The association between composite dietary antioxidant index and sarcopenic obesity Authors: Xi Luo, Weiwei Jin Published: 2025 ABSTRACT AbstractBackground and objectiveThe composite dietary antioxidant index (CDAI) is a tool to assess an individual’s comprehensive dietary antioxidant intake profiles. This study aimed to explore the association between energy-adjusted composite dietary antioxidant index (E-CDAI) and sarcopenic obesity (SO).MethodsParticipants from the NHANES cohort (n = 4388) and the Tongde Hospital of Zhejiang Province cohort (n = 276) were enrolled in the present study. Logistic regression analyses were applied to explore the association between E-CDAI (both continuous and quintiles) and SO. Moreover, the mediating effect of inflammatory/nutritional indicators were evaluated to investigate the association between E-CDAI and SO.ResultsIn the NHANES cohort, logistic regression demonstrated a significant inverse correlation between E-CDAI and SO (adjust odd ratio (OR)continuous =0.89, 95% CI = 0.81, 0.98, P = 0.041; adjust ORQ5vsQ1 =0.44, 95% CI = 0.29, 1.18, Ptrend =0.050) In the Chinese population cohort, an inverse correlation also existed between E-CDAI and SO (adjust ORcontinuous =0.72, 95% CI = 0.60, 0.87, P < 0.001; adjust ORQ5vsQ1 =0.30, 95% CI = 0.13, 0.74, Ptrend <0.001). Applying data from the NHANES cohort, the mediation analysis indicated that neutrophil-percentage-to-albumin ratio mediated 12.812% and C-reactive protein mediated 6.485% of the associations between E-CDAI and SO.ConclusionA decreased E-CDAI score is associated with the increased risk of SO in adults. The association appeared to be partially mediated through inflammatory/nutritional pathways, suggesting that the E-CDAI may serve as a valuable indicator for identification of individuals at risk for SO.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12986-025-01069-x.Keywords: Sarcopenic obesity, Composite dietary antioxidant index, Neutrophil-percentage-to-albumin ratio, C-reactive protein, NHANES Introduction Sarcopenic obesity (SO), a functional and clinical disorder, marked by coexistence of excess fat mass (FM) and low muscle mass/function [1]. It rapidly developing into a global epidemic and may lead to serious health outcomes, such as diminished quality of life, increased risks of frailty and disability, and even mortality [2, 3]. The pathophysiology of SO is complicated and multifactorial. To date, several mechanisms including sedentary lifestyle, insulin resistance, hormonal changes, chronic low-grade inflammatory state, and oxidative stress (Os) have been implicated in the onset and development of SO [4–8]. Os is a key factor that exacerbates SO through endoplasmic reticulum stress, mitochondrial dysfunction, and dysregulation in both fatty infiltration and muscular regeneration pathways [7]. Many factors contribute to Os, with diet being one of the most significant modifiable factors [9, 10]. Unhealthy dietary patterns are closely linked to decreased antioxidant capacity [10, 11], resulting in increased adverse health outcomes, such as sarcopenia, obesity, and various chronic non-communicable diseases [12–17]. The composite dietary antioxidant index (CDAI) is a scoring tool designed to assess the antioxidant capacity of an individual’s daily food intake profiles [17]. It was constructed based on their collective anti-inflammatory properties [18]. A higher CDAI score indicates an increased antioxidant capacity of diet, whereas a lower CDAI score reflects a decreased antioxidant capacity of diet. Nowadays, the implications of daily dietary antioxidant capacity on sarcopenia and obesity have garnered increasing attention. Many population-based studies have investigated the associations between the CDAI score and these conditions, highlighting the strong inverse connection between CDAI and the risk of probable obesity and sarcopenia [11–16]. However, there is limited research exploring the association between CDAI and SO. Contemporary epidemiological research has demonstrated that SO is not exclusive to the elderly. With changes in lifestyle such as poor dietary structures and sedentary behaviors, progressively younger populations are presenting pre-SO/SO manifestations [3]. Therefore, elucidating CDAI-SO interaction provides significant insight for formulating targeted nutritional strategies to optimize therapeutic efficacy in weight management. The present study aimed to explore the association between energy-adjusted composite dietary antioxidant index (E-) score for the daily diet and the SO in adults, using the datasets from the National Health and Nutrition Examination Survey (NHANES) cohort and the Tongde Hospital of Zhejiang Province cohort. Furthermore, we explored whether inflammatory and nutritional indicators have a potential role in mediating the association between E-CDAI and SO. Materials and methods Study population Participants from the NHANES cohort (n = 4388) and the Tongde Hospital of Zhejiang Province cohort (n = 276) were enrolled in the present study. The detailed research process was demonstrated in Fig. 1. Flow chart of study participants NHANES, National Health and Nutrition Examination Survey; E-CDAI, energy-adjusted composite dietary antioxidant index; SO, sarcopenic obesity; BIA, bioelectrical impedance analysis In the NHANES cohort, datasets including demographic information, physical examination results, health-related data, and dietary details were utilized