Changes in body composition commonly occur with aging in healthy individuals and in those with protein–energy malnutrition, poor physical conditions, or cognitive impairments such as Alzheimer’s disease (AD). Body composition changes over the aging process are characterized by increasing body fat mass, particularly by the reformation of the body’s fat distribution from appendicular fat to trunk fat, with decreasing lean mass and bone mineral density, especially in older adults (1–3). In middle-aged individuals, these changes increase both fat and lean mass, resulting in an increase in body weight (4, 5). Besides these changes, physiological changes in water metabolism and sodium balance occur, leading to alterations in plasma osmolality and changes in the volume within the body’s fluid compartments (6–10).

Mild cognitive impairment (MCI) is an age-related disorder which describes the symptomatic transitional phase between normal cognition and dementia. Clinically, MCI is defined as a set of criteria, including evidence of a change in cognition and lower performance in one or more cognitive domains despite the preservation of independent daily functioning [diagnostic guidelines from National Institute on Aging and the Alzheimer’s Association (11)]. Aside from age, several risk factors contribute to the occurrence and progression of MCI, such as negative lifestyle behaviors and poor dietary habits (12–14). These factors could also lead to chronic non-communicable diseases such as obesity, hypertension, cardiovascular disease, diabetes, and depression, which disrupt physiological homeostasis and hasten the progression of MCI (15, 16).

Recognizing and monitoring changes in body composition can reveal early signs of the body’s vital status, thereby allowing the application of appropriate therapies or lifestyle changes to help reverse the risk of MCI and progression to dementia. Anthropometrical measurements such as weight, height, body mass index (BMI), body circumference, and waist to hip ratio (WHR) have been commonly used in clinical practice to indirectly estimate fat or muscle mass (17–19). However, these measurements can only provide a rough estimate of body mass distribution and cannot discreetly quantify fat or lean mass percentages or intracellular/extracellular water levels (18). Bioelectrical impedance analysis (BIA) is a simple, portable, and relatively accurate technique (20, 21) that can be used to repeatedly measure the body composition of many participants to recognize early changes in body composition status and prognosticate the likelihood of having MCI.

Most previous studies employed specific bioelectrical impedance vector analysis in the later stage of AD rather than during early cognitive impairment. They reported a lower body cell mass, indicated by a lower reactance and phase angle or a shifting to the lower right half of the ellipse, in the AD or dementia group than in individuals with normal cognition (22, 23). These patterns were more prevalent in women with worse psycho-functional status (24, 25). Higher impedance, higher resistance, and longer impedance vector were found in AD patients than in normal controls, suggesting a lower proportion of fat free mass (FFM) (23, 24, 26). Nevertheless, few studies have applied conventional BIA techniques to the assessment of body composition changes in patients with MCI. Some examples include findings of a lower phase angle in women (27) or a higher visceral fat area associated with a lower risk of non-amnestic MCI in older women {mean [standard deviation (SD)] age of 75.0 ± 5.18} (28), though the number of these studies is limited.

Sex is another aspect to be considered for the evaluation and comparison of body composition in individuals with cognitive decline. Sex has a strong relationship with body composition, where women present with a relatively higher fat mass and lower FFM than men (29, 30). Buffa et al. reported a lower body cell mass with dehydration in women with severe AD than in those with mild-moderate AD, but these findings were not evident in men. Regarding bioimpedance, Cova et al. described significantly higher height-normalized values of impedance and resistance in men with MCI than in those without MCI, though this was not observed among women with MCI (26).

Regarding segmental body analysis, few studies have reported changes in the composition of different body segments in individuals with cognitive decline. Mereu et al. demonstrated that bioimpedance results obtained for only the right arm would be similar to whole-body composition in patients with AD (31). Reduction of skeletal muscle mass or muscle strength in the lower limbs was associated with a higher MCI risk in women (32, 33), whereas loss of skeletal muscle mass in the upper and lower limbs was associated with a higher risk of MCI in men (32). Lower extremity skeletal muscle mass was associated with lower cognitive performance and higher cortical beta-amyloid burden (34, 35), whereas upper extremity skeletal muscle mass did not show any association (34). Similarly, lower extremity motor function was found to be prominent in patients with MCI and AD (36, 37).

