Dementia is a serious health issue that occurs in later life in the aging population. Global estimates indicate that the number of older adults suffering from dementia will rise¹⁾ and a decline in cognitive function attributes to adverse health outcomes.²⁾ According to the 2020 report of the Lancet Commission on Dementia Prevention, Intervention, and Care, 40% of global dementia cases are attributable to modifiable risk factors. Therefore, the identification of modifiable risk factors for cognitive decline is critical.³⁾ Mild cognitive impairment (MCI) is a condition that occurs between dementia and age-related cognitive decline. Interventions to prevent and postpone dementia onset aimed at older adults with MCI have been developed.⁴⁾ Among the modifiable factors for dementia in the lifestyles of older adults are dietary patterns.⁵⁾

Several studies have examined the relationship between dietary factors and cognitive function. Otsuka et al.⁶⁾ reported that ingesting a variety of foods was positively associated with cognitive function. The World Health Organization’s guidelines recommend healthy and balanced dietary patterns to prevent cognitive decline and enhance cognitive function.⁷⁾ A prospective study of 11,179 Chinese community-dwelling older adults also showed that low dietary diversity increases the risk of cognitive impairment.⁸⁾ Other studies have shown that ingesting a variety of foods contributes to the maintenance and improvement of cognitive function.

Dietary habits, such as eating out, eating prepared, and home-cooked food helps maintain dietary diversity.⁹⁾ According to previous studies, individuals who eat meals at home are more likely to have a high-quality diet compared with those who eat meals outside the home.¹⁰⁾ However, Lachat et al.¹¹⁾ reported that eating outside was positively associated with dietary diversity and better dietary intake in terms of energy supply from fat among Vietnamese adolescents. Thus, the association between eating out and dietary diversity remains controversial. Particularly, any association between eating out and dietary diversity among community-dwelling older adults remains unclear.

Therefore, this study examined the association between eating out and dietary diversity among community-dwelling older adults. We also investigated the direct or indirect associations of eating-out habits with MCI based on dietary diversity. Clarifying these relationships may provide a better understanding of the association between these factors to develop intervention strategies to prevent cognitive decline.

This study included data from 597 older adults with a median age of 73.0 years (IQR 69.0–78.0), 62.6% of whom were women. According to the Kolmogorov–Smirnov test, the continuous variables in the present study were not normally distributed (p<0.01). Table 1 shows the participants’ demographic characteristics. The participants were divided into high (n=485; 81.2%) and low (n=112; 18.8%) dietary diversity groups. Compared with participants with high dietary diversity, those with low dietary diversity were less likely to be women, have less education, and have a higher BMI (all p<0.05). The overall prevalence of MCI was 20.4 % (n=122). The prevalence was higher among participants with low dietary diversity than those with high dietary diversity (28.6% vs. 18.6%; p<0.05). Furthermore, we classified 23.8% (n=142) of participants as not eating out, with a higher proportion in the low dietary diversity group than in the high dietary diversity group (33.0% vs. 21.6%; p<0.05).

The results of the adjusted multiple logistic regression analysis, which examined the relationship among MCI, not eating out, and low dietary diversity, are presented in Table 2. After adjusting for confounding variables such as age, sex, education, medication, BMI, and single status, participants who did not eat out had low dietary diversity (OR=1.97; 95% CI 1.20–3.20). Second, we examined the association between not eating out and MCI using a logistic regression analysis. The proportion of participants who did not eat out was not significantly associated with MCI (OR=0.76; 95% CI 0.46–1.25). Multiple logistic regression analysis indicated an association between low dietary diversity and MCI. The results showed that low dietary diversity at baseline was associated with a higher odds ratio of MCI compared with participants with high dietary diversity (OR=1.72; 95% CI 1.02–2.87).

Fig. 2 shows the results of the SEM, which examined whether the effect of not eating out on MCI was indirectly transmitted through dietary diversity status, or was directly related to MCI. This model indicated that not eating out was not directly associated with MCI (β=-0.037, p=0.36) but contributed significantly to low dietary diversity (β=0.108, p<0.01). Furthermore, low dietary diversity was significantly and directly associated with MCI (β=0.101 p<0.05). Moreover, not eating out was associated with MCI via low dietary diversity and provided an excellent fit to the data (GFI=0.999, AGFI=0.984, RMSEA=0.030, AIC=29.549).

The results of the present study indicated that older adults with non-eating out habits had low dietary diversity patterns, and that low dietary diversity was associated with MCI. Analysis of the SEM showed that not eating out was not directly associated with MCI but was indirectly associated with MCI through low dietary diversity.

Multiple analyses in the present study revealed low dietary diversity among community-dwelling older adults who did not eat out. Food literacy might explain these findings. To define the term "food literacy," Vidgen and Gallegos¹⁸⁾ identified an effect that was related to the areas of preparation, eating, planning and management, and selection. In a nationwide cross-sectional study in Japan, high food literacy was associated with high scores on the National Cancer Institute’s Healthy Eating Index 2015, which indicated high diet quality.¹⁹⁾ In another study, individuals with a low frequency of eating out demonstrated lower food literacy compared with those with a higher frequency.²⁰⁾ Thus, older adults who do not eat out have lower food literacy, and therefore, may have lower dietary diversity.