from two NHANES cycles (2015–2016 and 2017–2018). Participants were excluded based on the following criteria: individuals (I) aged < 20 years; (II) pregnancy status; and (III) those missing data on CDAI, energy intake, SO-related parameters, and inflammation/nutritional parameters, (IV) energy intake < 600 kcal/d or > 5000 kcal/d, resulting in a final cohort of 4388 subjects (Figure S1). The survey was conducted in accordance with the local legislation and institutional requirements, and the approved NHANES protocols were documented on https://www.cdc.gov/nchs/nhanes/about/erb.html. All participants provided written informed consent at enrollment. In the Chinese population cohort, participants were enrolled from the weight management program at Tongde Hospital of Zhejiang Province between January 2024 and February 2025. In this program, participants with overweight and obesity (OAO) were recruited but excluded if they met the following criteria: individuals (I) aged < 20 years; (II) pregnancy status; (III) congenital muscle or skeletal underdevelopment; (IV) a history of weight-loss or muscle-gain treatment within the past 3 months; (Ⅴ) suffering from chronic wasting diseases, including malignant tumors, poorly controlled diabetes, and complex heart disease; (Ⅵ) presence of metal implants, such as cardiac pacemakers, which could affect the accuracy of bioelectrical impedance analysis (BIA) results. From the cohort, twenty-eight subjects were excluded due to missing dietary intake data, body composition, or covariates, which resulted in a cohort of 276 subjects enrolled in analysis (Figure S2). The study was approved by the Medical Ethics Committee of Tongde Hospital of Zhejiang Province (NO.2023–182 K, NO.2023–183 K), and all participants signed informed consent. All methods were performed in accordance with the relevant guidelines and regulations, and a full description is online available in the China Trial Register (ChiCTR2400080592, ChiCTR2400080393). Assessment of E-CDAI The E-CDAI includes six antioxidant dietary components (carotenoids, selenium, zinc, vitamin A, vitamin C, and vitamin E). To estimate participants’ daily carotenoids, selenium, zinc, vitamin A, vitamin C, and vitamin E intake, the present study calculated the daily nutrient intake through 24-hour recall interviews. In the NHANES cohort, dietary intake information was collected at Mobile Examination Centers. We calculated the mean nutrient intake values from two reliable 24-hour recall interviews, specifically excluding supplements and medications [17]. In the Chinese population cohort, dietary intake information was evaluated using three reliable 24-hour recall interviews, comprising two weekdays and one weekend day. All participants engaged in face-to-face interviews, utilizing food models and pictures to improve the accuracy of the dietary intake data. The participants’ nutrient intake data was calculated using a computerized app NutriData with 21 food categories and more than 3000 foods in accordance with the China Food Composition (National Institute of Nutrition and food safety, China CDC) [19]. To calculate the E-CDAI score, the following steps were carried out: (I) we calculated the CDAI based on the followed formula: (II) CDAI score was expressed per 1000 kcal/day. The E-CDAI was used for all the analyses. The advantage of the E-CDAI over the original CDAI is that it accounts for inter-individual differences in total energy intake. Assessment of SO SO was defined as the coexistence of sarcopenia and obesity (OAO, or abdominal obesity) [1].In the NHANES cohort, appendicular skeletal muscle mass (ASM) was assessed using dual-energy x-ray absorptiometry. Body mass index (BMI) was calculated using the equation: the ratio of weight (kg)/height (m) squared. The sarcopenia index (SI) was calculated by follow formulas: SI = ASM (kg)/BMI (kg/m²). Sarcopenia was defined using the sex-specific SI cut-offs established by the Foundation for the National Institutes of Health, which are < 0.789 for males and < 0.512 for females [17]. OAO was categorized as overweight (25.0 kg/m² ≤ BMI < 30.0 kg/m²) and obesity (BMI ≥ 30.0 kg/m²) according to the WHO reference. Abdominal obesity was defined using waist circumference (WC) cut-offs, which are for males ≥ 102 cm and ≥ 88 cm for females. In the Chinese population cohort, body composition was assessed by multi-frequency BIA analyzer (InBody720, Biospace, Korea). Body composition, including body weight, ASM, FM, percent body fat (PBF) and visceral fat area (VFA). Height was measured with a height-weight scale (HCM-800, SHENGYUAN, China). WC was measured in a standing position, at the midpoint between the anterior superior iliac spine and the lower rib, after a normal expiration, using a non-stretchable tape. The appendicular skeletal muscle mass index (ASMI) was calculated by follow formulas: ASMI (kg/m²) = ASM (kg)/Height² (m²). Sarcopenia was defined using the sex-specific ASMI cut-offs (via BIA) established by the Asian Working Group for Sarcopenia 2019 Consensus, which are < 7.0 kg/m² for males and < 5.7 kg/m² for females [20]. BMI was categorized as overweight (24.0 kg/m² ≤ BMI < 28.0 kg/m²) and obesity (BMI ≥ 28.0 kg/m²) according to the Chinese BMI cut-off values. Abdominal obesity was defined using Chinese WC cut-offs, which are for males ≥ 90 cm and ≥ 85 cm for females. Assessment of inflammatory/nutritional indicators Following the NHANES cohort, laboratory parameters