Association between low-carbohydrate diet and low-fat diet scores and sarcopenia as determined by dual-energy X-ray absorptiometry: a cross-sectional study Authors: Sitong Xin, Zibo Wu, Yang Xu, Xiaojing Feng, Yuqi Gao, Yu Han, Jing Zhao, Yuangang Guo, Yan Liu, Fengdan Wang, Bo Li, Xinyao Zhang Published: 2025 ABSTRACT ObjectiveThis study aimed to investigate the effects of low-carbohydrate diets (LCD) and low-fat diets (LFD) on sarcopenia, as well as the source and quality of nutrients, to provide strong support for the development of a more scientifically based dietary guidance scheme.MethodsThe data of this study came from the National Health and Nutrition Examination Survey (NHANES). LCD and LFD scores were calculated based on macronutrient energy proportions. Logistic regression, isoenergetic substitution models were used to evaluate associations with sarcopenia, defined by dual-energy X-ray absorptiometry (DXA).ResultsA total of 1,745 participants were included in this study, with a mean age of 38.71 ± 0.46 years and a mean body mass index (BMI) of 28.87 ± 0.26 kg/m². After adjusting for covariates, the highest tertiles of overall LCD and healthful LCD scores were associated with a significantly lower risk of sarcopenia, with odds ratios (OR) of 0.48 (95% CI: 0.24–0.96) and 0.51 (95% CI: 0.32–0.81), respectively, compared with the lowest tertiles. In comparison, the highest tertile of the total LFD was associated with a 78% higher risk of sarcopenia (OR = 1.78; 95% CI: 1.03–3.11). Furthermore, replacing 5% of energy from carbohydrates with protein or fat, and with plant protein or unsaturated fat, was associated with reduced sarcopenia risk, with OR of 0.86 (95% CI: 0.74–0.99) and 0.80 (95% CI: 0.65–0.99), respectively.ConclusionIn this cross-sectional study, higher overall and healthful LCD scores are associated with a lower risk of sarcopenia, whereas higher overall and unhealthful LFD scores are associated with an increased risk. These findings suggest that the relationships between LCD, LFD, and sarcopenia may depend not only on the quantity of macronutrients but also on their dietary sources.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12937-025-01273-8. Introduction Sarcopenia is a progressive decline in muscle mass, strength, and fitness that occurs with age [1]. A meta-analysis indicated that the global prevalence of sarcopenia among people over 60 years old was 10% to 27%, with severe cases accounting for 2% to 9% [2]. Because muscle mass and strength increase in late adolescence and early adulthood but decline significantly after the age of 50, the prevalence of this disease rises considerably with age [3]. Although sarcopenia often occurs in the elderly, with lifestyle changes, sarcopenia also appears to be a younger trend, and sarcopenia that happens in youth would undoubtedly have far-reaching negative impacts on future muscle mass in old age [4]. In addition to aging, a series of well-established risk factors contributed to the development of sarcopenia. These included prolonged physical inactivity, which led to gradual atrophy of muscle tissue and loss of strength; nutritional deficiencies, such as insufficient intake of energy, high-quality protein, vitamin D, and omega-3 fatty acids, which impaired muscle protein synthesis and increased muscle degradation; chronic diseases like diabetes, chronic kidney disease, and cardiovascular disease, which disrupted metabolic balance and induced inflammatory responses that damaged muscle cells; obesity (sarcopenic obesity), where excessive fat accumulation triggered systemic inflammation and insulin resistance, further exacerbating muscle loss [5, 6]. The consequences of sarcopenia should not be ignored, as it could lead to a range of serious consequences such as falls, weakness, cognitive impairment, metabolic disease, fractures, and even death [7]. Studies have also found that sarcopenia was strongly associated with poor progression of liver disease [8] and increased risk of cardiovascular disease in patients with coronary arteries [9]. Therefore, sarcopenia was undoubtedly an important risk factor that affected the health and longevity of the population and deserved our great attention and in-depth research. In recent years, a low-carbohydrate diet (LCD) has become increasingly sought after in weight loss circles, while a low-fat diet (LFD) has also been considered conducive to weight loss [10]. At present, there is no clear consensus on the exact definitions of LCD and LFD. A typical LCD is characterized by a strict limitation of carbohydrates with an increased proportion of proteins and fats [11]. A frequently cited threshold is a carbohydrate intake of less than 40% of total energy. In contrast, an LFD emphasizes strict fat restriction, with a relatively increased intake of carbohydrates and proteins, and is typically defined as a fat intake of less than 30% of total