Subsequently, there is a lack of information about the relationship between body composition, and bioimpedance for each body segment and cognitive decline. Differences by sex in these components between the upper and lower extremities are not well established, especially for individuals with MCI. Therefore, using a segmental multi-frequency BIA technique, we aimed to examine how whole-body composition changes, segmental body fluids, and bioimpedance variables distinctly for the upper and lower extremities differed between cognitively normal (CN) people and those with MCI among both women and men. We hypothesized that lower extremity variables are more strongly linked to the incidence of MCI than upper extremity variables and that this pattern is more noticeable in women than in men. To test our hypotheses, linear logistic regression methods were conducted to examine odds ratios (ORs) before and after controlling for potential covariates, which was then repeated for the different sexes.

In this study, we aimed to examine the association between the BIA variables with the incidence of MCI, including seven whole-body composition variables and seven pairs of segmental BIA variables. The MCI group was shown to be slightly older, more depressive, and had significantly poorer cognitive abilities than CN group. The CN and MCI groups had approximately equivalent body anthropometric indices such as height, weight, and BMI. However, individuals with MCI had a slightly lower fat mass as indicated by lower WHR and PBF and higher PBCM and TBW_FFM than the CN participants (Table 2).

Segmental analysis revealed that MCI individuals had a higher amount of relative water volumes and extra- to intracellular water ratios, but lower resistance, reactance, and phase angle in the upper and/or lower extremities when compared with CN individuals (Figure 2). After accounting for age, sex, GDS score, and the comorbidities, although the whole-body composition variables failed to differentiate CN and MCI, segmental analysis found that segmental water, segmental lean mass, resistance and reactance variables of the lower extremity were significantly associated with MCI (Table 3).

First of all, the segmental water and segmental lean mass in the lower extremities were proportionally associated with MCI, but their ratio indicated by the Water_Lean_lower variable was not significantly associated with MCI (Table 3). This result suggests that the increase in segmental water was directly proportional to the increase in segmental lean mass in the lower extremity, reflecting an increase in segmental lean mass but equal water volume per lean mass in the cognitive decline subjects. Despite the water volume normality, the water distribution between extra- and intracellular spaces showed the tendency of a shifting of water from the intracellular space to the extracellular areas, demonstrating by the ORs of ECW_ICW_lower closely adjoin to the statistical significance threshold (Table 3). This tendency suggests an abnormal water distribution and cellular dehydration status in individuals with cognitive decline, independently of age and the other covariates.

Secondly, the reduction in lower extremity reactance was found to be associated with the higher prevalence of MCI, independent of age, sex, and the other covariates. Lower reactance indicates a lower cell membrane capacitance due to a reduction in body cell mass or a decrease in cell strength due to malnutrition (20, 46). Cova et al. reported a lower nutritional status as indicated by a lower Mini Nutritional Assessment score and a decrease in body cell mass as demonstrated by a shifting of the RXc graph on the right side in individuals with MCI when compared with the reference (26). In studies of dementia, lower reactance was found in women with severe AD than in patients with mild-to-moderate AD, as it was more evident in women with worse psycho-functional status (24, 25) and in men with dementia (23). Our finding of a reactance reduction in the MCI group is in alignment with these publications as we found a negative association between reactance and MCI, demonstrating a decrease in body cell mass or cell strength in the lower extremity in the individuals with cognitive decline. However, considering the positive association between SL_lower and MCI which implies an increase of segmental lean mass in lower extremity, the decrease in body cell mass might be due to the reduction in the volume of individual cell rather than decrease in total cell quantity. Most of living cells contain a large amount of water (47), so that ICW accounts for approximately 70% of total cell mass (48). The decrease in body cell mass or presumably reduced volume of individual cell therefore might be due to the reduction of ICW content in the MCI individuals.

Furthermore, resistance in lower extremity was also found to be linked with the higher prevalence of MCI, regardless of age and the other covariates. Resistance is a measure of resistivity and is inversely proportional to the relative amount of total body water, decrease in resistance indicates a higher proportion of water volume (20, 49). Saragat et al. reported a higher height-normalized resistance in both sexes with AD, but a preserved value in patients with lower psychological status (MMSE < 20.7 compared with MMSE ≥ 20.7 in CN individuals), suggesting a relative increase in fat in individuals with AD than that in CN individuals (24). The same team reported dehydration in women with severe AD as showed by a lengthening of the impedance vector on the RXc graph compared to the control (25). In this study, our MCI group did not have an increase in fat mass percentage or dehydration. In contrast, they had a higher segmental lean mass and a tendency of higher extracellular water compared with CN individuals (Figure 2 and Table 3). Consequently, the increase in segmental lean mass and ECW volume might cause the R_lower reduction in the MCI individuals.