including albumin, C-reactive protein (CRP), and complete blood counts (CBCs) such as white blood cell (WBC), lymphocytes, neutrophils, monocytes, and platelets were collected. The inflammatory and nutritional indicator, neutrophil-percentage-to-albumin ratio (NPAR), was calculated by following formulas [21]: Assessment of covariates In the NHANES cohort, the potential covariates considered in the present study included age, gender, race, marital status, education level, family poverty-to-income ratio (PIR), smoking status, drinking status, physical activity (PA) level, diabetes mellitus (DM), hypertension, and hyperlipidemia history. Smoking status was categorized into never, former, and current smokers. Never smokers refer to smoked < 100 cigarettes in their life. Former smokers refer to having smoked ≥ 100 cigarettes in their life and does not smoke cigarettes currently. Current smokers refer to having smoked ≥ 100 cigarettes in their life and still smoking now. Drinking status was categorized into mild, moderate, and heavy drinkers according to the average number of alcoholic drinks over the past year. Mild drinkers refer to ≤ 2 alcohol drinks per day for males or ≤ 1 alcohol drink per day for females. Moderate drinkers refer to 2–4 alcohol drinks per day for males or 1–3 alcohol drinks per day for females. Heavy drinkers refer to ≥ 5 alcohol drinks per day for males or ≥ 4 alcohol drinks per day for females. PA levels were categorized into low and high PA levels. Low PA levels refer to < 500 MET-min/week, and high PA levels refer to ≥ 500 MET-min/week [22]. DM was indicated by any of the following criteria: a self-reported history of DM, use of insulin or oral antidiabetic medications, fasting glucose ≥ 126 mg/dL, two-hour glucose ≥ 200 mg/dL, or glycosylated hemoglobin ≥ 6.5%. Hypertension was diagnosed based on self-reported history of hypertension, systolic blood pressure ≥ 140 mmHg, and/or diastolic blood pressure ≥ 90 mmHg. Hyperlipidemia was identified in participants who were taking lipid-lowering medications, TG ≥ 150 mg/dL, total cholesterol (TC) ≥ 200 mg/dL, low-density lipoprotein cholesterol (LDL-C) ≥ 130 mg/dL, and/or HDL-C ≤ 40 mg/dL for males or ≤ 50 mg/dL for females [23]. In the Chinese population cohort, the covariates including age, gender, marital status, education level, smoking status, PA level, DM, hypertension, and hyperlipidemia history. PA levels were assessed using the international physical activity questionnaire [22]. PA levels were calculated and categorized into low PA (< 500 MET-min/week) and high PA (≥ 500 MET-min/week). Data on DM, hypertension, and hyperlipidemia were collected through self-reported previous diagnosis. Statistical analysis We conducted a comparison of baseline characteristics between the SO and non-SO groups. Due to the intricate sampling methodology and sample weights, NHANES provided nationally representative data. Accordingly, we used the sample weight calculation method proposed by NHANES to appropriately weigh the data in this study. In the NHANES cohort, continuous variables were expressed as mean (standard error), while categorical variables were presented as number (percentage). In the Chinese population cohort, continuous variables were expressed as mean (standard deviation), and categorical variables were presented as number (percentage). First, we explored the association between E-CDAI and SO based on datasets from the NHANES cohort and the Tongde Hospital of Zhejiang Province cohort. In the NHANES cohort, we conducted analyses following three steps: (I) we performed linear regression and logistic regression analyses with univariate and multivariate models. Model 1 was an unadjusted model. Model 2 was adjusted for age, gender, race, marital status, education level, PIR, smoking status, and PA level. Model 3 was further adjusted for DM, hypertension, and hyperlipidemia history. (II) we evaluated the linear relationship between E-CDAI and risk of SO by restricted cubic spline (RCS) regression. (III) we implemented both sensitivity and subgroup analyses to strengthen the reliability of our data analysis. First, we performed stratified and interaction analyses across various parameters, including gender, BMI, abdominal obesity, race, marital status, education level, smoking status, drinking status, PA level, and histories of DM, hypertension and hyperlipidemia. Then, we employed multiple imputation techniques, to address missing values in variables such as marital status (0.02%), education level (0.02%), smoking status (0.05%), drinking status (27.83%), PIR (8.59%), DM (1.32%), and hypertension (0.09%). This process involved five replications and utilized the Markov-chain Monte Carlo method within the EmpowerStats procedure. Subsequently, we employed linear regression and logistic regression to validate our findings with data obtained from subjects recruited at Tongde Hospital of Zhejiang Province, and applied RCS regression to further explore the association between E-CDAI and SO. In the second part of present study, the “mediation” package in R was used to conduct mediation analysis evaluating the mediating effects of inflammatory/nutritional indicators on the association between E-CDAI and SO, adjusted by age, gender, marital status, education level, smoking status, drinking status, PA level, DM, hypertension, and hyperlipemia history. All analyses were conducted using R version 4.4.0 (The R Foundation, http://www.R-project.org), along