Consensus‐Defined Sarcopenic Obesity and a Related Dynapenic Obesity Phenotype in Relation to Early Cognitive Impairment Mu‐Cyun Wang, Tao‐Chun Peng, Jeng‐Min Chiou, Yen‐Ching Chen, Jen‐Hau Chen Abstract ABSTRACTBackgroundPrevious studies examining sarcopenic obesity in relation to cognitive function have utilized heterogeneous operational definitions. Some studies have evaluated low muscle strength combined with obesity, more appropriately described as dynapenic obesity, under the broader term sarcopenic obesity. We examined the associations of consensus‐defined sarcopenic obesity and a related strength‐based obesity phenotype with global and domain‐specific cognitive function in older adults and evaluated the robustness of these associations across alternative adiposity definitions.MethodsThis prospective cohort study (2015–2019) included non‐demented older adults from the Taiwan Initiative for Geriatric Epidemiological Research. Participants underwent cognitive assessments at baseline and at the 2‐year follow‐up. Handgrip strength was measured using a dynamometer, and total muscle mass was measured by bioelectrical impedance analysis. Obesity was defined primarily as a body mass index ≥ 27 kg/m2. Sarcopenic obesity was defined as the coexistence of low muscle mass, low muscle strength and obesity, whereas dynapenic obesity was defined as low muscle strength and obesity. Multivariable‐adjusted generalized linear mixed models were used to examine the associations of sarcopenic obesity and a related dynapenic obesity phenotype with cognitive function across repeated measurements and over time.ResultsAt baseline, 366 older adults aged ≥ 69 years (54.6% women, mean ± standard deviation [SD] = 75.7 ± 4.8 years) were included. All 16 participants with dynapenic obesity also met the criteria for sarcopenic obesity. Sarcopenic obesity was associated with poor global cognition (MoCA‐T: β^ = −0.96; 95% confidence interval [CI] −1.80 to −0.13), memory (immediate free recall: β^ = −0.38; 95% CI, −0.72 to −0.04) and attention performance (Trail Making Test A: β^ = −0.43; 95% CI, −0.73 to −0.12). Dynapenic obesity was associated with poor global cognition (MoCA‐T: β^ = −1.15; 95% CI, −1.99 to −0.30), memory (immediate free recall: β^ = −0.43; 95% CI, −0.77 to −0.09) and attention performance (Trail Making Test A: β^ = −0.44; 95% CI, −0.75 to −0.14). The associations of sarcopenic obesity and dynapenic obesity with global cognition and attention performance remained evident among women and apolipoprotein E (APOE) ε4 non‐carriers.ConclusionsSarcopenic obesity and a related dynapenic obesity phenotype were associated with poor global and domain‐specific cognitive function in non‐demented older adults. Muscle strength and obesity measures may facilitate early identification of these phenotypes and help identify individuals at increased risk of cognitive impairment. Future studies are needed to determine whether targeted interventions for sarcopenic and dynapenic obesity can prevent or delay cognitive decline.Keywords: cognitive impairment, dynapenic obesity, handgrip strength, muscle mass, obesity, sarcopenic obesity Introduction Sarcopenic obesity, defined as the coexistence of excess adiposity and reduced muscle mass and function [], has gained attention as a distinct condition with adverse health implications (e.g., disability, functional decline, cardiometabolic diseases and mortality) []. While obesity is typically associated with increased cardiometabolic risk, its impact in older adults remains controversial, with some studies suggesting better survival, a phenomenon known as the “obesity paradox” []. However, such findings may mask important heterogeneity in body composition; in particular, the presence of sarcopenia in obese individuals likely affects health outcomes []. Age‐related changes in body composition, including reduced lean mass and increased fat mass, contribute to sarcopenic obesity through a bidirectional relationship: Obesity accelerates muscle loss via inflammation, insulin resistance and oxidative stress [], whereas sarcopenia may promote fat accumulation by reducing energy expenditure []. Shared risk factors, including aging, inactivity and chronic disease, may further perpetuate this cycle. Despite growing interest, prevalence estimates of sarcopenic obesity vary widely because of inconsistent definitions. A recent meta‐analysis reported a global prevalence of 11% in adults aged ≥ 60 years, which increased to 23% in those aged ≥ 75 years []. The lack of consensus until the recent definition released by the European Society for