energy [12, 13]. Additionally, studies have shown that obesity may lead to a series of inflammatory reactions through the accumulation of lipids and their derivatives, which in turn can lead to sarcopenia—a condition that has been termed sarcopenic obesity [6]. Therefore, we hypothesized that LCD and LFD might have an impact on the occurrence of sarcopenia. Studies have also indicated that both the quality and quantity of nutrients significantly influence metabolism and health [14]. Thus, we planned to explore not only the quantity of nutrients but also to distinguish between high-quality and low-quality carbohydrates, as well as to classify fats based on their sources as saturated or unsaturated fatty acids. However, to date, there are no studies on the association between LCD and LFD scores and sarcopenia. Therefore, we aim to investigate the association between sarcopenia and the influence of nutrient source and quality. We expect that the results of this study will provide new epidemiological insights into the role of macronutrients in sarcopenia and offer a scientific basis for developing targeted dietary guidance plans in the future. Methods Study sample The National Health and Nutrition Examination Survey (NHANES) is a research program designed to provide an in-depth profile of the health and nutritional status of adults and children in the United States. Led by the Centers for Disease Control and Prevention (CDC), the project collects data covering a wide range of demographics, socioeconomic status, dietary habits, and body composition (bones and soft tissues) through a series of cross-sectional surveys of a nationally representative, non-institutionalized U.S. population. To ensure the timeliness and consistency of the data as well as to minimize measurement error, we have chosen to use only the 2017–2018 data from NHANES, totaling 5,569 participants older than 20 years of age. And 3,328 of them have missing dual-energy X-ray absorptiometry (DXA) data. In addition, we excluded those participants with missing dietary information, energy intake outliers, and covariate data (data on poverty-to-income ratio, smoking status, drinking status, activity level, etc.), and ultimately 1,745 eligible participants were enrolled in this rigorous study. The study’s specific flow is illustrated in Fig. 1.Fig. 1Flow diagram of study sample size Flow diagram of study sample size DXA-assessed sarcopenia DXA scans can accurately measure bone and soft tissue composition throughout the body, including the arms, legs, trunk, and head. It is the gold standard for measuring bone mineral density (BMD), diagnosing osteoporosis, and monitoring BMD [15]. It is now one of the most commonly used techniques for measuring body composition due to its non-invasive nature, low radiation dose, and ease of application to both healthy individuals and patients [16]. In this study, we focused specifically on Appendicular Lean Mass (ALM), which is the sum of the muscle mass of the extremities. It is essential to note that Lean Mass, as used here, does not include Bone Mineral Content (BMC). The determination of sarcopenia is based on the criteria of the Foundation for the National Institutes of Health (FNIH). It is defined by calculating the ratio of ALM to body mass index (BMI) (ALM: BMI), i.e., sarcopenia index = ALM (kg)/BMI (kg/m2). In determining sarcopenia, we considered both gender and sarcopenia index. The presence of sarcopenia was determined if the sarcopenia index was less than 0.789 for men or 0.512 for women [17]. This determination criterion provides us with strong data support for an in-depth analysis of adult muscle health in the United States. LCD and LFD scores We used the NHANES database to analyze in-depth participants’ dietary data, which provided detailed records of energy intake and various nutrients during the 24 h before the interview. Most participants completed two 24-hour dietary interviews, one on-site at a mobile examination center (MEC) on day one and a second interview by phone 3 to 10 days later. The LCD and LFD scoring systems subdivided the nutritional intake of the study population into 11 gender-specific tiers based on the proportion of protein, carbohydrate, and fat in total energy intake. Specific ranges of fat, protein, and carbohydrate energy supply ratios were set for each tier (as shown in Supplementary Table S1, S2). In the LCD score, the lower the carbohydrate energy ratio, the correspondingly higher the score (10 for the highest tier and 0 for the lowest), and conversely, the lower the fat and protein energy ratios, the lower the score. The scores for these three macronutrients are added together to obtain an overall LCD score, which ranges from 0 to 30. Thus, a higher total score indicates that the participant’s diet is closer to the overall LCD pattern [18]. Given the importance of food sources and quality to health, we further