with the EmpowerStats (www.empowerstats.net, X&Y Solutions, Inc., Boston, Mass achusetts) and Zstats 1.0 online platform (www.zstats.net). A P-value of < 0.05 was considered statistically significant. Results Baseline characteristics of study population A total of 4388 subjects from the NHANES cohort and 276 subjects from the Tongde Hospital of Zhejiang Province cohort were enrolled in the present study. In the NHANES cohort, 48.81% were males, and the mean age was 39.21 (0.30) years. The mean E-CDAI score was − 0.06 (0.06). Among the 4388 participants, the mean SI was 0.81 (0.01), and 10.03% were diagnosed with SO. Significant differences were observed between the two groups in terms of gender, age, BMI, WC, waist-to-height ratio (WHtR), relative fat mass (RFM), VFA, SI, SVR, race, education level, drinking status, PA level, histories of DM, hypertension, and hyperlipemia, WBC, neutrophils, lymphocytes, albumin, CRP, NPAR, energy intake, and E-CDAI (Table S1, P < 0.05). In the Chinese population cohort, 51.81% were males, and the mean age was 42.98 (12.49) years. The mean E-CDAI score was − 0.10 (1.57). Among the 276 OAO participants, the mean ASMI was 7.23 (1.43) kg/m2 and 35.87% were diagnosed with SO. Significant differences were observed between the two groups in terms of gender, age, BMI, WC, WHtR, PBF, ASMI, marital status, history of hypertension, energy intake, and E-CDAI (Table S2, P < 0.05). Associations of E-CDAI with body composition index The associations of the E-CDAI with body composition index and were evaluated (Table 1). First, we explored the relationship between the E-CDAI and body composition index in the NHANES cohort, Our findings revealed the positive association of E-CDAI with SI (Model 3: β = 0.02, 95% CI = 0.01, 0.02, P < 0.001) in the fully adjusted model. In contrast, there were negative association of E-CDAI with BMI (Model 3: β =−0.25, 95% CI =−0.39, −0.11, P = 0.004), WC (Model 3: β =−0.64, 95% CI =−0.99, −0.29, P = 0.004), WHtR (Model 3: β =−0.01, 95% CI =−0.99, −0.01, P = 0.002), RFM (Model 3: β =−0.28, 95% CI =−0.41, −0.14, P = 0.002), and VFA (Model 3: β =−2.33, 95% CI =−3.62, −1.04, P = 0.005) in the fully adjusted model. Table 1The association between E-CDAI and body composition indexVariablesModel 1β (95% CI) P Model 2β (95% CI) P Model 3β (95% CI) P The NHANES cohort SI0.03 (0.02, 0.03)< 0.0010.01 (0.01, 0.01)< 0.0010.01 (0.01, 0.01)< 0.001 SVR0.02 (0.01, 0.02)< 0.0010.02 (0.01, 0.02)< 0.0010.02 (0.01, 0.02)< 0.001 BMI−0.25 (−0.35, −0.15)< 0.001−0.30 (−0.45, −0.15)0.002−0.25 (−0.39, −0.11)0.004 WC−0.47 (−0.73, −0.21)0.001−0.77 (−1.15, −0.39)0.001−0.64 (−0.99, −0.29)0.004 WHtR−0.01 (−0.99, −0.01)< 0.001−0.01 (−0.99, −0.01)< 0.001−0.01 (−0.99, −0.01)0.002 RFM−0.96 (−1.13, −0.79)< 0.001−0.32 (−0.47, −0.17)0.001−0.28 (−0.41, −0.14)0.002 VFA−1.79 (−2.91, −0.66)0.004−2.79 (−4.26, −1.31)0.002−2.33 (−3.62, −1.04)0.005The Chinese cohort ASMI0.34 (0.24, 0.44)< 0.0010.24 (0.15, 0.32)< 0.0010.24 (0.16, 0.32)< 0.001 SVR0.03 (0.02, 0.04)< 0.0010.02 (0.01, 0.04)< 0.0010.02 (0.01, 0.04)< 0.001 BMI0.40 (0.06, 0.74)0.0210.31 (−0.00, 0.63)0.0530.22 (−0.01, 0.45)0.063 WC1.18 (0.48, 1.88)0.0010.70 (0.04, 1.35)0.0370.33 (−0.11, 0.77)0.140 WHtR−0.00 (−0.01, 0.00)0.097−0.00 (−0.01, 0.00)0.067−0.01 (−0.99, −0.01)0.047 PBF−1.30 (−1.79, −0.80)< 0.001−0.81 (−1.18, −0.43)< 0.001−0.77 (−1.14, −0.41)< 0.001 VFA−2.66 (−6.38, 1.07)0.163−2.15 (−5.84, 1.53)0.254−1.92 (−5.48, 1.64)0.291In the NHANES cohort, Model 1, non-adjusted model. Model 2, adjusted for age, gender, race, marital status, education level, PIR, smoking status, drinking status, and PA level. Model 3, adjusted for adjusted for age, gender, race, marital status, education level, PIR, smoking status, drinking status, PA level, DM, hypertension, and hyperlipemia history. In the Chinese cohort, Model 1, non-adjusted model. Model 2, adjusted for age, gender, marital status, education level, smoking status, drinking status, and PA level. Model 3, adjusted for adjusted for age, gender, marital status, education level, smoking status, drinking status, PA level, DM, hypertension, and hyperlipemia history. NHANES, National Health and Nutrition Examination Survey; E-CDAI, energy-adjusted composite dietary antioxidant index; PIR, family poverty-to-income ratio; PA, physical activity; BMI, body mass index; WC, waist circumference; WHtR, waist to height ratio; RFM, relative fat mass; PBF, percent body fat; VFA, visceral fat area; ASMI, appendicular skeletal muscle mass index; SVR, skeletal muscle mass to visceral area ratio; DM, diabetes mellitus; CI, confidence interval The association between E-CDAI and body composition index In the NHANES cohort, Model 1, non-adjusted model. Model 2, adjusted for age, gender, race, marital status, education level, PIR, smoking status, drinking status, and PA level. Model 3, adjusted for adjusted for age, gender, race, marital status, education level, PIR, smoking status, drinking status, PA level, DM, hypertension, and hyperlipemia history. In the Chinese cohort, Model 1, non-adjusted model. Model 2, adjusted for age, gender, marital status, education level, smoking status, drinking status, and PA level. Model 3, adjusted for adjusted for age, gender, marital status, education level, smoking status, drinking status, PA level, DM, hypertension, and hyperlipemia history. NHANES, National Health and Nutrition Examination Survey; E-CDAI, energy-adjusted composite dietary antioxidant