Lower fluid and fruits / vegetable intake in questionable dementia among older Hong Kong Chinese

Abstract

Objective

Nutrition plays a role in the ageing process of the brain and suboptimal nutrient intake might precede clinical cognitive impairment. A diet high in fruits and vegetables has been recommended while little has been said about the influence of fluid intake in cognitive function. We examine the dietary pattern of community-dwelling older individuals with questionable dementia and compared that with normal individuals.

Design

Cross-sectional.

Setting

Community.

Participants

285 community-dwellers aged 60 or older.

Measurements

Dietary habits were recorded using the Mini-Nutritional Assessment (MNA). Questionable dementia was diagnosed by psychogeriatricians and relevant demographic and dietary factors were examined using univariate then multivariate analyses.

Results

146 questionable dementia and 139 cognitively normal subjects were interviewed. Both groups were not at risk of malnutrition (MNA score 26.1 vs. 26.7 respectively, p = 0.02). The former were older, had fewer years of education, lower MMSE and ADAS-cog as well as lower MNA scores. In univariate analysis, questionable dementia was associated with decline in food intake and appetite, eating less vegetables and fruits, and drinking less fluid. After adjustment for age, gender and education level, eating > 2 servings of vegetables / fruits per day (OR 0.26, 95% CI 0.067, 0.973) and taking > 5 cups of fluids per day (OR 0.4, 95% CI 0.204, 0.792) was associated with a lower prevalence of questionable dementia.

Conclusion

Older people with questionable dementia have lower intakes of vegetables, fruits and fluid than those who were cognitively normal. This may pose additional health risks, and increase their chance of progressing into dementia.

Introduction

Nutrition plays a role in the ageing process of the brain. Fruits and vegetables high in antioxidants have been found to be able to counteract oxidative stress vulnerability of the aging brain (1). Older people having higher dietary intake of vitamins and micronutrients beneficial to brain function were found to maintain higher cognitive function ( 2., 3., 4., 5., 6.), yet randomized controlled trials using supplements of vitamins and micronutrient rich in fruits and vegetables had yielded mixed results ( 7., 8., 9.). Likewise, foods rich in polyunsaturated fats and antioxidants were important in reducing the progression into dementia (10, 11), but the role of antioxidants in preventing cognitive decline has remained largely elusive, with conflicting results from different reports (3, 12, 13).

On the other hand, relatively little has been said about the relationship between fluid intake and cognitive function in older people. Gopinathan et al. studied the psychological functions of young soldiers under dehydration and found that short term memory, arithmetic ability, motor speed and attention all began to deteriorate progressively when body weight loss by dehydration reached 2% (14). Szinnai et al. reported that in young adults, cognitive-motor function was preserved in up to 2.6% of body weight loss by dehydration, but increased subjective task-related effort and concentration was necessary for the same cognitive tasks to be accomplished during dehydration than in euhydration status (15). However, all these results were obtained from young subjects. Little is known about how hydration will affect long-term cognitive function and in particular, in older persons.

The Mini- Nutritional Assessment (MNA) (16, 17) is a simple tool designed to screen for malnutrition and risk for malnutrition among older people in healthy or disease states. Using the MNA, a high percentage of dementia patients have been identified to be either malnourished or at risk of malnutrition (17). As the onset of dementia is insidious, and nutrients or the lack of them contribute to brain health by affecting anything from neuronal membrane regeneration to cognitive performance (18, 19), it would be of interest to see whether a state of nutritional deficiency or suboptimal dietary pattern actually precedes the actual symptomatic cognitive decline. If that is the case, identifying and correcting the eating pattern in individuals in the prodromal stage of dementia may delay the onset of clinical dementia, or at least reduce the rate of further cognitive decline.

By using the MNA, we examine the association between dietary pattern and the presence of questionable dementia (QD) – a possible prodromal stage of dementia, among a representative sample of community-dwelling older Chinese in Hong Kong, to explore whether diet and fluid intake bears any relationship to very early cognitive impairment.

Methods

Study Sample

The study was conducted from October 2005 to July 2006 as part of a large population study examining the prevalence of very mild and mild dementia among the community-dwelling older population in Hong Kong. In the population study, people aged 60 or older were identified in household surveys and invited to be screened for cognitive impairment by the Mini-Mental State Examination (MMSE) (20, 21), and a short memory questionnaire (Abbreviated Memory Inventory for the Chinese, AMIC) by trained interviewers (22). In total, 2073 subjects were screened positive.