The present cross-sectional study examined the association between low dietary diversity and MCI among community-dwelling older adults. The potential mechanisms behind the present findings may involve the complex effects of various nutrients on cognitive function due to dietary diversity. First, calcium and magnesium have favorable effects against vascular diseases by inhibiting free radical formation and platelet aggregation, improving dyslipidemia, and increasing insulin sensitivity.^21-23) Calcium and magnesium are abundant in milk and dairy products. A prospective cohort study showed that ingesting milk and dairy products reduced the risk of developing dementia after a follow-up of >17 years.²⁴⁾ Second, the intake of diverse amino acids has protective effects on cognitive function. A longitudinal study showed that very low meat consumption increases the long-term risk of dementia²⁵⁾ and that the intake of amino acids, especially lysine, phenylalanine, threonine, and alanine, is positively associated with cognitive function in late life.²⁶⁾ Thus, the intake of foods containing proteins and amino acids, such as meat, is important to enhance cognitive function. One potential explanation for these effects is that the ingestion of amino acids contributes to neurotransmitter synthesis and growth hormone levels.^27,28) A previous large prospective cohort study reported that high fish consumption was associated with a lower risk of dementia.²⁹⁾ Fish contain nutrients, including eicosapentaenoic acid,³⁰⁾ docosahexaenoic acid,³¹⁾ vitamins B,^32,33) and vitamin D.³⁴⁾ Furthermore, vitamins C³⁵⁾ and vitamin E,³⁶⁾ and carotenoids³⁷⁾ are also present in green and yellow vegetables and fruits. However, because food and nutrients are consumed in combination rather than separately, increasing and improving dietary diversity may be a more successful strategy for delaying cognitive decline.

In the current study, we used structural equation modeling to clarify whether not eating out was directly or indirectly associated with MCI. Our results indicated that not eating out was not directly associated with MCI. However, it did have an indirect effect owing to low dietary diversity. Nutritional knowledge is an important factor in maintaining dietary diversity and balance.³⁸⁾ A previous cross-sectional study reported that lower nutritional knowledge was associated with a lower frequency of eating out,³⁹⁾ in which older adults with a high frequency of eating out may be more nutritionally knowledgeable. Lee et al. reported higher dietary diversity and intake of protein, fat, calcium, phosphorus, iron, potassium, vitamin B₁, vitamin B₂, and niacin among adults who favored non-homemade meals, including eating out.⁹⁾ Furthermore, social isolation, including eating alone, was associated with a decline in cognitive function in a 3-year prospective study.⁴⁰⁾ Moreover, eating outside the home was positively associated with dietary diversity and better dietary intake in terms of energy supply from fat among Vietnamese adolescents.¹¹⁾ Thus, eating out could be a key factor in ensuring a positive nutritional status and dietary diversity. A positive nutritional status and consumption of a variety of foods have favorable effects on cognitive function.⁵⁾ According to these findings, eating-out behavior was associated with nutritional status and dietary diversity status and may be related to MCI through nutritional status.

The present study examined the association between participants who did not eat out and social relationships (going out less frequently than the previous year, visiting friends sometimes, and talking with someone every day)⁴¹⁾ (Supplementary Table S1). The results suggest that non-eating out participants were less likely to visit friends and converse with others compared with older adults who ate out. A previous large cohort study showed that persistent loneliness in midlife is an independent risk factor for dementia and Alzheimer’s disease⁴²⁾ and that social relationships such as social networks, social activity participation, and social support were negatively related to incident dementia.⁴³⁾ Based on previous studies and our results, we believe that similar aspects may link eating habits to social relationships and loneliness. Future research should further clarify the relationship between eating habits, social engagement, and MCI among community-dwelling older adults.

This study has several limitations. First, although we evaluated the frequency of intake, we could not assess the quantity of each food item. Future studies should consider the quantity of food ingested. Second, because this was a cross-sectional study, we were unable to determine the causal relationships between MCI, eating-out practices, and dietary diversity. Third, we excluded socioeconomic and economic status domains from the multivariate analyses. Other possible covariates (e.g., lifestyle activity and socioeconomic status) should also be considered as they may be connected to eating out, dietary diversity, and MCI. Fourth, this cohort study used self-reported diet assessments. Although we excluded older adults with dementia and severe cognitive impairment, recall bias was still possible. Additional research using objective food records such as pictures is required. Fifth, while we asked participants about the frequency of their eating-out habits, we could not evaluate what they ate when eating out. Eating habits such as meal content, frequency of eating out, food environment, and the cost of meals vary by country and region. Our sample was obtained from a single country (Japan) and the target population was older adults living in Tarumizu City, Japan. Further investigation is required to clarify whether our findings can be generalized to other countries and regions.

In conclusion, the results of this cross-sectional study revealed that non-eating habits were not directly associated with MCI. In addition, we applied structural equation models to identify that non-eating habits were indirectly associated with MCI due to low dietary diversity. These findings suggest that promoting eating-out behaviors may lead to enhanced dietary diversity, resulting in positive effects on cognitive function.