index; PIR, family poverty-to-income ratio; PA, physical activity; BMI, body mass index; WC, waist circumference; WHtR, waist to height ratio; RFM, relative fat mass; PBF, percent body fat; VFA, visceral fat area; ASMI, appendicular skeletal muscle mass index; SVR, skeletal muscle mass to visceral area ratio; DM, diabetes mellitus; CI, confidence interval In the study subjects recruited from Tongde Hospital of Zhejiang Province cohort, there also existed positive association between of E-CDAI with ASMI and SVR, and negative association of E-CDAI with WHtR, and PBF in the unadjusted and multivariable-adjusted models (Table 1). Association between E-CDAI and SO We further explored the relationship between E-CDAI and SO (Table 2). In the NHANES cohort, a significant inverse association between E-CDAI and SO was observed in fully adjusted model (Model 3), with an odds ratio (OR) of 0.89 (95% CI = 0.81, 0.98, P = 0.041). We also converted CDAI into quintiles, and categorical CDAI still displayed significant correlation with SO. The categorical E-CDAI existed significant correlation with SO in the fully adjusted models (Model 3: ORQ5vsQ1 =0.44, 95% CI = 0.29, 1.18, Ptrend =0.050). In the RCS regression analysis (Fig. 2A), we found that the association between E-CDAI and risk of probable SO was monotonically decreasing. Table 2Association between E-CDAI and SOVariablesThe NHANES cohortThe Chinese cohortModel 1OR (95% CI), PModel 2OR (95% CI), PModel 3OR (95% CI), PModel 1OR (95% CI), PModel 2OR (95% CI), PModel 3OR (95% CI), PE-CDAI continuous0.86 (0.82, 0.91), < 0.0010.87 (0.79, 0.96), 0.0180.89 (0.81, 0.98), 0.0410.66 (0.55, 0.59), < 0.0010.72 (0.59, 0.87), < 0.0010.72 (0.60, 0.87), < 0.001E-CDAI quintiles Q1Ref.Ref.Ref.Ref.Ref.Ref. Q20.62 (0.44, 0.88), 0.0120.79 (0.50, 1.26), 0.3470.85 (0.53, 1.38), 0.5310.69 (0.33, 1.47), 0.3400.72 (0.32, 1.64), 0.4360.79 (0.34, 1.85), 0.590 Q30.65 (0.45, 0.93), 0.0260.65 (0.40, 1.05), 0.1090.71 (0.44, 1.16), 0.2080.29 (0.13, 0.65), 0.0020.33 (0.14, 0.76), 0.0100.34 (0.14, 0.80), 0.014 Q40.47 (0.32, 0.68), < 0.0010.39 (0.26, 0.58), < 0.0010.44 (0.30, 0.66), 0.0040.18 (0.08, 0.42), < 0.0010.23 (0.09, 0.55), 0.0010.23 (0.09, 0.57), 0.002 Q50.45 (0.31 0.65), < 0.0010.53 (0.27, 1.04), 0.0940.44 (0.29, 1.18), 0.1710.22 (0.10, 0.49), < 0.0010.29 (0.12, 0.70), 0.0060.30 (0.13, 0.74), 0.009P for trend< 0.0010.0210.050< 0.001< 0.001< 0.001The quintile cut-off values of the E-CDAI are − 1.33, −0.30, 0.45, and 1.24 in the NHANES cohort. The quintile cut-off values of the E-CDAI are − 1.35, −0.57, 0.25, and 1.31 in the Chinese cohort. In the NHANES cohort, Model 1, non-adjusted model. Model 2, adjusted for age, gender, race, marital status, education level, PIR, smoking status, drinking status, and PA level. Model 3, adjusted for adjusted for age, gender, race, marital status, education level, PIR, smoking status, drinking status, PA level, DM, hypertension, and hyperlipemia history. In the Chinese cohort, Model 1, non-adjusted model. Model 2, adjusted for age, gender, marital status, education level, smoking status, drinking status, and PA level. Model 3, adjusted for adjusted for age, gender, marital status, education level, smoking status, drinking status, PA level, DM, hypertension, and hyperlipemia history. NHANES, National Health and Nutrition Examination Survey; E-CDAI, energy-adjusted composite dietary antioxidant index; SO, sarcopenic obesity; PIR, family poverty-to-income ratio; PA, physical activity; DM, diabetes mellitus; OR, odd ratio; CI, confidence interval; Q, quintile Association between E-CDAI and SO The quintile cut-off values of the E-CDAI are − 1.33, −0.30, 0.45, and 1.24 in the NHANES cohort. The quintile cut-off values of the E-CDAI are − 1.35, −0.57, 0.25, and 1.31 in the Chinese cohort. In the NHANES cohort, Model 1, non-adjusted model. Model 2, adjusted for age, gender, race, marital status, education level, PIR, smoking status, drinking status, and PA level. Model 3, adjusted for adjusted for age, gender, race, marital status, education level, PIR, smoking status, drinking status, PA level, DM, hypertension, and hyperlipemia history. In the Chinese cohort, Model 1, non-adjusted model. Model 2, adjusted for age, gender, marital status, education level, smoking status, drinking status, and PA level. Model 3, adjusted for adjusted for age, gender, marital status, education level, smoking status, drinking status, PA level, DM, hypertension, and hyperlipemia history. NHANES, National Health and Nutrition Examination Survey; E-CDAI, energy-adjusted composite dietary antioxidant index; SO, sarcopenic obesity; PIR, family poverty-to-income ratio; PA, physical activity; DM, diabetes mellitus; OR, odd ratio; CI, confidence interval; Q, quintile RCS regression analysis for E-CDAI and SO in the NHANES cohort. (A) Pre-imputation; (B) Imputation 1; (C) Imputation 2; (D) Imputation 3; (E) Imputation 4; (F) Imputation 5. Adjusted for adjusted for age, gender, race, marital status, education level, PIR, smoking status, drinking status, PA level, DM, hypertension, and hyperlipemia history. RCS, restricted cubic spline; E-CDAI, energy-adjusted composite dietary antioxidant index; SO, sarcopenic obesity; PIR, family poverty-to-income ratio; PA, physical activity; DM, diabetes mellitus To ensure the robustness of our findings, we used gender, BMI, abdominal obesity, race, marital status, education level, drinking status, PA level, DM, hypertension, and hyperlipidemia