Three groups of elderly subjects were invited for further assessment by psychogeriatricians and to undergo cognitive testing including the Alzheimer’s Disease Assessment Scale – Cognitive subscale (ADAS-Cog) (23, 24). The groups were: (1) those with MMSE scores below the local cutoff for dementia (< 18 with no education, < 20 with 1 to 2 years education, and < 22 for > 2 years education); 2) those with MMSE scores above the cutoff for dementia but with significant memory complaints according to the AMIC. The combined MMSE and AMIC cutoffs were previously determined to have 97% sensitivity and 72% specificity in identifying subjects with MCI and clinical dementia in Hong Kong (25); and 3) 5% of cognitively normal elderly subjects randomly selected. Details of the procedure were further described elsewhere (25).

Diagnosis and rating of dementia

Dementia was diagnosed according to the DSM-VI classification (American Psychiatric Association, 1994) and graded according to the Clinical Dementia Rating Scale (CDR) (26) by psychogeriatricians. The CDR is a semi-structure interview to determine the diagnosis and severity of dementia, giving rise to 5 global ratings (0 = normal cognition, 0.5 = questionable dementia, 1 = mild dementia, 2 = moderate dementia and 3 = severe dementia). Six subscales covering different clinical aspects are used to form the global ratings.

Dietary assessment

Subjects assessed by psychogeriatricians were invited to attend a dietary assessment at their local elderly centre. Dietary habits were recorded using the Mini-Nutritional Assessment (Chinese version) (MNA), modified for the use among Chinese. (27, 28) It is an 18-item questionnaire covering four areas: arthropometric assessment (weight, height, body mass index (BMI) and recent weight changes); global assessment (mobility, recent hospitalization, medications, neuropsychological symptoms, independent living; presence of sores); dietary assessment (amount of foods and fluid intake); and subjective assessment (self-perception of nutrition and health). All questions have weighted answers and the maximum score is 30. Subjects were considered malnourished if the score was <18.5, at risk of malnutrition if between 18.5 and 23.5, and not at risk of malnutrition if the score is > 23.5.

Ethics Approval

The study has been approved by the ethics committees of the Chinese University of Hong Kong and the Department of Health of the Hong Kong SAR. Written informed consent was obtained for all subjects. For moderate to severely demented subjects who were unable to obtain consent, the first degree relatives were contacted for consent by proxy.

Statistical Analysis

Statistical analysis was performed by using the SPSS version 12.0. Possible associated factors were examined in the first step using univariate analysis. In the second step, factors with p<0.05 were entered into multivariate analyses.

Results

After evaluation by psychogeriatricians, 316 subjects consented to be assessed using the MNA. Among them, 139 were cognitively normal (CDR 0), 146 had QD (CDR 0.5), 25 were diagnosed to have early dementia (CDR 1), 5 had moderate dementia (CDR 2) and 1 had severe dementia (CDR 3). Only MNA data from subjects with normal cognitive function and QD were included for analysis, as subjects with more significant cognitive impairment might have recall errors on the MNA.

Subjects with QD were significantly older than their normal counterparts and received less education. Living arrangement, personal and household incomes were not different between the normal and QD subjects (Table 1). When compared with normal subjects, QD subjects had similar body mass index (BMI) but significantly lower mean MNA scores (26.1 ± 1.9 versus 26.8 ± 1.7, p<0.005). On univariate analysis of diet related items in the MNA, reduced food intake and appetite, eating less vegetables and fruits, and drinking less fluid were associated with QD (Table 1).

Table 1.