subdivided the LCD into healthful LCD scores and unhealthful LCD scores. Both scores were calculated in the same way as the LCD score. The healthful LCD score focuses on the ratio of inferior carbohydrates, vegetable protein, and unsaturated fat (UFA). The higher the ratio of poor-quality carbohydrates to energy, the lower the score (10 in the highest tier and 0 in the lowest tier); Conversely, the higher the ratio of plant proteins and unsaturated fats to energy, the higher the score. In particular, poor-quality carbohydrates are mainly derived from potatoes, other starchy vegetables, fruit juices, and refined grains, while plant proteins are primarily derived from legumes, nuts, seeds, grains, fruits, and vegetables. The unhealthful LCD score, on the other hand, focuses on food sources of quality carbohydrates, animal protein, and saturated fat (SFA). High-quality carbohydrates are primarily derived from non-starchy vegetables, whole fruits, whole grains, and legumes, and animal proteins are mainly derived from dairy products, meat, poultry, fish/seafood, and eggs [14, 18, 19]. Specific food source information was obtained from the Food and Nutrition Database for Dietary Studies (FNDDS) database. LFD scores were constructed similarly to the LCD, taking into account the health effects of different food sources and nutrients, by subdividing the LFD scores into healthful LFD scores and unhealthful LFD scores (as shown in Supplementary Table S1, S2). Covariate assessment Age, gender, race, marital status, educational level, poverty income ratio (PIR), smoking status, drinking status, physical activity, diabetes, hypertension, cardiovascular diseases and calorie intake were included as covariates, with race categorized into two categories: non-Hispanic white, and other race. Marital status was categorized as widowed, divorced or separated, married/living with a partner, and never married. Educational levels were classified as high school or below and above high school. Household incomes were categorized into three groups based on the household poverty income ratio, ≤ 1, >1 to ≤ 3 or >3. Smoking status was defined as a three-categorical variable, with non-smokers defined as those who had never smoked at least 100 cigarettes in their lifetime; former smokers were defined as those who had at least 100 cigarettes but did not smoke now; and current smokers were defined as participants who had at least 100 cigarettes and reported a non-zero number of cigarettes smoked in the past 30 days [20]. Drinking status was categorized as a non-drinker (a person who had < 12 drinks in their lifetime and any one year), former drinker (a person who had ≥ 12 drinks in their lifetime and in any one year but had not consumed alcohol in the past 12 months), and a current drinker (a person who had ≥ 12 drinks in their lifetime and in any one year and had consumed alcohol at least once in the past 12 months) [21]. In terms of physical activity, we determined physical activity based on standard metabolic equivalents (MET) and physical activity time, with (1): total weekly MET-min value ≥ 599 (2): intense physical activity time ≥ 74 min/week (3): moderate physical activity time ≥ 149 min/week, and one of the three being met to be defined as active physical activity [22]. Diabetes was defined as follows: (1): the self-reporting of a diagnosis by a physician or healthcare professional, or (2): a glycosylated hemoglobin A1c (HbA1c) level ≥ 6.5% (47.5 mmol/mol), or (3): fasting plasma glucose (FPG) level ≥ 126 mg/dL (7.0mmol/L). Those who met one of these criteria were diagnosed with diabetes. Individuals with SBP ≥ 140 mmHg and/or DBP ≥ 90 mmHg or participants who answered ‘yes’ to the question ‘Have you ever been told you have high blood pressure?’ were also classified as having high blood pressure [23]. Has a doctor or other healthcare provider ever diagnosed you with any of the following: congestive heart failure, coronary heart disease, angina, heart attack, or stroke? A ‘yes’ response to any of these conditions was considered indicative of cardiovascular disease [24]. Statistical analysis Due to the complexity of the NHANES sampling design, sample weights, clustering, and stratification were incorporated in all analyses. Tertiles were used to describe LCD and LFD scores. Continuous variables were presented by means and standard errors (SE), while categorical variables were presented by unweighted frequencies and weighted percentages. Binary logistic regression was used to assess the odds ratio (OR) and 95% confidence intervals (CI) for the association between tertiles of LCD versus LFD scores and sarcopenia. Model 1 was not adjusted for covariates. Model 2 was adjusted for age, sex, and race. Based on Model 2, Model 3 was further adjusted for marital status, educational level, PIR, physical activity, smoking status, drinking status, diabetes, hypertension, cardiovascular