Clinical Nutrition and Metabolism (ESPEN) and the European Association for the Study of Obesity (EASO) in 2022 has limited comparability across earlier studies []. Moreover, dynapenic obesity, defined as the coexistence of excess adiposity and low muscle strength, has been used in the literature as a related but more readily assessed muscle–adiposity phenotype. However, it has often been used inconsistently or interchangeably with sarcopenic obesity, although sarcopenic obesity additionally requires evidence of reduced muscle mass under current consensus definitions []. Both sarcopenic obesity and cognitive impairment are common in aging populations. However, studies of sarcopenic obesity have used heterogeneous operational definitions. These include low muscle strength combined with high body mass index (BMI) or waist circumference [, , ], low muscle mass combined with high waist circumference or fat mass [, , ], and the coexistence of both low muscle mass and low muscle strength with high BMI or fat mass [, ]. Such variation in the muscle and adiposity components used to define sarcopenic obesity complicates comparisons across studies []. Although sarcopenia has been consistently associated with cognitive decline [], findings concerning obesity and dementia have been mixed [, ] and may vary according to apolipoprotein E (APOE) ε4 status [] and inflammatory profiles []. Studies of sarcopenic obesity have likewise yielded inconsistent results. Several studies have reported associations between different muscle–adiposity phenotypes and cognitive impairment [, , , , , , ], whereas two studies have reported no association between low muscle mass combined with high fat mass and cognitive impairment [, ]. Another study revealed that low muscle mass combined with low muscle strength and/or physical performance in the presence of a high fat mass index was not related to cognitive impairment, whereas low muscle strength combined with a high fat mass index was associated with poor global and domain‐specific cognition []. Moreover, relatively few cohort studies have applied standardized definitions to examine these relationships longitudinally [, ]. Because muscle strength is easier to assess than muscle mass in routine clinical and community settings, dynapenic obesity may represent a more readily assessed muscle–adiposity phenotype. However, muscle strength reflects muscle function rather than muscle quantity, and dynapenic obesity should therefore not be considered equivalent to sarcopenic obesity. Evidence remains limited regarding the associations of consensus‐defined sarcopenic obesity and related strength‐based obesity phenotypes with global and domain‐specific cognitive function in older adults. The aim of this study was to examine the associations of sarcopenic obesity and the related, more readily assessed dynapenic obesity phenotype with global and domain‐specific cognitive function in older adults and to evaluate the robustness of these associations across alternative adiposity definitions. Methods Study Design and Participants This prospective cohort study was based on the Taiwan Initiatives for Geriatric Epidemiological Research (TIGER), in which 605 community‐dwelling older adults were initially recruited from the senior health check‐up programme at National Taiwan University Hospital in 2011. In the study, 446 eligible participants were enrolled between 2015 and 2017. Of these participants, 80 were excluded because of prevalent dementia, neurological conditions, missing cognitive or sarcopenic‐obese data, or a BMI < 18.5 kg/m2, yielding a final sample of 366 for data analyses. We conducted follow‐up assessments 2 years later (2017–2019). The study flowchart is shown in Figure . This study was approved by the Research Ethics Committee of National Taiwan University Hospital (REC number: 201101039RB; 201112047RIB; 201312156RINC; 201412213RINC; 201712220RIN; 201712218RIN; 201812102RIN; 202012214RIN; 202012285RINA; 202112042RINA; 202312052RINC; 202401172RINC; 202501209RINB), and written informed consent was obtained from all participants. This study complies with the Helsinki Declaration of 1975. Sarcopenic Obesity and Dynapenic Obesity Sarcopenic obesity is defined as the coexistence of low muscle mass, low muscle strength and obesity [], whereas dynapenic obesity is defined as the coexistence of low muscle strength and obesity []. Total muscle mass (kg) was assessed using bioelectrical impedance analysis (BIA; Tanita BC420‐MA, Japan) and normalized by body weight (kg); participants in the lowest sex‐specific quartile were classified as having