Demographics and univariate analyses of dietary factors of QD and normal older people

Factor	QD (n=146)	Normal (n=139)	p value*
Demographics
Age	72.4 (6.7)	68.6 (6.1)	<0.001
Female	81 (55.5)	68 (48.9)	0.268
Socioeconomic factors
Years of education, mean (sd)	3.7 (3.8)	7.0 (4.8)	<0.001
Living with someone	124 (85.5)	123 (88.5)	0.457
Marital status			0.534
Married	98 (67.6)	97 (69.8)
Widowed	43 (29.7)	38 (27.3)
Divorced / Separated / Single	4 (2.7)	4 (2.9)
Monthly household Income			0.119
<HK$ 10,000	75 (51.7)	66 (47.5)
10,000 – 19,999	33 (22.8)	24 (17.3)
20,000 – 29,999	18 (12.4)	25 (18.0)
30,000 or above	17 (11.7)	24 (17.2)
refused to answer	2 (1.4)	0 (0.0)
Monthly Personal Income			0.163
<HK$5,000	131 (90.3)	117 (84.2)
5,000 – 9,999	10 (6.9)	19 (13.7)
10,000 or above	4 (2.8)	3 (2.1)
Cognitive measurements
ADASCOG	15.5 (5.7)	8.3 (4.0)	<0.001
MMSE	24.4 (2.8)	27.2 (2.0)	<0.001
Nutritional measurements and MNA items
BMI (kg/m2)	23.4 (3.3)	23.1 (3.1)	0.331
MNA score, mean (sd)	26.1 (1.9)	26.8 (1.7)	0.002
Decline in food intake and appetite			0.045
severe loss	3 (2.1)	1 (0.7)
moderate loss	9 (6.2)	3 (2.2)
no loss	133 (91.7)	135 (97.1)
Weight loss in past 6 months			0.323
>3 kg	3 (2.1)	4 (2.9)
1-3 kg	16 (11.0)	17 (12.2)
none	105 (72.4)	107 (77.0)
not certain	21 (14.5)	11 (7.9)
Rice, noodles, congee per day			0.642
< 1.5 bowls	6 (4.1)	1 (0.7)
1.5 to 2.5 bowls	63 (43.2)	63 (45.3)
2.5 or more bowls	77 (52.7)	75 (54.0)
Servings of vegetables / fruits per day			0.001
< 1	13 (8.9)	4 (2.9)
1 to 2	58 (39.7)	39 (28.1)
2 or more	75 (51.4)	96 (69.1)
Fish, meat, chicken, egg or tofu per day			0.444
< 2 taels	6 (4.1)	3 (2.1)
2 to 4 taels	57 (39.0)	48 (34.5)
4 or more taels	83 (56.8)	88 (63.3)
Milk or milk supplement per day			0.444
< 1/2 cup	105 (71.9)	95 (68.3)
1/2 to 3/4 cup	16 (11.0)	14 (10.1)
3/4 cup or more	25 (17.1)	30 (21.6)
Fluids consumed per day			<0.001
<3 cups	5 (3.4)	0 (0.0)
3-5 cups	44 (30.1)	21 (15.1)
>5 cups	97 (66.4)	118 (84.9)
Is your intake adequate?			0.286
No	3 (2.1)	6 (4.3)
Yes	142 (97.3)	133 (95.7)
Uncertain	1 (0.7)	0 (0.0)

^*p value or p for trend for variables with 3 categories or more. Student t test and Chi-square test for continuous and categorical data respectively; QD = questionable dementia, ADASCOG = Alzheimer’s Disease Assessment Scale cognitive subscale, MMSE = Mini-Mental State Exam score, BMI = body mass index, MNA = Mini Nutrition Assessment

On multivariate analyses by logistic regression, using age, sex and education as covariates, eating more than 2 servings of fruits and vegetables per day as compared to only 1-2 servings and less, and drinking 5 or more cups of fluids per day as compared to having less than 5 cups were associated with a lower chance of being diagnosed with QD (Table 2).

Table 2.

Multivariate analyses of dietary factors associated with QD

Factor	N(%)	Adjusted 95%		p value
		OR	confidence interval
Age (per yr increment)	285 (100)	1.08	1.034, 1.128	0.001*
Female gender	149 (52.3)	0.88	0.485, 1.601	0.678
Education levelf
None	61 (21.7)	1	-
Primary	127 (45.2)	0.494	0.238,,1.025	0.058
Secondary	73 (26.0)	0.211	0.090, 0.496	0.000*
Tertiary	20 (7.1)	0.029	0.005, 0.182	0.000*
No decline in food	268 (94.0)	0.448	0.127,1.586	0.213
intake and appetite
Servings of vegetables
/ fruits per day
<1	17 (5.9)	1	-
1 to 2	94 (33.9)	0.32	0.083, 1.242	0.100
>2	169 (60.2)	0.26	0.067, 0.973	0.045*
Fluids per day
≤ 5 cups	67 (23.9)	1	-
> 5 cups	213 (76.1)	0.40	0.204, 0.792	0.008*