history as stratification variables, and performed stratified analysis to evaluate the association between E-CDAI and SO in stratified populations. As shown Table 3, the above factors were not interactive factors in stratified populations (all P for interaction > 0.05). Then, we examined the association between E-CDAI and SO using imputed data. The findings indicated no significant differences between the datasets before and after imputation (Fig. 2, B-F). Table 3Subgroup analysis between E-CDAI and SO in the NHANES cohortVariablesn (%)OR (95% CI) P P for interactionAll patients4388 (100.00)0.89 (0.81, 0.98)0.041Gender0.618 Male2099 (47.84)0.87 (0.73, 1.02)0.116 Female2289 (52.16)0.92 (0.81, 1.03)0.171BMI0.748 < 25.01320 (30.08)0.99 (0.94, 1.03)0.629 ≥ 25.0, < 30.01364 (31.08)0.85 (0.68, 1.05)0.158 ≥ 30.01704 (38.83)0.91 (0.82, 1.02)0.137Abdominal obesity0.954 Yes2263 (51.57)0.91 (0.82, 1.01)0.089 No2125 (48.43)0.84 (0.69, 1.02)0.098Race/ethnicity0.021 Non-Hispanic White1356 (30.90)0.80 (0.69, 0.93)0.014 Non-Hispanic Black862 (19.64)0.89 (0.75, 1.07)0.268 Other Race2170 (49.45)0.98 (0.88, 1.09)0.707Education level0.742 Less than High School763 (17.39)0.95 (0.74, 1.22)0.702 High School or Equivalent981 (22.36)0.90 (0.77, 1.04)0.181 Some College or AA degree1453 (33.12)0.91 (0.78, 1.06)0.257 College Graduate or above1190 (27.13)0.80 (0.62, 1.03)0.109Marital status0.279 Married/Living with partner2689 (61.29)0.86 (0.74, 1.00)0.070 Widowed/Divorced/Separated598 (13.63)0.95 (0.74, 1.22)0.690 Never married1100 (25.07)0.91 (0.77, 1.08)0.290Smoking status0.038 Never2741 (62.49)0.80 (0.70, 0.90)0.003 Former737 (16.80)0.97 (0.84, 1.13)0.735 Current908 (20.70)0.99 (0.79, 1.28)0.982Drinking0.606 Mild1396 (44.08)0.92 (0.79, 1.08)0.333 Moderate1169 (36.91)0.87 (0.74, 1.02)0.119 Heavy602 (19.01)0.95 (0.75, 1.19)0.645PA status0.878 Low PA1289 (29.38)0.89 (0.75, 1.07)0.237 High PA3099 (70.62)0.90 (0.81, 1.02)0.115Hypertension0.082 Yes1245 (28.40)0.77 (0.63, 0.94)0.027 No3139 (71.60)0.98 (0.88, 1.08)0.651DM0.077 Yes527 (12.17)0.78 (0.64, 0.95)0.037 No3803 (87.83)0.95 (0.85, 1.07)0.442Hyperlipemia0.613 Yes2705 (61.65)0.91 (0.81, 1.01)0.097 No1683 (38.35)0.84 (0.73, 0.97)0.038Adjusted for adjusted for age, gender, marital status, education level, smoking status, drinking status, PA level, DM, hypertension, and hyperlipemia history. The model is not adjusted for the stratification variable. E-CDAI, energy-adjusted composite dietary antioxidant index; SO, sarcopenic obesity; PIR, family poverty-to-income ratio; PA, physical activity; BMI, body mass index; DM, diabetes mellitus; OR, odd ratio; CI, confidence interval Subgroup analysis between E-CDAI and SO in the NHANES cohort Adjusted for adjusted for age, gender, marital status, education level, smoking status, drinking status, PA level, DM, hypertension, and hyperlipemia history. The model is not adjusted for the stratification variable. E-CDAI, energy-adjusted composite dietary antioxidant index; SO, sarcopenic obesity; PIR, family poverty-to-income ratio; PA, physical activity; BMI, body mass index; DM, diabetes mellitus; OR, odd ratio; CI, confidence interval Subsequently, we investigated the association between E-CDAI and SO in the Chinese population cohort. The findings demonstrated significant correlations between the E-CDAI and SO, which were evident in both unadjusted and multivariable-adjusted models (Model 3, ORcontinuous =0.72, 95% CI = 0.60, 0.87, P < 0.001; ORQ5vsQ1 =0.30, 95% CI = 0.13, 0.74, Ptrend <0.001; Table 2). In the RCS regression analysis (Fig. 3), the result showed that the association between E-CDAI and risk of probable SO also was monotonically decreasing. The stratified and interaction analyses across various parameters, including gender, abdominal obesity, education level, marital status, drinking status, PA level, and hyperlipemia showed there were no interactive factors in stratified populations (all P for interaction > 0.05, Table 4). All findings further emphasized the robustness and reliability of the observed trend: as E-CDAI scores increased, the risk of SO correspondingly decreased. RCS regression analysis for E-CDAI and SO in the Chinese cohort Adjusted for adjusted for age, gender, marital status, education level, smoking status, drinking status, PA level, DM, hypertension, and hyperlipemia history. RCS, restricted cubic spline; E-CDAI, composite dietary antioxidant index; SO, sarcopenic obesity; PA, physical activity; DM, diabetes mellitus Table 4Subgroup analysis between E-CDAI and SO in the Chinese cohortVariablesn (%)OR (95% CI) P P for interactionAll patients276 (100.00)0.72 (0.60, 0.88)< 0.001Gender0.911 Male143 (51.81)0.75 (0.53, 1.04)0.087 Female133 (48.19)0.73 (0.57, 0.93)0.012Abdominal obesity0.407 Yes169 (61.23)0.75 (0.59, 0.96)0.019 No107 (38.77)0.66 (0.41, 1.06)0.085Education level0.363 High School or below99 (35.87)0.53 (0.36, 0.77)< 0.001 College or above177 (64.13)0.76 (0.59, 0.97)0.027Marital status0.985 Couple221 (80.07)0.72 (0.59, 0.88)0.002 Single or separated55 (19.93)0.66 (0.33, 1.32)0.235Drinking0.753 Yes211 (76.45)0.72 (0.58, 0.89)0.003 No65 (23.55)0.80 (0.49, 1.32)0.381PA status0.352 Low PA217 (78.62)0.69 (0.55, 0.87)0.001 High PA59 (21.38)0.92 (0.61, 1.38)0.670Hyperlipemia0.323 Yes204 (73.91)0.77 (0.61, 0.95)0.018 No72 (26.09)0.52 (0.32, 0.83)0.007Adjusted for adjusted for age, gender, race, marital status, education level, PIR, smoking status, drinking status, PA level, DM, hypertension, and