low muscle mass [, ]. Muscle strength was assessed using handgrip strength (Lafayette dynamometer, USA), with the Asian Working Group for Sarcopenia (AWGS) cutoffs (< 28 kg for older men and < 18 kg for older women) []. Obesity was defined primarily as a BMI ≥ 27 kg/m2 according to the adult obesity criterion established by the Taiwan Ministry of Health and Welfare []. This ethnicity‐ and population‐specific threshold was selected instead of the conventional BMI ≥ 30 kg/m2 criterion. Alternative definitions of obesity were based on central adiposity and fat mass. Central obesity was defined using Asian‐specific waist circumference cutoffs (≥ 90 cm for men and ≥ 80 cm for women) []. Fat mass index was calculated as body impedance analysis (BIA)‐assessed fat mass divided by height squared (kg/m2), with obesity defined as the highest sex‐specific quartile []. Because evidence supporting height‐adjusted fat mass measures remains limited, this definition was used only in sensitivity analyses []. Cognitive Assessments Cognitive function was evaluated at baseline and follow‐up using a battery of neuropsychological tests administered by trained interviewers. Global cognition was assessed using the Montreal Cognitive Assessment‐Taiwanese version (MoCA‐T) []. Domain‐specific measures included immediate and delayed recall (Wechsler Logical Memory), attention (Trail Making Test A and digit span tests), executive function (Trail Making Test B) and verbal fluency (semantic categories). Domain scores were standardized into z scores. The Trail Making Test scores were reverse‐coded so that lower values indicated poorer performance. Covariates Sociodemographic data, lifestyle factors and comorbidities, including smoking, alcohol consumption, hypertension, diabetes, dyslipidemia, cardiovascular disease and depressive symptoms, were collected via structured questionnaires. Depressive symptoms were considered present if participants met any of the following criteria: a self‐reported depression history, use of antidepressant medication or a score ≥ 16 on the 20‐item Center for Epidemiologic Studies Depression (CES‐D) scale. Physical activity was measured using the International Physical Activity Questionnaire (IPAQ), expressed in metabolic equivalent of task (MET) minutes a week []. Functional status was assessed using the Barthel Index for basic activities of daily living (ADL) and the Lawton–Brody instrumental activities of daily living (IADL) scale. Diet quality was evaluated using the modified Alternative Healthy Eating Index []. Biomarkers Serum inflammatory markers, such as C‐reactive protein (CRP) and interleukin‐6 (IL‐6), were measured in blood samples. Baseline CRP and IL‐6 concentrations were dichotomized at upper‐tertile cutoffs of > 1.2 mg/L and > 1.65 pg/mL, respectively. These thresholds were cohort‐specific and represented relatively high levels of low‐grade inflammation within this generally healthy, community‐dwelling population. The APOE ε4 genotype was determined on the basis of two single‐nucleotide polymorphisms, rs429358 and rs7412, using TaqMan SNP assays on an ABI 7900HT Fast Real‐Time PCR system (Applied Biosystems Inc., CA, USA). Individuals with at least one ε4 allele were categorized as APOE ε4 positive. Statistical Analyses Baseline characteristics were compared across sarcopenic obesity status and dynapenic obesity status using the Kruskal–Wallis test for continuous variables and the chi‐square test or Fisher's exact test for categorical variables. Generalized linear mixed models (GLMMs) with random intercepts were used to examine the associations of sarcopenic obesity and dynapenic obesity with cognitive outcomes over the 2‐year follow‐up period. Models were adjusted for potential confounders, including demographic characteristics, comorbidities, lifestyle factors, inflammatory markers and practice effects (defined as the number of completed cognitive assessments). Two sets of models were evaluated. Model 1 included sarcopenic obesity status or dynapenic obesity status as a time‐varying exposure to examine associations between phenotype status and cognitive performance across repeated measurements. Model 2 included baseline phenotype status, follow‐up time and their interaction to assess whether cognitive performance over time differed according to baseline phenotype status. Stratified analyses were conducted by sex, APOE ε4 status, diabetes status, and IL‐6 and CRP levels using multivariable‐adjusted GLMMs; these analyses were not adjusted for multiple comparisons. Several sensitivity analyses were performed. First, we repeated