^*p<0.05; f Education data missing for 4 subjects; OR = odds ratio

Discussion

We found that elderly community dwellers with questionable dementia eat less fruits and vegetables. The role of antioxidants on cognitive function had been extensively studied. Studies focusing on nutrients abundant in fruits and vegetables (vitamins C, E, folate, carotinoids and flavinoids) had generally demonstrated better cognitive profile for those with higher intake of these, either as food or supplement, or in the plasma (2, 4, 5, 8, 13, 29, 30). Prospective studies on dietary patterns also showed that older people with high intakes of fruit and vegetables had higher cognitive scores, lower rate of cognitive decline, and lower incidence of dementia (32, 33). However, results from randomized controlled trials were inconclusive: vitamins C and E as supplements had failed to give cognitive benefits in some (7, 9), while others reported a lower dementia incidence with these supplements (8, 34). Our finding suggests a difference in dietary pattern among the cognitively normal and marginally impaired older people, prior to clinical dementia. This difference may compromises the intake of nutrients beneficial to brain health.

The intake of fruits and vegetables in our subjects are related to their chance of being diagnosed as QD. This suggests that either those who took less of these foods were more prone to QD, or that those having this pre-clinical state of cognitive impairment had some underlying causes yet unknown causing them to take less of these foods. In the elderly, poor dentition is a common cause for poor fruits/ vegetable intake (35) due to difficulties in chewing. Poorer masticatory function as well as fewer teeth was found to be related to poorer cognitive function (36), which might be a reason for poorer fruits/ vegetable intake in this group. In QD and in particular those with poor dentition therefore, supplementary folate, antioxidants or flavinoids might be of extra value.

When compared with the group with very low fruits and vegetable intake (< 1 serving per day), even a slightly higher intake (>2 servings per day) short of the widely recommended 5 servings per day were associated with a lower chance of QD (OR 0.26). Low folate intake is a likely result of a diet poor in fruit and vegetables. Folate, together with vitamin B12, is important in the methylation reactions in the productions of neurotransmittors, phospholipids and myelin in brain tissue. It has been postulated that age-related cognitive decline may be partly due to impairment of these reactions (18, 37). In addition, hyperhomocysteinaemia in low folate states may contribute to vascular disease of the brain, causing further damage. A recent randomized controlled trial indicated that subjects taking supplemental folic acid for 3 years performed significantly better in cognitive tests than those on placebo. (38). Our results suggested that among older persons with a diet very low in fruits and vegetables, any slight increase in these antioxidants might already be beneficial in preserving cognitive function.

In our results, the different intake level of fruits and vegetables between the cognitively impaired and normal groups had persisted after adjustment to gender, age and educational level. Larrieu et al had reported that women and those of higher education tend to take more fruits and vegetables, and that older subjects tend to take less (39). Prevalence studies in the Hong Kong populations had reported a much higher prevalence of cognitive impairment among women, those with older age, lower educational level and income (25, 40). We found that despite the effect of age, gender, education and income, very low fruits and vegetables intake was still independently associated with QD.

We found fluid intake in excess of five cups per day in accordance with current dietary recommendations were associated with a much lower chance of QD (OR 0.4). The relationship between fluid intake and cognitive function had been much less explored compared with that in nutrients. Gopinathan et al. (14) and Cian et al. (41) had demonstrated experimentally that cognitive functions including short term memory, arithmetic ability, motor speed and attention, and perceptive discrimination were impaired with heat or exercise-induced dehydration in young healthy adults. A more recent study demonstrated that though neurocognitive performance remained the same during dehydration up to 2.6%, subjects felt more tired and require more effort and concentration to perform the tasks (15). Although no similar work had been done in chronic mild dehydration, it is possible that the same phenomena occurs and results in or worsens underlying borderline cognitive impairment in the elderly.