In this study, those with MCI were in the early stage of cognitive decline, with a mean (SD) K-MMSE score of 26.2 (2.5) that is somewhat closer to the mean (SD) of 27.6 (1.9) seen in the CN group than those in the aforementioned studies. Similarly, our MCI participants had better cognitive test scores than the cognitively impaired individuals in the previous studies (24, 25). Therefore, they might not yet have substantial changes in physical activity or nourishment and thus did not experience solid changes in body composition. Furthermore, age was reported to be an important factor to influence the fluctuation of cellular water distribution, such that ECW to ICW ratio increases with age and might be due to the steeper decrease in the ICW content than in the ECW compartment (9, 50). After controlling for age and other covariates, most body composition variables became insignificant; only segmental BIA variables such as resistance and reactance in the lower extremities and the related body composition/water compartment variables such as SL_lower and SW_lower remained to be significantly associated with the incidence of MCI.

After stratifying by sex, we found that women with MCI showed a significant reduction in terms of reactance and resistance of the lower extremity, whereas this pattern did not emerge in men (Table 4). The body composition differences between men and women were diversely reported in previous BIA publications (22, 24, 28). Sex hormones have been known to strongly influence body fat distribution, particularly estrogen and progesterone in women during menopause (49, 51). A common finding is that there is a higher possibility of obesity with predominant abdominal fat accumulation in aging women with a decrease in estrogen secretion (52, 53). However, Collelueri et al. described a U-shaped relationship between estradiol (a major estrogen in females) and body fat mass, such that total body fat mass was found to be decreased when estradiol levels were between 14.0 and 17.4 pg/ml and increased elsewhere (54). The complex relationship between hormonal changes during aging and body composition might regulate the relationship between impedance variables and cognitive decline. Nonetheless, our findings suggest that women with MCI experience more dramatic changes in their body composition and water compartment than men with MCI.

In this study, whole-body composition measurements, such as FFM or PBCM, were investigated as indicators of muscle volume in relation to MCI. Although individuals with MCI showed slightly more muscle volume due to larger FFM and PBCM compared with CN individuals (Table 2), these variables were insignificantly linked with MCI when taking into account age, sex, GDS score, and comorbidities (Table 3). However, segmental reactance in the lower extremity, which is associated with segmental cell mass or segmental muscle volume was significantly associated with the incidence of MCI, especially in women (Tables 3, 4). Individuals with cognitive impairment are likely to experience protein–energy malnutrition and/or decreased physical activity, thus having a decrease in body lean mass or muscle volume. Myokines, which are synthesized and secreted from muscle mass during muscular contractions, regulate the metabolism in the muscle itself (autocrine) as well as in other tissues and organs, such as the brain, through their receptors (para/endocrine) (55). A lower muscle volume reduces myokine secretion and therefore lowers the amount of myokines crossing the blood–brain barrier. This imbalance in the myokine level upregulates pro-inflammatory cytokine production and impairs glucose and lipid metabolism in the brain (56). As a result, individuals with lower muscle volume or muscle atrophy are susceptible to neuronal damage which induces abnormal cognitive ability (57). In this study, a significant reduction in lower extremity reactance indicating a lower body cell mass or lower muscle volume and therefore a decrease in myokine secretion, might be one of the causes of cognitive impairment in the individuals with MCI (58). This finding revealed the importance of segmental body cell mass or muscle volume when examining the association with cognitive decline, especially in women (59).