the analyses using alternative definitions of obesity, including central obesity defined as a waist circumference ≥ 90 cm for men and ≥ 80 cm for women and a high fat mass index defined as the highest sex‐specific quartile. Second, because the primary definition of low muscle mass was based on the lowest sex‐specific quartile of weight‐adjusted muscle mass, we conducted additional sensitivity analyses using alternative sex‐specific thresholds, including the lowest tertile and lowest quintile, to evaluate whether the findings were robust to the choice of internal cutoffs. Third, participants who had been excluded solely because of a BMI < 18.5 kg/m2 were re‐included, and the main GLMM analyses were repeated in this expanded analytic sample (n = 373). Because muscle mass was standardized by body weight, the resulting measure was structurally related to BMI‐defined obesity. We explored whether the muscle mass component could be examined separately by presenting the distribution of these combinations descriptively rather than through a formal comparison between mass‐based and strength‐based phenotypes. As previously mentioned, participants whose baseline cognitive data were missing or whose information was insufficient to determine sarcopenic obesity status were excluded. The mixed‐effects models included all available repeated observations; observations with missing outcome, exposure or covariate data at a given visit were excluded from the analysis for that visit. The number of missing values for each covariate is reported in Table . All the statistical tests were two‐sided, and a p value < 0.05 was considered to indicate statistical significance. Analyses were conducted using SAS version 9.4 (SAS Institute, Cary, NC, USA), and graphs were generated using R version 4.5.1. Results Characteristics of the Study Population Among 366 older adults (median age 75 years, 55% women), 303 (83%) had neither sarcopenia nor obesity, 20 (5%) had sarcopenia alone, 27 (7%) had obesity alone and 16 (4%) had sarcopenic obesity (Table ). Among the four groups (sarcopenic‐obese, nonsarcopenic‐obese, sarcopenic‐nonobese and nonsarcopenic‐nonobese), participants with sarcopenia alone were the oldest and exhibited the lowest handgrip strength and gait speed. Individuals with obesity alone had the greatest waist circumference. Participants with sarcopenic obesity had the lowest muscle mass, the highest BMI and the highest fat mass and fat mass index. The number of missing values for each covariate was small at baseline. In terms of dynapenic obesity status, 204 participants (56%) had neither dynapenia nor obesity, 119 (33%) had dynapenia only, 27 (7%) had obesity only and 16 (4%) had dynapenic obesity (Table ). Among the 303 participants classified as having neither sarcopenia nor obesity, 99 had dynapenia alone. All 16 obese participants with dynapenia also met the criteria for sarcopenia, resulting in complete overlap between the two phenotypes. Therefore, the corresponding estimates should be interpreted as findings for related, non‐independent muscle–adiposity phenotypes rather than as independent associations for two distinct groups. When obesity was alternatively defined using waist circumference, 23.8% of participants classified as having obesity alone under the BMI‐based definition were reclassified as having dynapenic obesity (data not shown). Sarcopenic Obesity and Cognitive Function In the multivariable GLMMs (Table ), Model 1 (upper part) revealed that participants with sarcopenic obesity was associated with poorer global cognition (MoCA‐T: β^ = − 0.96; 95% CI: −1.80 to −0.13), memory (immediate free recall: β^ = − 0.38; 95% CI: −0.72 to −0.04) and attention performance (Trail Making Test A: β^ = − 0.43; 95% CI: −0.73 to −0.12) compared with those without sarcopenia or obesity. Obesity alone was associated with poorer executive function, as assessed by the Trail Making Test B (β^ = − 0.29; 95% CI: −0.56 to −0.01). In Model 2 (lower part), baseline sarcopenic obesity was associated with poorer global cognition, memory and attention performance, whereas baseline sarcopenia alone was associated with poorer attention. Baseline sarcopenic obesity showed a significant interaction with follow‐up time for the Trail Making Test B, suggesting that executive function over time was poorer in participants with baseline sarcopenic obesity. Baseline sarcopenia alone also showed a significant interaction with follow‐up time for the immediate theme and free recall, suggesting that memory performance over time was poorer in participants with baseline sarcopenia alone. In