Alternatively, being a cross-sectional study, the association between less fluid intake and cognitive dysfunction could be the reverse: cognitively impaired persons might experience less sense of thirst hence take less fluid. In the elderly, dehydration is an established precipitant of delirium. Several theories have been proposed for dehydration related cognitive deficits, supported by animal and in vitro studies (42). Hypercortisolaemia associated with dehydration has been linked to poorer active learning and short term memory (43). Nitric oxide, an important neurotransmitter facilitating long-term memory, may also come into play in dehydration. Older rats were found to have reduced nitric oxide (NO) production (44), which suggests that the normal response of increased NO secretion in response to dehydration might be impaired in older persons, resulting in susceptibility to cognitive dysfunction in hypovolaemic states. A small human study involving nine Alzheimer’s disease (AD) patients and nine normal controls showed that AD subjects had a greater degree of dehydration after overnight fluid restriction, had lower vasopressin levels for their osmolalities, and had a reduced sense of thirst compared with the controls. They also drank less fluid to quench their thirst in comparison with the controls (45). As the cerebral cholinergic system was impaired in AD, and vasopressin is also under cholinergic control, a pre-clinical state of cholinergic deficiency could present as reduced thirst and fluid intake prior to full blown AD in questionable dementia subjects. Our results showed that there was a strong association between fluid intake and QD, suggesting a possible vicious cycle between reduced thirst and persistent chronic mild dehydration, which might perpetuate further cognitive decline in those already in this marginal pre-clinical state.

Nevertheless, the study pointed out a difference in the dietary pattern, in particular of fluid intake, between normal elderly and those with questionable dementia. The implication could be important as older people tend to restrict their fluid intakes due to fear of urinary frequency or prostatism in men, and stress or urge incontinence in women. This self-inflicted fluid restriction could lead to a chronic mild dehydration state, especially in warm weather, subtropical or tropical regions. On the other hand, poor dentition in old age might be a cause of lower fruits/ vegetable intake and dentition is a largely neglected area in old age care. Our result indicates the need for further investigations in these areas and stresses the importance of adequate fluid and vegetable/ fruits intake in the older population for brain health.

In this study, the MNA had proved useful in describing the difference in dietary patterns of normal and borderline cognitively impaired older people. As far as we know, this is one of the first studies using this instrument to look at how dietary habits might be related to borderline cognitive impairment in the elderly. Prospective studies using the MNA might be important in studying how deficiencies in fruit/ vegetable and fluid intakes might alter the progress of very early dementia. However, the MNA, although being a very important screening tool for the risk of malnutrition, is not an instrument constructed (or validated) to analyse past eating behaviour. Our results therefore should be interpreted with caution as they described the eating behaviour of questionable dementia subjects using items originally designed to detect malnutrition.

It should however be noted that that the age of the normal group was significantly younger than that of the cognitively impaired group. The mean age of the QD group was 72.4 years while that of the normal group, 68.6 years (p<.0000). As the prevalence of dementia increases with age, many people from the younger ‘control’ group might develop QD as they become a few years older. This age discrepancy might also have affected the association between less fluid intake and cognitive impairment. It might be possible that the normal group, being a few years younger, was more engaged in physical and other outdoor activities, and as a result their fluid intake may be higher. Although we have statistically adjusted for this age difference in the multivariate model, nevertheless these points should be taken into account when interpreting the results.

Limitations

Being cross-sectional, our results could not be taken as conclusive in indicating a direct causal relationship between very early cognitive impairment and lower fruit/ vegetable or fluid intakes. Fruits and vegetables intake in the MNA was categorized with the aim to detect malnutrition, with the highest category set as more than two servings, rather than the recommended five servings. Our results might have been diluted by those having intakes between 2 to 5 servings per day, and the actual protective effect of taking the recommended amount might have been even more pronounced. Similarly, an ordinal protective effect according to the three categories of fluid intake would be more desirable. However, owing to the very few numbers of subjects in the lowest fluid intake category, the lower two groups were grouped into one for analysis. The relationship between fluid intake and cognitive impairment might be shown more clearly with a larger sample, and the message would then be more obvious. Lastly, as subjects were assessed in elderly centres, those too frail to travel outside their homes were likely to have been excluded.

Conclusion

Older people with QD have lower intakes of vegetables, fruits and fluid and this dietary pattern may put them at risk of further cognitive decline. The MNA might be useful in detecting these dietary differences between those in the pre-clinical stage of cognitive impairment and the normal elderly. Prospective studies are needed to investigate how significant these effects could be and to examine whether changing these suboptimal dietary habits could indeed reduce the chance of progressing into clinical dementia.

Financial disclosure

None of the authors had any financial interest or support for this paper.