hyperlipemia history. The model is not adjusted for the stratification variable. E-CDAI, energy-adjusted composite dietary antioxidant index; SO, sarcopenic obesity; PIR, family poverty-to-income ratio; PA, physical activity; DM, diabetes mellitus; OR, odd ratio; CI, confidence interval Subgroup analysis between E-CDAI and SO in the Chinese cohort Adjusted for adjusted for age, gender, race, marital status, education level, PIR, smoking status, drinking status, PA level, DM, hypertension, and hyperlipemia history. The model is not adjusted for the stratification variable. E-CDAI, energy-adjusted composite dietary antioxidant index; SO, sarcopenic obesity; PIR, family poverty-to-income ratio; PA, physical activity; DM, diabetes mellitus; OR, odd ratio; CI, confidence interval Mediating role of inflammatory/nutritional indicators In the NHANES cohort, the associations of inflammatory and nutritional indicators with E-CDAI and SO were demonstrated in Table S3. After fully adjustment, E-CDAI maintained significant associations with neutrophils (β =−0.04, 95% CI =−0.08, −0.01, P = 0.049), CRP (β =−0.30, 95% CI =−0.45, −0.15, P = 0.002), albumin (β = 0.10, 95% CI = 0.01, 0.18, P = 0.046), and NPAR (β = −0.10, 95% CI =−0.17, −0.02, P = 0.020). Results from logistic regression analysis displayed that neutrophiles, CRP, albumin, and NPAR were significantly associated with SO. Then, we explored the mediation effect of inflammatory/nutritional indicators of the association between E-CDAI and SO. The mediation analysis results indicated that NPAR mediated 12.812% of association between E-CDAI and SO, while CRP mediated 6.485% (Fig. 4, Table S4). Analysis of the mediation by inflammatory/nutritional indicators of the association between E-CDAI and SO. (A) mediation by NPAR, (B) mediation by CRP. Adjusted for adjusted for age, gender, race, marital status, education level, PIR, smoking status, drinking status, PA level, DM, hypertension, and hyperlipemia history. NPAR, neutrophil-percentage-to-albumin ratio; CRP, C-reactive protein; E-CDAI, energy-adjusted composite dietary antioxidant index; SO, sarcopenic obesity; CI, confidence interval Discussion This study is the first to explore the association between E-CDAI and SO in U.S. and Chinese populations. The results demonstrated a significant positive correlation between reduced E-CDAI scores and elevated SO risk across both study cohorts. Further investigation into potential mediation mechanisms revealed that inflammatory and nutritional profiles partially account for these established connections. Os has been widely acknowledged as a key factor in triggering skeletal muscle atrophy and fat deposition [24–26]. The progression of skeletal muscle atrophy involves disrupted oxidative-antioxidative equilibrium, with Os serving as a pathological driver of oxidation-reduction homeostasis collapse [27]. OS also could initiate and promote obesity through mitochondrial redox imbalance, cytokine network dysregulation in adipogenesis, ectopic lipid accumulation, adipose precursor differentiation, and central energy homeostasis circuit remodeling [28]. Moreover, muscle progenitor cells may differentiate into an adipocyte-like phenotype as a result of paracrine signaling from cytokines in adipose tissue (AT) [6, 29, 30]. This differentiation results in increased fatty infiltration and decreased muscular regeneration, thereby perpetuating a vicious cycle [6]. The interaction between fatty infiltration and muscle loss could potentially initiate and exacerbate SO [8, 30]. Dietary antioxidants may protect the AT and skeletal muscle from the injury caused by Os. Previous studies have elucidated the mechanistic pathways and therapeutic efficacy of macronutrients and phyto- chemical antioxidants in modulating Os-mediated pathogenesis of sarcopenia and obesity [28, 31]. Therefore, it is necessary to explore the correlation between dietary antioxidant components and SO to develop dietary treatment strategies for managing SO. The E-CDAI is a validated scoring system developed to evaluate the antioxidant capacity of individual dietary patterns [17]. This index was formulated based on the synergistic anti-inflammatory properties of dietary components [18]. Recent investigations into the implications of dietary antioxidant potential on sarcopenia and obesity pathogenesis have attracted considerable attention. Multiple population-based epidemiological studies consistently demonstrate a significant inverse association between higher E-CDAI scores and reduced risks of both obesity and sarcopenia [11–16]. However, research examining E-CDAI’s relationship with SO remains limited. Our cross-sectional analysis utilizing nationally representative data from NHANES and clinical cohort from Zhejiang Province Tongde Hospital revealed a robust positive correlation between diminished E-CDAI scores and elevated SO risk among U.S. and Chinese adult populations. These findings align with previous observations in older U.S. adults [17], while extending the epidemiological significance by demonstrating that this inverse E-CDAI-SO association persists across younger demographic strata. The findings indicated that dietary regimens characterized by elevated E-CDAI scores exhibit therapeutic potential for SO mitigation. Representing integrated antioxidant capacity, the E-CDAI system contains carotenoids, selenium, zinc, vitamin A, vitamin C, and vitamin E [17]. In vitro and in vivo studies demonstrated that