exploratory stratified analyses, sarcopenic obesity was associated with poorer MoCA‐T performance among women, APOE ε4 non‐carriers and participants in the highest tertile of IL‐6 or CRP (β^ = − 1.93 to −0.89; all p < 0.05; Figure ). It was also associated with poorer immediate free recall among men, APOE ε4 non‐carriers, participants without diabetes and participants in the highest tertile of CRP (β^ = − 0.82 to −0.36; all p < 0.05; Figure ) and with poorer Trail Making Test A performance among women, APOE ε4 non‐carriers, participants without diabetes and participants in the lower two tertiles of IL‐6 (β^ = − 0.56 to −0.43; all p < 0.05; Figure ). Obesity alone was associated with poorer Trail Making Test B performance in selected subgroups (β^ = − 1.14 to −0.32; all p < 0.05; data not shown). Sensitivity analyses using alternative obesity definitions (waist circumference and fat mass index) yielded similar results. When waist circumference exceeded the cutoffs (≥ 90 cm for men and ≥ 80 cm for women), sarcopenic obesity was associated with poorer global cognition (MoCA‐T: β^ = − 0.68; 95% CI: −1.29 to −0.06) and attention performance (Trail Making Test A: β^ = − 0.24; 95% CI: −0.46 to −0.03) (Figure and Table ). When the highest quartile of fat mass index (≥ 6 kg/m2 for men and ≥ 9.26 kg/m2 for women) was used, sarcopenic obesity remained associated with poorer global cognition (MoCA‐T: β^ = − 0.64; 95% CI: −1.25 to −0.03) and memory (immediate free recall: β^ = − 0.26; 95% CI: −0.50 to −0.01) (Figure and Table ). When alternative sex‐specific thresholds for weight‐adjusted muscle mass (the lowest tertile and lowest quintile) were used, sarcopenic obesity remained associated with poorer global cognition, memory and attention performance (Tables and ). After the seven participants who had been excluded solely because of a BMI < 18.5 kg/m2 were re‐included, the expanded analytic sample (n = 373) yielded estimates similar to those from the original sample (n = 366), and the conclusions remained unchanged (Table ). Dynapenic Obesity and Cognitive Function In the multivariable analyses (Table ), Model 1 (upper part) revealed that participants with dynapenic obesity were associated with poorer global cognition (MoCA‐T: β^ = − 1.15; 95% CI: −1.99 to −0.30), memory (immediate free recall: β^ = − 0.43; 95% CI: −0.77 to −0.09) and attention performance (Trail Making Test A: β^ = − 0.44; 95% CI: −0.75 to −0.14) compared with those without dynapenia or obesity. Dynapenia alone was associated with poorer global cognition (MoCA‐T: β^ = − 0.48; 95% CI: −0.85 to −0.10), memory (delayed free recall: β^ = − 0.18; 95% CI: −0.32 to −0.03), executive function (Trail Making Test B: β^ = − 0.17; 95% CI: −0.31 to −0.03) and attention (digit span forward: β^ = − 0.15; 95% CI: −0.30 to −0.01). Obesity alone was associated with poorer executive function, as assessed by the Trail Making Test B (β^ = − 0.34; 95% CI: −0.62 to −0.06). In Model 2 (lower part), baseline dynapenic obesity was associated with poorer global cognition, memory, attention performance and executive function. Baseline dynapenia alone was associated with poorer memory and attention. Baseline dynapenic obesity was associated with poor Trail Making Test B performance and showed a significant interaction with follow‐up time, suggesting that executive function over time was poorer in participants with baseline dynapenic obesity. Baseline dynapenia alone also showed a significant interaction with follow‐up time for the MoCA‐T, suggesting that global cognition over time was poorer in participants with baseline dynapenia alone. Dynapenic obesity showed similar patterns in exploratory stratified analyses. It was associated with poorer MoCA‐T performance among women, APOE ε4 non‐carriers, participants without diabetes and those in the highest tertile of IL‐6 or CRP (β^ = − 1.94 to −1.00; all p < 0.05; Figure ); with poorer immediate free recall among men and the corresponding APOE ε4, diabetes and inflammatory‐marker subgroups (β^ = − 0.81 to −0.40; all p < 0.05; Figure ); and with poorer Trail Making Test A performance among women, APOE ε4 non‐carriers, participants without diabetes and those in the lower two tertiles of IL‐6 (β^ = − 0.54 to −0.45; all p < 0.05; Figure ). Dynapenia alone was associated with poorer global cognition, delayed free recall, Trail Making Test B performance and digit span forward in selected subgroups (β^ = − 1.11 to −0.18; all p < 0.05; data not shown). Sensitivity analyses using alternative obesity definitions revealed that dynapenic obesity remained associated with poorer