diseases, and calorie intake. We also conducted trend tests, in which the median of each tertile was considered as a continuous variable. In addition, we estimated the OR for sarcopenia with a 5-point increase in each score. After adjusting for potential confounders, the dose-response associations between LCD scores, LFD scores, and sarcopenia were examined using restricted cubic spline modelling. The risk of sarcopenia was estimated using an isoenergetic substitution model, which estimated the risk of sarcopenia per five-percentage-point increase in energy from different types of fat (UFAs versus SFAs) or various protein food sources (plant-derived protein versus animal-derived protein) when substituting these nutrients for carbohydrates (each with a five-percentage-point increase in energy). For example, we used 5% energy protein to replace 5% energy carbohydrate (not included in the model), while keeping energy intake and fat intake (5% energy) constant in the model [25]. Analyses were also stratified according to age (< 40 or ≥ 40 years), sex (male or female), race/ethnicity (non-Hispanic white or other), educational level (high school or below, above high school), PIR (≤ 1, > 1 to ≤ 3 or > 3), physical activity (active or inactive) and diabetes (yes or no). The significance of interactions was estimated using the p-value of the product between each dietary score and the stratified variables. All statistical analyses were performed using IBM SPSS version 24.0 and R 4.3.1 with the “rms” and “forestploter” software packages. Two-sided P-values less than 0.05 were considered statistically significant. Results A total of 1745 participants were included in this study, and the results showed that the prevalence of sarcopenia reached 7.3%. In Table 1, we detailed the baseline characteristics of the participants in the different sarcopenia states. Overall, the age range of the participants was 20–59 years, with a mean age of 38.71 ± 0.46 years. The mean BMI was 28.87 ± 0.26 kg/m². The percentage of females was slightly higher, reaching 51.2%. Patients with sarcopenia were older, more likely to have high blood pressure, cardiovascular diseases and diabetes, more likely to be less physically active, had lower household income to poverty income ratios, lower educational level, lower overall LCD scores, healthful LCD scores, and higher overall LFD scores and unhealthful LFD scores, compared to those without sarcopenia compared to those with non-sarcopenia. However, gender, marital status, smoking status, drinking status, and calorie intake did not show significant differences between sarcopenic and non-sarcopenic individuals.Table 1The characteristics of the participants and association with sarcopeniaVariablesGroupsTotal(N = 1745)SarcopeniaPYesN = 178(7.3%)NoN = 1567(92.7%)Sex0.168Male807(48.8)94(58.8)713(48.0)Female938(51.2)84(41.2)854(52.0)Age(year)38.71 ± 0.4644.29 ± 1.4038.28 ± 0.470.002Race/ethnicity< 0.001Non-Hispanic White564(56.3)47(38.4)517(57.7)Other1181(43.7)131(61.6)1050(42.3)Educational level< 0.001High school or below645(33.8)102(60.0)543(31.8)Above high school1100(66.2)76(40.0)1024(68.2)Marital status0.078Widowed/divorced/separated255(13.1)34(21.7)221(12.5)Married/living with a partner1046(59.9)106(59.3)940(59.9)Never married444(27.0)38(19.0)406(27.6)Physical activity0.004Active physical activity1248(76.8)103(61.5)1145(78.0)Inactive physical activity497(23.2)75(38.5)422(22.0)Poverty income ratio (PIR)0.017≤ 1325(14.3)44(19.8)281(13.9)> 1 to ≤ 3740(35.2)85(47.6)655(34.2)> 3680(50.5)49(32.5)631(52.0)Smoking status0.921Non-smoker1091(61.6)112(60.9)979(61.6)Former smoker304(20.2)31(19.7)273(20.2)Current smoker350(18.3)35(19.4)315(18.2)Drinking status0.051Non-drinker159(6.8)26(10.2)133(6.5)Former drinker236(10.7)32(16.8)204(10.2)Current drinker1350(82.6)120(73.0)1230(83.3)Hypertension0.042Yes509(27.0)78(48.1)431(25.3)No1236(73.0)100(51.9)1136(74.7)Cardiovascular diseases0.002Yes79(3.2)15(7.6)64(2.8)No1666(96.8)163(92.4)1503(97.2)DiabetesYes188(7.5)39(23.4)149(6.3)< 0.001No1557(92.5)139(76.6)1418(93.7)Calorie intake(kcal/day)2075.57 ± 21.251996.49 ± 87.592081.84 ± 24.010.643Overall LCD Score0.018Tertile 1609(31.9)78(46.5)531(30.8)Tertile 2579(33.8)59(31.7)520(33.9)Tertile 3557(34.3)41(21.8)516(35.3)Unhealthful LCD Score0.576Tertile 1609(30.7)68(36.4)541(30.2)Tertile 2576(35.7)55(34.0)521(35.9)Tertile 3560(33.6)55(29.6)505(33.9)Healthful LCD Score0.006Tertile 1635(33.8)78(49.9)557(32.5)Tertile 2551(33.0)55(27.9)496(33.4)Tertile 3559(33.3)45(22.2)514(34.1)Overall LFD Score0.010Tertile 1629(38.6)49(24.1)580(39.7)Tertile 2581(34.1)60(36.8)521(33.9)Tertile 3535(27.3)69(39.1)466(26.4)Unhealthful LFD ScoreTertile 1681(39.6)63(32.1)618(40.2)0.034Tertile 2556(31.2)46(24.2)510(31.8)Tertile 3508(29.2)69(43.7)439(28.0)Healthful LFD Score0.845Tertile 1622(36.9)61(35.1)561(37.0)Tertile 2604(36.8)64(39.6)540(36.6)Tertile 3519(26.3)53(25.3)466(26.4) Table 