nutrient antioxidant components, including carotenoids, selenium, zinc, vitamin A, vitamin C, and vitamin E, could protect against Os and inflammation damages [32–37]. Although these six antioxidant components demonstrate protective roles against Os-driven skeletal muscle atrophy and lipid metabolism dysregulation [38–46], and our study revealed the significant inverse association between E-CDAI and SO, the underlying biological mechanisms mediating these relationships remain unclear. Both inflammation and malnutrition are critical risk factors for SO [3, 4, 8]. Therefore, recognizing the critical roles of Os and inflammatory/nutritional status in the pathogenesis and prevention of SO is imperative. CBC markers, including lymphocytes and neutrophils, are traditional biomarkers used to evaluate immune function. Lymphocytes exhibit a strong link to muscle satellite cells (MUSCs), which play a critical role in skeletal muscle regeneration [47, 48]. Neutrophils, as the first immune cells to infiltrate adipose tissue, could release inflammatory factors and thereby contribute to systemic inflammation [49, 50]. CRP is another widely used clinical indicator that reflects inflammation levels, and previous studies demonstrated that a high level of CRP was independently associated with SO [51, 52]. The NPAR index combines neutrophil percentage and albumin levels [21]. The present study demonstrated that the NPAR and CRP showed negative associations with SO. A mediation analysis in our study identified significant mediating roles of NPAR and CRP indices in the CDAI-SO association, with NPAR for 12.812%, and CRP for 6.485% of the total mediation effect. The primary strength of the present study lies in methodological robustness anchored in dual-cohort validation incorporating NHANES population data (U.S.) and clinical specimens from Zhejiang Tongde Hospital (China). The nationally representative sampling framework of the NHANES cohort ensures epidemiological generalizability, while the Chinese cohort analysis enables a detailed examination of the relationship between E-CDAI and SO. However, the study has several limitations. First, the cross-sectional study design may limit the ability to draw definitive conclusions about the causal relationship between E-CDAI and SO outcomes. Second, the sample size of the clinic cohort was relatively moderate. Third, inherent differences in baseline characteristics between clinic patients and community survey participants introduce confounding effects that may influence the results. Therefore, further research in different populations is necessary for validation. Conclusion A decreased E-CDAI score is associated with the increased risk of SO in adults. This association appears to be partially mediated through inflammatory/nutritional pathways, indicating that the E-CDAI score may serve as a valuable indicator for identifying individuals at risk for SO. Adults may reduce the presence of SO by decreasing the antioxidant capacity of their daily diets. Supplementary Information Supplementary Material 1 Supplementary Material 2 Supplementary Material 3 Acknowledgements The authors are grateful to the participants and the relevant staff of the National Health and Nutrition Examination Survey and Tongde Hospital of Zhejiang Province. Abbreviations Appendicular skeletal muscle mass Appendicular skeletal muscle mass index Adipose tissue Bioelectrical impedance analysis Complete blood count Energy-adjusted composite dietary antioxidant index Confidence interval C-reactive protein Diabetes mellitus Fat mass National Health and Nutrition Evaluation Surveys Neutrophil-percentage-to-albumin ratio Overweight and obesity Odds ratio Oxidative stress Physical activity Percent body fat Poverty-to-income ratio Restricted cubic spline Relative fat mass Sarcopenia index Sarcopenic obesity Skeletal muscle mass to visceral area ratio Visceral fat area white blood cell Waist circumference Waist to height ratio Author contributions The study design was conceived by X.L., X.L., and W-W. J. organized the data, conducted the analyses, and wrote and edited the manuscript. X.L., and W-W. J. contributed to the interpretation of the results, revision, and finalization of the manuscript. All authors have reviewed and approved the final version of the manuscript. Funding This work is supported by Chinese Medicine Research Program of Zhejiang Province (Grant No.2024 ZL349) and Disease Control and Prevention Program of Zhejiang Province (Grant No.2025JK207). Data availability The datasets used and analyzed during the current study available from the corresponding author on reasonable request. Declarations Ethics approval and consent to participate In the NHANES cohort, the study protocol was approved by the National Center for Health Statistics Research Ethics Review Board, and all participants provided their written informed consent to participate in this study. In the Chinese cohort, the study was approved by the Medical Ethics Committee of Tongde Hospital of Zhejiang Province (NO.2023–182 K, NO.2023–183 K), and all participants signed informed consent. The survey was were performed in accordance with the relevant guidelines and regulations. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary Material 1 Supplementary Material 2 Supplementary Material 3 Data Availability Statement The datasets used and analyzed during the current study available from the corresponding author on reasonable request.