global cognitive function, whereas dynapenia alone remained associated with poorer memory and executive function (Figure , Tables and ). After re‐including participants who had been excluded solely because of a BMI < 18.5 kg/m2, the estimates were similar to those from the original analytic sample, and the conclusions remained unchanged (Table ). Muscle Mass–Strength Combinations Among Participants With Obesity Among the 43 participants with BMI‐defined obesity at baseline, 16 had both low muscle mass and low muscle strength, 25 had low muscle mass with normal muscle strength, 2 had normal muscle mass and normal muscle strength and none had low muscle strength with normal muscle mass. Thus, sarcopenic obesity completely overlapped with dynapenic obesity in this cohort. In addition, nearly all participants with obesity met the internally defined criterion for low weight‐adjusted muscle mass. Given the resulting sparse and highly unbalanced cells, formal comparisons among the muscle mass–strength combinations were not considered statistically reliable. Discussion This longitudinal study of community‐dwelling older adults revealed that sarcopenic obesity was associated with poor global cognition, memory and attention performance, particularly among APOE ε4 non‐carriers. Importantly, this investigation employed the consensus‐defined sarcopenic obesity and applied multiple obesity definitions with ethnicity‐specific cutoffs when associations with cognitive function were examined. Our findings reinforce the growing body of evidence linking sarcopenic obesity to cognitive impairment [, , ]. We additionally examined dynapenic obesity as a related, more readily assessed strength‐based obesity phenotype with potential relevance to clinical and community screening. Under the ESPEN–EASO framework, sarcopenic obesity is conceptually nested within dynapenic obesity. In our data, however, all participants with dynapenic obesity also fulfilled the criteria for sarcopenic obesity, precluding a direct comparison between the two phenotypes. Accordingly, our findings should not be interpreted as evidence that dynapenic obesity is equivalent to sarcopenic obesity or that muscle strength is more important than muscle mass. Rather, they suggest that a strength‐based obesity phenotype, assessed using handgrip strength and BMI, may help identify older adults with a clinically relevant muscle–adiposity imbalance who are vulnerable to poor cognitive function. Confirmation in larger cohorts with more balanced distributions of muscle mass and muscle strength is needed to determine whether dynapenic obesity provides information distinct from consensus‐defined sarcopenic obesity. The reasons for inconsistent results across previous studies include variations in age distribution, follow‐up duration and use of different cognitive assessment tools. For example, whereas one cohort study evaluated incident dementia for more than a decade [], our study focused on preclinical and prodromal cognitive impairment in older adults without prevalent dementia. The stronger evidence observed in our cohort may therefore reflect an association of sarcopenic obesity with earlier cognitive impairment that precedes overt dementia. Moreover, heterogeneity in the diagnostic criteria of sarcopenic obesity complicates direct comparisons of prevalence and association with cognitive performance across studies []. Associations between dynapenic obesity and cognitive decline have been reported in the US, Japanese and Mexican cohorts [, , ]. These findings support the potential clinical relevance of dynapenic obesity as a pragmatic screening phenotype when direct assessment of muscle mass is unavailable. However, our study did not differentiate dynapenic obesity from sarcopenic obesity with respect to cognitive impairment, and comparisons between muscle strength and muscle mass cannot be made because of the highly unbalanced combinations of muscle mass and strength. Further confirmation in larger cohorts with mutually exclusive phenotype groups is needed. Our results also suggest potential sex‐based differences: Sarcopenic obesity and dynapenic obesity are associated with impairment in global cognition and attention performance in women. This pattern is partly consistent with a previous study from Japan, which revealed that the coexistence of decreased handgrip strength and elevated BMI was associated with mild cognitive impairment as assessed by the Montreal Cognitive Assessment in women []. These observations may reflect sex‐specific vulnerabilities related to muscle strength decline and differences