1The characteristics of the participants and association with sarcopeniaVariablesGroupsTotal(N = 1745)SarcopeniaPYesN = 178(7.3%)NoN = 1567(92.7%)Sex0.168Male807(48.8)94(58.8)713(48.0)Female938(51.2)84(41.2)854(52.0)Age(year)38.71 ± 0.4644.29 ± 1.4038.28 ± 0.470.002Race/ethnicity< 0.001Non-Hispanic White564(56.3)47(38.4)517(57.7)Other1181(43.7)131(61.6)1050(42.3)Educational level< 0.001High school or below645(33.8)102(60.0)543(31.8)Above high school1100(66.2)76(40.0)1024(68.2)Marital status0.078Widowed/divorced/separated255(13.1)34(21.7)221(12.5)Married/living with a partner1046(59.9)106(59.3)940(59.9)Never married444(27.0)38(19.0)406(27.6)Physical activity0.004Active physical activity1248(76.8)103(61.5)1145(78.0)Inactive physical activity497(23.2)75(38.5)422(22.0)Poverty income ratio (PIR)0.017≤ 1325(14.3)44(19.8)281(13.9)> 1 to ≤ 3740(35.2)85(47.6)655(34.2)> 3680(50.5)49(32.5)631(52.0)Smoking status0.921Non-smoker1091(61.6)112(60.9)979(61.6)Former smoker304(20.2)31(19.7)273(20.2)Current smoker350(18.3)35(19.4)315(18.2)Drinking status0.051Non-drinker159(6.8)26(10.2)133(6.5)Former drinker236(10.7)32(16.8)204(10.2)Current drinker1350(82.6)120(73.0)1230(83.3)Hypertension0.042Yes509(27.0)78(48.1)431(25.3)No1236(73.0)100(51.9)1136(74.7)Cardiovascular diseases0.002Yes79(3.2)15(7.6)64(2.8)No1666(96.8)163(92.4)1503(97.2)DiabetesYes188(7.5)39(23.4)149(6.3)< 0.001No1557(92.5)139(76.6)1418(93.7)Calorie intake(kcal/day)2075.57 ± 21.251996.49 ± 87.592081.84 ± 24.010.643Overall LCD Score0.018Tertile 1609(31.9)78(46.5)531(30.8)Tertile 2579(33.8)59(31.7)520(33.9)Tertile 3557(34.3)41(21.8)516(35.3)Unhealthful LCD Score0.576Tertile 1609(30.7)68(36.4)541(30.2)Tertile 2576(35.7)55(34.0)521(35.9)Tertile 3560(33.6)55(29.6)505(33.9)Healthful LCD Score0.006Tertile 1635(33.8)78(49.9)557(32.5)Tertile 2551(33.0)55(27.9)496(33.4)Tertile 3559(33.3)45(22.2)514(34.1)Overall LFD Score0.010Tertile 1629(38.6)49(24.1)580(39.7)Tertile 2581(34.1)60(36.8)521(33.9)Tertile 3535(27.3)69(39.1)466(26.4)Unhealthful LFD ScoreTertile 1681(39.6)63(32.1)618(40.2)0.034Tertile 2556(31.2)46(24.2)510(31.8)Tertile 3508(29.2)69(43.7)439(28.0)Healthful LFD Score0.845Tertile 1622(36.9)61(35.1)561(37.0)Tertile 2604(36.8)64(39.6)540(36.6)Tertile 3519(26.3)53(25.3)466(26.4) The characteristics of the participants and association with sarcopenia LCD, LFD scores, and sarcopenia The univariate and multivariate analyses of LCD and LFD scores, as well as sarcopenia, are presented in Table 2. To validate the effect of different scoring levels on sarcopenia, we categorized the scores into tertiles. In Model 2, participants in the highest tertile of unhealthful LFD scores showed increased odds of sarcopenia compared to those in the lowest tertile (OR: 1.81; 95% CI: 1.11–2.94). Similarly, Model 3 revealed that both the second (OR: 1.71; 95% CI: 1.05–2.77) and third tertiles (OR: 1.78; 95% CI: 1.03–3.11) of overall LFD scores were associated with elevated sarcopenia risk relative to the lowest tertile, suggesting that higher unhealthful LFD and overall LFD scores may increase sarcopenia risk. In contrast, the highest tertiles of both overall LCD (OR: 0.48; 95% CI: 0.24–0.96) and healthful LCD scores (OR: 0.51; 95% CI: 0.32–0.81) demonstrated protective effects against sarcopenia compared to their respective lowest tertiles, with significant linear trends (P for trend < 0.05). Unhealthful LCD scores and healthful LFD scores showed no significant association with sarcopenia risk. The detailed OR values, 95% CI values and trend test P values of each dietary score quantile in Model 3 were summarized in Supplementary Table S3, which provides a concise reference for comparing the association between different dietary patterns and the risk of sarcopenia after adjustment for all covariates. Furthermore, restricted cubic spline analyses (Fig. 2) revealed no evidence of non-linear relationships between these dietary scores and sarcopenia (P for nonlinearity > 0.05). The RCS curves for each dietary pattern were shown in the Supplementary Figures S1-S6.Table 2Logistic regression models for LCD scores and LFD scores with sarcopeniaTertiles of diet scoresP for trendPer Five-Point IncreaseTertile 1Tertile 2Tertile 3Overall LCD scoreMedian score (IQR)7 (4–9)15 (14–17)24 (21–26) Model 1aReference0.62 (0.33, 1.16)0.41 (0.21, 0.78)0.0110.77 (0.63,0.93) Model 2bReference0.71 (0.37, 1.35)0.45 (0.25, 0.83)0.0160.79 (0.66,0.95) Model 3cReference0.76 (0.39,1.48)0.48 (0.24,0.96)0.0420.81 (0.66,0.99)Unhealthful LCD scoreMedian score (IQR)8 (5–11)16 (14–17)22 (20–24) Model 1Reference0.79 (0.36, 1.75)0.73 (0.33, 1.61)0.4160.89 (0.66,1.20) Model 2Reference0.87 (0.41,1.86)0.84 (0.39, 1.80)0.6320.94 (0.71,1.25) Model 3Reference0.90 (0.43,1.88)0.86 (0.43,1.74)0.6660.95(0.73,1.23)Healthful LCD scoreMedian score (IQR)9 (6–11)15 (14–17)22 (20–24) Model 1Reference0.54 (0.29,1.01)0.42 (0.27, 0.67)0.0020.71 (0.58,0.86) Model 2Reference0.57 (0.32, 