In terms of muscle quality, phase angle and ECW/ICW have been recommended as indicators of muscle quality (60, 61). Phase angle is a major parameter previously investigated as a predictor of morbidity and mortality for several pathological conditions or critical illnesses, such as mortality in intensive care or in postoperative patients (62–64), sarcopenia, frailty (65), and Human Immunodeficiency Virus infection (66). It is also suggested as a potential indicator for diabetes (67), hyperlipidemia (68), and hypertension (69), which are known risk factors of cognitive impairment. The phase angle cutoff threshold varies across samples of different ages, sexes, and morbidities (65, 70, 71). In this study, MCI patients of both sexes had just a slightly lower phase angle than the normal controls, and this difference disappeared after controlling for covariates in both women and men; the muscle quality expressed by phase angle might not be a sensitive marker of cognitive impairment in its early stage. Moreover, ECW/ICW is another sensitive parameter that is positively correlated with age as well as cognitive function such as executive function, inhibition, or attention ability in elderly people (9, 50). In this study, after controlling for age, sex and the comorbidities, ECW/ICW in upper and lower extremities showed a tendency of associations with MCI prevalence. Further investigations with individuals in the early cognitive impairment stages are needed to confirm the relation between ECW/ICW and MCI.

In the correlation analysis, age was the primary factor in the correlations between cognitive tests (SNSB-II and K-MMSE) and the BIA variables (Figure 3). After adjusting for age and sex, the cognitive tests showed only weak correlations with some BIA variables; resistance and reactance variables did not or only weakly correlated with the cognitive test scores. However, it was still able to produce significant associations with the incidence of MCI, regardless of age and the other covariates. Since the cognitive impairment in the older individuals is a degenerative symptom, it seems to be accompanied by overall decline in physiological functions with an early change of cellular water volume especially in lower extremity.

Daily activities involve physical strengths in both arms and both legs. In this study, bioimpedance changes in the lower extremity were observed and were more significantly associated with MCI than those in the upper extremity. A reduction in physical function in the lower extremity is likely to be related to a higher risk of cognitive decline (32, 33); it suggests that more focus should be placed on impedance variable changes in the lower extremity in relation to the early stage of cognitive impairment.

In this study, among multiple frequencies of 5, 50, and 250 kHz, the frequency of 50 kHz was selected to simplify the data analysis and be representative of the impedance variables. Indeed, the bioimpedance variables at other frequencies, such as 5 and 250 kHz, behave similarly as those at 50 kHz; lower extremity resistance and reactance variables at those frequencies were associated with MCI prevalence as well (Supplementary Table 4).

Compared to precedent publications (22–27, 36), our study contained a sufficient sample size of MCI and CN participants and thus can provide additional insight into body composition/impedance changes occurring in the early stage of cognitive decline. Regarding the transition from CN to dementia, MCI individuals represented a closer-to-normal stage based on cognition test scores. Despite this limitation, segmental bioimpedance variables such as resistance and reactance in the lower extremity and the related water compartment/body composition components were still effective in recognizing and distinguishing MCI patients from CN individuals. These lower extremity features might be used as an early sign of cognitive decline.

There are some limitations in this study. First, confounding factors, such as daily activity hours and nutritional states, may affect body composition and bioimpedance characteristics (72, 73). Second, segmental impedance variables were derived directly from BIA measurements, whereas body composition and segmental water compartments were estimated based on these impedance variables; therefore, accuracy may vary between ethnicity, age, sex, and manufacturer. Third, longitudinal studies are needed to monitor and confirm the changes in body composition, water compartment and bioimpedance features even during the earlier stages of cognitive decline (e.g., from CN or subjective memory impairment).

The selected BIA variables, which reflect the changes in body composition, water compartment and bioimpedance for upper and lower extremities, have weak correlations with the K-MMSE score as well as with the SNSB-II domains, especially when taking age and sex into consideration. Nevertheless, multiple logistic regression analysis reveals significant associations between the lower extremity BIA variables with MCI. Specifically, we found an increase in segmental lean mass and a decrease in body cell volume due to an abnormal cellular water distribution in the lower extremities with MCI, which were demonstrated by the reductions in both resistance and reactance. These characteristic changes of BIA variables may be considered as an early sign accompanied by cognitive impairment in the older population.

The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author.

The studies involving human participants were reviewed and approved by the Institutional Review Board of the Chonnam National University Hospital. The patients/participants provided their written informed consent to participate in this study.

DD analyzed the data and wrote the manuscript. BK assisted in the data analysis. KK handled the Institutional Review Board approval and managed the data. MJ participated in writing the manuscript. KC and KL took care of the data collection and curation. JK designed the study and wrote the manuscript. All authors revised and approved the contents of the manuscript, contributed to the article, and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.