in inflammatory and hormonal profiles [, ]. However, these findings should be interpreted cautiously because of small subgroup cell sizes and multiple comparisons. Further studies are needed to elucidate the underlying biological mechanisms involved. At baseline (2015–2017), the prevalence of sarcopenic obesity in our study was 4.4%, which was comparable with the prevalence of 4.3%–5.3% reported among older Japanese adults using the ESPEN–EASO criteria []. When obesity was alternatively defined using waist circumference and fat mass index, the prevalence of sarcopenic obesity increased to 8.7% and 8.5%, respectively, approximating that obtained using a BMI threshold of ≥ 25 kg/m2. All of these estimates were lower than the reported global prevalence of approximately 11% among adults aged ≥ 60 years [], although direct comparisons should be made with caution because of heterogeneity in study populations and operational definitions. Our cohort consisted of community‐dwelling participants (mean age of 75.7 years) who were generally healthy at baseline, which may explain the lower prevalence of this condition compared with clinical samples. As the cohort ages, their health profile is expected to converge towards that of the general older population. We adopted the ESPEN–EASO conceptual framework, under which sarcopenic obesity requires both low muscle mass and strength in the presence of obesity. This approach differs from that used in earlier studies in which sarcopenia was defined on the basis of low muscle mass or low muscle strength alone. The ESPEN–EASO criteria do not include gait speed because of potential confounding from knee osteoarthritis, which is common among obese older adults []. Although the Asia–Oceania consensus definition for sarcopenic obesity adopts the AWGS 2019 criteria, which incorporate physical performance measures (e.g., 6‐m gait speed, Short Physical Performance Battery and the 5‐Times Sit‐to‐Stand Test) [], the latest AWGS 2025 has simplified the diagnostic criteria of sarcopenia as concurrent low muscle mass and strength []. Therefore, applying the ESPEN–EASO conceptual framework to this population was methodologically appropriate. Our sensitivity analyses, in which obesity was alternatively defined using waist circumference and fat mass index, revealed associations with global cognition that were consistent with those based on BMI‐defined obesity, supporting the robustness of our findings across different adiposity measures. These results also aligned with prior evidence suggesting that both central and overall adiposity might be associated with adverse cognitive outcomes when accompanied by low muscle strength or mass [, ]. Although the associations based on the BMI definition were generally stronger than those based on waist circumference or fat mass index, no formal tests were conducted to compare the estimates across adiposity definitions. One possible explanation is that the stricter BMI threshold of ≥ 27 kg/m2 resulted in the identification of a smaller group with a more pronounced muscle–adiposity phenotype, as reflected by the lower prevalence of sarcopenic obesity under this definition. This interpretation should nevertheless be considered exploratory. The mechanisms linking sarcopenic obesity to cognitive impairment are likely multifactorial. Age‐related muscle decline and physical inactivity may promote adiposity, insulin resistance and low‐grade systemic inflammation, which may in turn impair muscle metabolism and contribute to vascular dysfunction and neuroinflammation. In our stratified analyses, sarcopenic obesity and dynapenic obesity were associated with poor global cognition, as assessed by the MoCA‐T, among participants in the upper tertile of IL‐6 or CRP. Although the concentrations of these markers were generally below commonly used clinical thresholds for overt inflammation, these findings suggest that relative differences in low‐grade inflammation may be relevant to cognitive vulnerability []. Both phenotypes were also associated with poor memory and attention performance among participants without diabetes. These patterns are compatible with possible inflammatory and metabolic pathways linking muscle–adiposity imbalance to cognitive health; however, given the observational design and exploratory nature of the subgroup analyses, they do not establish causal mechanisms. From a clinical perspective, our findings may inform a pragmatic first‐step assessment based on obesity and muscle strength. Dynapenic obesity, identified using simple measures of handgrip strength and BMI, may represent a readily