1.02)0.44 (0.30, 0.65)0.0010.72 (0.61,0.85) Model 3Reference0.61 (0.32,1.15)0.51 (0.32, 0.81)0.0120.76 (0.62,0.93)Overall LFD scoreMedian score (IQR)9 (7–11)16 (14–17)21 (20–23) Model 1Reference1.79 (1.32, 2.42)2.45 (1.41, 4.26)0.0041.45 (1.15,1.83) Model 2Reference1.65 (1.17, 2.32)1.95 (1.22, 3.13)0.0071.32 (1.09,1.60) Model 3Reference1.71 (1.05,2.77)1.78 (1.03,3.11)0.0331.28(1.02,1.59)Unhealthful LFD scoreMedian score (IQR)10 (8–12)16 (15–17)21 (20–23) Model 1Reference0.97 (0.62, 1.54)1.95 (1.14, 3.36)0.0301.36 (1.04,1.79) Model 2Reference1.00 (0.60, 1.68)1.81 (1.11, 2.94)0.0281.31 (1.03,1.67) Model 3Reference1.01 (0.62,1.65)1.68 (0.98, 2.87)0.0711.27(0.98,1.65)Healthful LFD scoreMedian score (IQR)8 (5–10)15 (13–17)23 (21–26) Model 1Reference1.14 (0.57, 2.28)1.01 (0.58, 1.78)0.9351.01 (0.84,1.21) Model 2Reference1.02 (0.49, 2.11)0.82 (0.47, 1.43)0.4820.94 (0.78,1.13) Model 3Reference1.02 (0.53,1.96)0.81 (0.50, 1.33)0.3950.94 (0.80,1.10)Note: a-Model1: no adjustments were madeb-Model2: sex, age, race/ethnicityc-Model3: Model2, marital status, PIR, and education level, physical activity, smoking status, drinking status, diabetes, hypertension, cardiovascular diseases and calorie intakeFig. 2The dose–response relationships of overall, healthful, and unhealthful LCD (A–C) and overall, healthful, and unhealthful LFD (D–F) with sarcopenia in the United States. Covariates adjusted in the models were the same as those in model 3 in Table 2. The solid lines represent the estimated OR, and the shaded areas indicate the 95% confidence intervals Table 2Logistic regression models for LCD scores and LFD scores with sarcopeniaTertiles of diet scoresP for trendPer Five-Point IncreaseTertile 1Tertile 2Tertile 3Overall LCD scoreMedian score (IQR)7 (4–9)15 (14–17)24 (21–26) Model 1aReference0.62 (0.33, 1.16)0.41 (0.21, 0.78)0.0110.77 (0.63,0.93) Model 2bReference0.71 (0.37, 1.35)0.45 (0.25, 0.83)0.0160.79 (0.66,0.95) Model 3cReference0.76 (0.39,1.48)0.48 (0.24,0.96)0.0420.81 (0.66,0.99)Unhealthful LCD scoreMedian score (IQR)8 (5–11)16 (14–17)22 (20–24) Model 1Reference0.79 (0.36, 1.75)0.73 (0.33, 1.61)0.4160.89 (0.66,1.20) Model 2Reference0.87 (0.41,1.86)0.84 (0.39, 1.80)0.6320.94 (0.71,1.25) Model 3Reference0.90 (0.43,1.88)0.86 (0.43,1.74)0.6660.95(0.73,1.23)Healthful LCD scoreMedian score (IQR)9 (6–11)15 (14–17)22 (20–24) Model 1Reference0.54 (0.29,1.01)0.42 (0.27, 0.67)0.0020.71 (0.58,0.86) Model 2Reference0.57 (0.32, 1.02)0.44 (0.30, 0.65)0.0010.72 (0.61,0.85) Model 3Reference0.61 (0.32,1.15)0.51 (0.32, 0.81)0.0120.76 (0.62,0.93)Overall LFD scoreMedian score (IQR)9 (7–11)16 (14–17)21 (20–23) Model 1Reference1.79 (1.32, 2.42)2.45 (1.41, 4.26)0.0041.45 (1.15,1.83) Model 2Reference1.65 (1.17, 2.32)1.95 (1.22, 3.13)0.0071.32 (1.09,1.60) Model 3Reference1.71 (1.05,2.77)1.78 (1.03,3.11)0.0331.28(1.02,1.59)Unhealthful LFD scoreMedian score (IQR)10 (8–12)16 (15–17)21 (20–23) Model 1Reference0.97 (0.62, 1.54)1.95 (1.14, 3.36)0.0301.36 (1.04,1.79) Model 2Reference1.00 (0.60, 1.68)1.81 (1.11, 2.94)0.0281.31 (1.03,1.67) Model 3Reference1.01 (0.62,1.65)1.68 (0.98, 2.87)0.0711.27(0.98,1.65)Healthful LFD scoreMedian score (IQR)8 (5–10)15 (13–17)23 (21–26) Model 1Reference1.14 (0.57, 2.28)1.01 (0.58, 1.78)0.9351.01 (0.84,1.21) Model 2Reference1.02 (0.49, 2.11)0.82 (0.47, 1.43)0.4820.94 (0.78,1.13) Model 3Reference1.02 (0.53,1.96)0.81 (0.50, 1.33)0.3950.94 (0.80,1.10)Note: a-Model1: no adjustments were madeb-Model2: sex, age, race/ethnicityc-Model3: Model2, marital status, PIR, and education level, physical activity, smoking status, drinking status, diabetes, hypertension, cardiovascular diseases and calorie intake Logistic regression models for LCD scores and LFD scores with sarcopenia Note: a-Model1: no adjustments were made b-Model2: sex, age, race/ethnicity c-Model3: Model2, marital status, PIR, and education level, physical activity, smoking status, drinking status, diabetes, hypertension, cardiovascular diseases and calorie intake The dose–response relationships of overall, healthful, and unhealthful LCD (A–C) and overall, healthful, and unhealthful LFD (D–F) with sarcopenia in the United States. Covariates adjusted in the models were the same as those in model 3 in Table 2. The solid lines represent the estimated OR, and the shaded areas indicate the 95% confidence intervals Isoenergetic substitution model In the substitution model, replacing 5% of carbohydrates with total protein and total fat, plant protein and unsaturated fat was associated with a 14% (OR: 0.86; 95% CI: 0.74–0.99) and 20% (OR: 0.80; 95% CI: 0.65–0.99) reduction in the risk of sarcopenia, respectively. Detailed information is shown in Fig. 3.Fig. 3Association between isoenergetic replacement (as 5% of energy) of carbohydrates with other macronutrients and sarcopenia in the United States. The covariates adjusted in the model are the same as those in Model 3 of Table 2 Association between isoenergetic