This study examined the differences in PA levels and EI according to the presence or absence of MetS and sex among ESH in Korea to understand the correlations between these factors. The total MetS group engaged in significantly less recreational moderate PA and total PA than the non-MetS group. In particular, the recreational moderate and total PA levels were significantly lower in males with MetS, whereas females with MetS had only significantly lower recreational moderate PA. "Active" was associated with lowering high waist circumference and "Very active" with lower MetS occurrence and low HDL-C as in Model 2 using adjustments for sex, smoking, alcohol consumption, and body mass index. Examination of differences in energy intake according to the presence or absence of MetS showed that there was a significant difference only in fat intake in females, which was lower in those with MetS; no significant difference was found in the total group or in males. According to the odds ratio of EI, "moderate energy intake" was found to be associated with only HDL-C in Model 2, and there was no association between components of EI and MetS. Taken together, our results suggest that single-household MetS prevention is more strongly associated with PA than EI.
In the present study, we found that higher recreational moderate PA was associated with low MetS morbidity in both males and females. A previous study of 477 people (aged 55–80 years) in Spain found that the MetS group had lower energy expenditure and less leisure-time PA (< 4 MET) than the non-MetS group . Jung et al.  investigated 3,720 participants in the Korea National Health and Nutrition Examination Survey (KNHANES) from 2016 to 2018, aged > 65 years irrespective of household type, and compared MetS risk with PA level. They reported that the extent of PA according to the presence or absence of MetS differed more in terms of recreational PA than occupational PA. Our study also showed that MetS was more inactive than non-MetS in the total group; therefore, it appears that recreational moderate PA is important to lower the risk factors for MetS in elderly people. In addition, in relation to PA level, Smith et al. (2017)  and Sarkar et al. (2016)  reported that having someone (e.g. family and friends) is positively correlated with PA level, whereas living alone may promote a decrease in total PA, which may be as a risk factor for increased occurrence of MetS. Comparing our study with ESH and with mixed households (single and mixed-household type) , PA at the total PA and recreational moderate levels of elderly people with MetS in the mixed-household type was higher than in our study; therefore, it can be observed that PA among ESH is low. Moreover, when the total PA level of mixed-households and ESH in our study was compared by sex, there was no significant difference in the presence or absence of MetS in males by mixed-household type, but there was a significant difference in females. Conversely, in our study, there was a significant difference in the presence or absence of MetS in males, but not in females. Previous studies have shown that older women living alone engage in more activities and contact with friends through social relationships than older men living alone . There appeared to be no difference in PA levels between the MetS and non-MetS groups due to the social characteristics of women. Furthermore, in the study by Jung et al. , according to sex, the male total PA in the MetS group of mixed households was higher and female total PA and recreational moderate PA in the MetS group of mixed households was higher than in our study. This difference can be interpreted as a result according to the type of household, and it can be expected that single households are more exposed to risk factors of MetS because the PA level is lower than that of mixed households.
In our study, an analysis of dietary intake in relation to the presence or absence of MetS was also performed in elderly people. This showed that total EI and carbohydrate, fat, and protein intakes did not show significantly different between MetS and non-MetS in both sexes, with the exception of fat intake in females. According to the results for a mixed household , carbohydrate, fat, and protein intakes were also not significantly different between MetS and non-MetS groups in males. However, in females, the total energy intake (carbohydrate, fat, and protein) was significantly different between the MetS and non-MetS groups. This result is supported by a previous study, which reported that the dietary quality and food diversity of females was better than that of males . Moreover, the overall low nutrient intake and low nutrient density of meals were the major nutritional problems in the group of ESH. As in the previous study mentioned above, ESH also consumed less than the recommended dietary intake. Therefore, malnutrition rather than a nutrient excess appears to be the problem in ESH, and consumption of a balanced diet may be more important than deficient intake of a single nutrient. These dietary intake patterns of the elderly have been described in more detail in previous studies. Giezenaar et al. reported that low EI is a strong predictor of poor outcomes, including the development of pathological undernutrition and sarcopenia, as well as reduced functioning and frailty; this low EI in the elderly affects the decline in PA . Therefore, ESH appear to experience a greater reduction in PA compared with mixed households.
The results of this study should be interpreted with the following limitations in mind. First, we assessed ESH with MetS but did not consider the timing of MetS development or the duration of MetS. Second, PA levels were not determined using heart rate measurements or an accelerometer but were based on survey results, which are prone to errors. Third, this study found only simple differences without establishing causality of underlying the association between PA and nutrition. Finally, the data obtained using the 24-hour reminder may not reflect long-term dietary habits. The 24-hour recall is essentially a retrospective method of diet assessment, in which an individual is asked about their food and beverage consumption during the previous day or the 24 hours. However, a single 24 hour-recall may not be representative of the habitual diet at an individual level. The strength of this study is that it analyzed PA levels and energy intake of ESH as a function of the prevalence of MetS. Most studies examining the relationship between MetS, PA levels, and energy intake have not considered household types in the elderly. In particular, this study classified them according to sex and investigated the PA levels and energy intake of ESH. Therefore, the risk of MetS depends on lifestyle habits, such as PA level and energy intake, and the risk of MetS can be prevented by increased PA levels, and a balanced dietary and healthy energy intake. In this study, we identified sex-specific aspects of PA levels and energy intake with and without MetS in ESH. In addition, this information can be used to reduce the incidence of MetS in ESH according to sex.