assessable screening phenotype to identify older adults who could benefit from further cognitive evaluation. When feasible, body composition assessment may provide additional information by characterizing the muscle mass component of the broader muscle–adiposity imbalance represented by sarcopenic obesity. However, the clinical utility of this approach requires validation in larger studies. One of the major strengths of this study is its longitudinal design, which included repeated measurements of both sarcopenic obesity status and cognition. We applied the ESPEN–EASO conceptual framework, providing a more explicit and consistent operational definition of sarcopenic obesity. This study has several limitations. First, low muscle mass was defined using the lowest sex‐specific quartile of BIA‐estimated, weight‐adjusted muscle mass. This internally derived threshold may depend on the distribution of the study sample and may limit comparability with studies using externally validated clinical cutoffs. Furthermore, BIA provides an indirect estimate of body composition rather than a direct anatomical measurement of skeletal muscle and is influenced by hydration status and device‐specific prediction equations. Appendicular skeletal muscle mass assessed using dual‐energy X‐ray absorptiometry (DXA) or other imaging modalities was unavailable. Nevertheless, sensitivity analyses using alternative thresholds yielded similar conclusions. Second, obesity was defined primarily using BMI, which is widely used in clinical and epidemiological settings, has established ethnicity‐specific cutoffs in Taiwan and is readily applicable to community‐based screening. However, BMI is only an indirect proxy for adiposity and does not capture fat distribution or distinguish fat mass from lean mass, which may lead to the misclassification of adiposity. To address this limitation, we performed sensitivity analyses using waist circumference and fat mass index as alternative adiposity measures. The associations with global cognition were generally consistent across these definitions, supporting the robustness of our findings. Third, although participants with dementia at baseline were excluded to reduce the possibility of reverse causation, reverse causality cannot be completely ruled out in an observational study. Early cognitive impairment may have contributed to reduced physical activity, poor nutritional status, low muscle strength or changes in body composition. In addition, physical activity was assessed by the IPAQ, which may be subject to recall and reporting bias. Future studies with longer follow‐up periods and repeated assessments of cognition, nutritional status, muscle strength and body composition, together with objective measures of physical activity such as accelerometers or pedometers, are needed to clarify the temporal sequence of these relationships. Finally, although sarcopenic and dynapenic obesity were relatively uncommon in this cohort, the associations were generally consistent across the different cognitive outcomes and in the sensitivity analyses. Nevertheless, the small number of affected participants may limit the statistical power and precision for phenotype‐specific inference, particularly in stratified analyses, where subgroup cell sizes were further reduced. In addition, multiple subgroup comparisons were performed without adjustment for multiplicity, increasing the possibility of chance findings. These subgroup findings should therefore be considered exploratory and require confirmation in larger cohorts with more participants in each phenotype category. Taken together, our findings suggest that sarcopenic obesity and a more readily assessed dynapenic obesity phenotype may represent clinically relevant conditions associated with early cognitive impairment. Handgrip strength and BMI may provide a practical first‐step, community‐based approach for identifying older adults who may benefit from further muscle mass and cognitive evaluations. Larger cohorts with longer follow‐up periods and integration of relevant biomarkers are needed to clarify the relationship between muscle–adiposity imbalance and neurocognitive health. Ethics Statement All the authors of this manuscript comply with the guidelines of ethical authorship and of publishing in the Journal of Cachexia, Sarcopenia and Muscle. Consent All participants provided informed consent for their participation in the Taiwan Initiatives for Geriatric Epidemiological Research (TIGER). Contributor Information Yen‐Ching Chen, Email: karenchen@ntu.edu.tw. Jen‐Hau Chen, Email: jhhchen@ntu.edu.tw.