replacement (as 5% of energy) of carbohydrates with other macronutrients and sarcopenia in the United States. The covariates adjusted in the model are the same as those in Model 3 of Table 2 Stratified analysis Supplementary Figures S7-S10 present the stratified analyses and interaction effects for overall LCD, healthful LCD, overall LFD, and unhealthful LFD scores. Among non-Hispanic whites and diabetic populations, higher overall LCD and healthful LCD scores were significantly associated with reduced sarcopenia risk, with corresponding OR (95% CIs) of 0.19 (0.06, 0.58) and 0.25 (0.11, 0.59) in non-Hispanic whites. Conversely, elevated overall LFD and unhealthful LFD scores showed increased sarcopenia risk, with OR (95% CIs) of 2.89 (1.37, 6.10) and 2.01 (1.01, 2.96), respectively. While most interaction terms were non-significant, a statistically significant interaction was observed between diabetes status and overall LCD score in relation to sarcopenia risk (P for interaction < 0.05), suggesting potential effect modification by diabetes status on the LCD-sarcopenia association. Discussion This study aimed to preliminarily explore the association between overall LCD and LFD scores and sarcopenia. The results of this national cross-sectional study showed that higher overall LCD scores and healthful LCD scores were negatively correlated with a lower prevalence of sarcopenia. However, a higher overall LFD score and an unhealthful LFD score were positively correlated with a higher prevalence of sarcopenia. The results of substitution models, restricted cubic splines, and subgroup analyses further supported the above findings. It is crucial to emphasize that due to the cross-sectional design, these findings indicate an association and cannot establish causality. Some studies have shown that increased protein intake (for the elderly, 1.0 to 1.6 g/kg/day, which should be allocated to 3 to 4 meals daily) is associated with a reduced incidence of sarcopenia [26, 27]. However, a study on elderly people in South Korea revealed an inverse association between carbohydrate intake and sarcopenia risk, wherein an average daily intake below 245.9 g was linked to a greater likelihood of developing the condition [28]. This finding contradicted our research results. This obvious difference might be attributed to the fact that the cited studies focused solely on absolute carbohydrate mass, whereas our analysis considered its proportion of total energy intake. A nutritional study based on a population of children and adolescents suggested that increased fat intake was associated with an elevated risk of sarcopenia, which contrasts with the conclusion of this study that a higher LFD score increases the risk of sarcopenia. The reason for the difference might be the different physiological stages and muscle metabolic characteristics of the study population [29]. Food source quality was found to significantly influence sarcopenia risk [30, 31]. Healthful dietary patterns typically emphasized plant-based nutrients—plant proteins, unsaturated fats, and high-quality carbohydrates from non-starchy vegetables, fruits, and whole grains [32]. For example, the Mediterranean diet (emphasizes the intake of fresh fruits, vegetables, whole grains, nuts, fish and olive oil, while reducing the consumption of red meat and processed foods), with its unique dietary structure and nutritional ratios, was shown in several studies to be positively correlated with higher muscle mass [33–35]. This benefit extended to other culturally specific patterns, as observed in studies of Korean and Hong Kong populations, where diets rich in vegetables, fruits, and quality protein sources were associated with a lower incidence of sarcopenia [36, 37]. Furthermore, the dietary inflammation index (DII)—an important indicator of dietary anti-inflammatory and pro-inflammatory effects—revealed the impact of diet on inflammation by considering the scores of 36 anti-inflammatory components (such as caffeine, ethanol, β-carotene, n-3 fatty acids, n-6 fatty acids, etc.) and 9 pro-inflammatory components (such as sugar, cholesterol, energy, total fat, etc.). The meta-analysis results showed that a 1-point increase in the DII score was associated with a 1.22-fold increase in the risk of sarcopenia [38]. In line with this evidence, our findings demonstrated that a higher healthful LCD score might be protective against muscle loss, whereas a higher unhealthful LFD score appeared to increase the risk. This consistency across studies underscores the pivotal role of nutrient quality, in addition to quantity, in preserving muscle health. The protective effects of high-quality carbohydrates, plant proteins, and unsaturated fatty acids against sarcopenia were mainly achieved through the specific nutrients they were rich in, which improved the core pathological mechanisms of sarcopenia, including chronic inflammation, insulin