61 0.86 0.74 18.20 0.97 40.33 14.93 29.53 0.91  No (n = 35) 69.97 10.61 28.57 22.24 0.88 0.99 20.19 0.52 78.42 16.39 25.91 2.82   t/Z 1.97 1.94 −0.33 −2.40 −7.72 6.80   p .056 .059 .740 .016 < .001 < .001 Need for assistive tools  Yes (n = 12) 75.41 9.74 38.50 19.40 1.08 0.66 18.20 0.97 41.25 14.63 29.58 0.99  No (n = 38) 70.73 10.87 30.44 22.43 0.81 0.98 20.19