Role of sarcopenia, obesity and sarcopenic obesity in frailty, frailty transitions and death Authors: Alejandro Álvarez-Bustos, Jose A Carnicero, Walter Sepúlveda-Loyola, Francisco J Garcia-Garcia, Leocadio Rodríguez-Mañas Published: 2026 ABSTRACT AbstractBackground & aimsThe role of sarcopenic obesity in the transitions of frailty has been largely neglected. The aim of this study is to assess the role of sarcopenic obesity with prevalent frailty, frailty transitions and mortality.MethodsThis is a cohort-based retrospective study including 1538 (74.73 years, 45.51% men) community-dwelling ≥ 65 years. Frailty status (baseline and at follow-up) was evaluated with the Frailty Phenotype (FP) and the Frailty Trait Scale-5 (FTS5). Sarcopenic obesity (SO) status was defined by Body Mass Index ≥ 30 and the Foundation for the National Institutes of Health criteria, standardized to our population. Logistic/multilevel/Poisson regression models were used to assess the role of categories according to the presence of sarcopenia and obesity [Non-sarcopenic, non-obese (NS-NO); Non-sarcopenic, obese (NS-O); Sarcopenic, non-obese (S-NO); Sarcopenic obese (SO)] on frailty, frailty transitions (median follow-up 2.99 years). Mortality (median follow-up 6.29 years) was assessed with Cox proportional hazard models.ResultsCross-sectionally, NS-O, S-NO and SO were significantly associated with frailty across FP and FTS5, compared with the NS-NO category. Longitudinally, S-O was significantly associated with increased rates of frailty progression according to the FP [IRR (95% CI) 1.67 (1.05, 2.68) in robust; 3.35 (1.15, 9.71) in prefrail] and FTS5 [IRR (95% CI): 6.56 (3.68, 11.68)], and reduced the probability of reversing prefrailty status [FP IRR (95% CI): 0.63 (0.41, 0.98)]. S-NO [FP robust, IRR (95% CI) 1.87 (1.17, 2.98); FTS5 non-frail IRR 2.80 (1.46, 5.37)] and NS-O [FTS5 non-frail IRR (95% CI) 2.26 (1.25, 4.09)] also exhibited higher rates of frailty worsening. When sarcopenic obesity was used as the reference category, NS-NO showed a significantly lower incidence of frailty worsening [FP robust IRR 0.60 (0.37, 0.96); FP prefrail IRR 0.30 (0.10, 0.87)], while according to the FTS5, all non–S-O phenotypes were associated with a lower risk of frailty progression, with the strongest associations observed for NS-NO [IRR 0.15 (0.09, 0.27)]. In addition, S-NO and S-O presented a significantly lower likelihood of maintaining robustness or non-frail status across frailty definitions. No category was independently associated with mortality after considering frailty status.ConclusionsSarcopenic obesity is consistently associated with a poor functional evolution in frailty status, but it does not modify the odds of death when frailty status is considered. In addition, it may influence the likelihood of improvement in frailty status among prefrail individuals. Further studies are needed to determine whether obesity provides an additional effect on sarcopenia in shaping frailty transitions.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12877-026-07756-5.Keywords: Frailty transitions, Sarcopenic obesity, Mortality, Community-dwelling Background Several factors have been shown to influence frailty status, being two of them sarcopenia and obesity. Sarcopenia, defined as the presence of low muscle mass and function [1], has itself been shown to predispose the individual to adverse events (such as falls, quality of life [2], disability [3–5] and mortality [3, 5] and to be a potent risk factor for worsening frailty status independently of age, gender or comorbidity [6]. Although frailty and sarcopenia are independent entities [7, 8], it seems that the sarcopenia is one of the most relevant factors predicting an evolution to functional impairment in frail individuals while it does not influence the odds of death associated to frailty [5, 6]. Obesity has been associated in a more controversial way to functional deterioration and death in older people. Concerns about the severity of obesity needed to promote unfavourable frailty transitions, the role of general or abdominal obesity [9], or its association with several outcomes [10] still remain. This lack of agreement regarding the role for sarcopenia and obesity in determining the outcomes of frailty is also the case when both conditions are present jointly in an old person. The so-called sarcopenic obesity, or the presence of low muscle mass and function accompanied by excess adipose tissue [11–13], has been recognized as a relevant clinical condition with important implications for health and treatment [13]. Although current evidence suggests an association between