resistance, mitochondrial dysfunction, and oxidative stress [39, 40]. Specifically, Vitamin D (current dietary therapies advocate for 800 IU of vitamin D twice daily for those with sarcopenia) was known to play an important role in promoting bone re-modelling and enhancing immunity [41]. Insufficient vitamin D intake was shown to reduce the sensitivity of the vitamin D receptor (VDR) to 1,25-dihydroxyvitamin D, thereby impairing intramuscular protein synthesis, reducing type II muscle fiber proportion, and ultimately contributing to sarcopenia [42, 43]. Dietary fiber can increase the diversity of intestinal flora and promote the growth of beneficial bifidobacteria, thus preventing intestinal inflammation [44]. It also served as a major source of short-chain fatty acids (SCFAs)—metabolites produced by anaerobic bacterial fermentation of fiber and resistant starch in the colon. SCFAs were associated with improved insulin sensitivity and anti-inflammatory effects [45]. Flavonoids acted as both inhibitory and regulatory agents: they suppressed the activity of inflammatory cytokines and modulated biological pathways involved in inflammation. They also mitigated the accumulation of nitric oxide (NO) and reactive oxygen species (ROS). Furthermore, the consumption of flavonoid-rich foods helped prevent insulin resistance by modulating the function of pancreatic β-cells, hepatocytes, and adipocytes [46, 47]. Amino acids played a complex role in the occurrence and development of sarcopenia. Studies have shown that branched-chain amino acids and others promote the synthesis of muscle proteins by activating anabolic signaling pathways such as mTOR and regulate the output of motor signals in the central nervous system through their neurotransmitter functions [48, 49]. In addition to amino acids, fatty acids play a crucial role in promoting muscle synthesis and inhibiting muscle metabolism, particularly Omega-3 fatty acids. This group of polyunsaturated fatty acids provided a healthful growth environment for muscles through its anti-inflammatory, antioxidant, and insulin-sensitizing properties [50]. In addition, Omega-3 polyunsaturated fatty acids can enhance muscle cell response to anabolism and metabolism stimulation by affecting related signaling pathways and gene expression, thereby promoting protein synthesis [51]. A randomized controlled study showed that supplementation of n-3 fatty acids (2.1 g EPA + 0.6 g DHA) can effectively improve muscle mass and increase maximal isometric torque [52]. In contrast, saturated fatty acids had an inhibitory effect on skeletal muscle formation. Specifically, research by Heitor O Santos et al. indicated that saturated fatty acids not only increased the expression of atrophy-promoting genes in the body but also inhibited protein synthesis by downregulating the mTOR/p70-S6 kinase 1-dependent signaling pathway, leading to reduced muscle strength and mass [53]. The research in this paper had some strengths and weaknesses. As for the advantages, firstly, this paper used only one cycle of survey data, which largely avoided measurement errors due to changes in investigators, measurement methods, or measurement instruments. Secondly, this paper employed a complex sampling design to enhance the national representativeness of the data. However, this study has several limitations. First, as a cross-sectional study, we could only identify associations between dietary scores and sarcopenia, not establish a causal relationship. Second, dietary intake was assessed via 24-hour recalls, which were subject to measurement error and recall bias. Third, although we adjusted for numerous potential confounders, residual confounding by unmeasured or imperfectly measured factors could not be ruled out. Fourth, the generalizability of our findings might be limited to the U.S. adult population and might not extend to other populations. Fifth, the exclusion of a substantial number of participants due to missing DXA data (n = 3,328) may have introduced selection bias. Finally, the dietary scores studied in this paper could not be directly translated to current popular dietary patterns to assess their risks and harms. Conclusions In a nationally representative sample of U.S. adults, our analysis demonstrated that higher overall and healthful LCD scores were associated with significantly lower odds of sarcopenia. Conversely, a higher overall LFD and an unhealthful LFD score were associated with greater odds of the condition. These findings suggest that the influence of macronutrient patterns on muscle health is substantially dependent on the quality and sources of nutrients. However, given the cross-sectional nature of this study, future longitudinal or interventional research is warranted to confirm these associations and elucidate the underlying causal mechanisms. Supplementary Information Supplementary Material 1. Supplementary Material 1.