sarcopenic obesity and frailty, this evidence comes mainly from cross-sectional studies [14–16]. However, no study has evaluated the role of sarcopenic obesity in frailty longitudinally, including its role in determining frailty evolution. Likewise, whether the presence of sarcopenic obesity modifies the odds of adverse events, such as mortality, within the same frailty state is also unknown. Therefore, the aim of this study is to evaluate the role of sarcopenia, obesity and sarcopenic obesity in frailty, frailty transitions, as well as in mortality, once frailty status is considered. Methods Data for the present study were gathered from the Toledo Study of Healthy Aging (TSHA, ethics committee approval protocol number 15/07/2010.-93) [17], a longitudinal, prospective, population-based cohort of individuals aged 65 years residing in the province of Toledo, Spain. TSHA study was approved by the Clinical Research Ethics Committee of the Toledo Hospital in Spain 15/07/2010.-93), fulfilled according to the ethical standards defined in the 1964 Declaration of Helsinki, and each participant provided written informed consent form prior to enrolment. TSHA includes three measurements or waves. Since DEXA assessments were not performed in the first wave, in this manuscript, measurements from the second (2011–2013) and the third (2014–2017) waves were used (Supplementary Fig. 1). The third wave was carried out 2.99 (range 2.0-5.4) median years after of the second one. Study variables Obesity, sarcopenia and sarcopenic obesity (SO) were measured at baseline and defined as the coexistence of the criteria of sarcopenia and obesity. Obesity Body height and weight were measured through an analogue medical scale and a stadiometer, respectively. The Body Mass Index (BMI) was calculated according to standard procedures (body weight in kg/ height2 in meters). Obesity was met when the BMI was ≥ 30 kg/m2 [18]. Sarcopenia Sarcopenia was assessed through the Foundation for the National Institutes of Health (FNIH) [19], adapted to the TSHA (standardized FNIH [sFNIH]) [7]. Muscle mass was evaluated by Dual-Energy X-ray Absorptiometry (DEXA) (Hologic, Serie Discovery QDR, Bedford, MA, USA) and analyzed by the software Physician’s Viewer (apex System Software, version 3.1.2: Bedford, USA). Appendicular Lean Muscle Mass (ALMM) was obtained from the sum of the appendicular mass of the lower and upper limbs. The ALMM cut-off points used were < 0.65 in men and < 0.54 in women [7]. Gait speed encompasses walk at a normal pace over 3 m. The best of two performances in m/s was considered. Slow gait speed cut-point was < 0.8 m/s. Isometric handgrip strength (IHS) was assessed by a Jamar dynamometer (Sammons Preston Rolyan, Bolingbrook, IL). The best of three attempts performed according to international procedures was chosen [20]. At least one minute of resting between attempts was allowed. IHS cut-off points were < 25.51 kg and < 19.19 kg for men and women, respectively. An individual was considered sarcopenic when he/she presented low appendicular muscle mass, low grip strength and low gait speed. Sarcopenic obesity Sarcopenic obesity was defined by the presence of obesity (BMI ≥ 30 kg/m2) and sarcopenia (operationalized as low ALMM, slow gait speed and weak IHS). Frailty status The Frailty Phenotype (FP) [21] and the Frailty Trait Scale 5 (FTS5) [22] were implemented for assessing frailty status. The FP [21] assesses frailty status through five domains: unintentional weight loss, low levels of physical activity, low walking speed, low IHS and exhaustion. FP cut-off points were fitted to our population [17] (Supplementary Table 1). Participants were classified as robust (zero criteria), prefrail (one or two criteria), or frail (three or more), depending on the number of criteria met. FTS5 [22] evaluates five domains to identify frailty: gait speed, IHS, self-reported physical activity, BMI and a progressive balance test (Supplementary Table 2). Each domain punctuation ranges from 0 (the lowest) to 10 (the highest), consequently, FTS5 total score ranges from 0 to 50. Frailty was diagnosed in those participants whose score was > 25. This tool was created under the rationale proposed by Fried and collaborators. However, it is more inclusive than the previous one, including more individuals as frail and presenting a greater ability to predict negative health events [22]. Moreover, a significant dose-response relationship between FTS5 score and adverse events has been found [23]. Frailty transitions Frailty status was reassessed at follow up, after a median of 2.99 (range 2.0-5.4) from the baseline examination. Accordingly, individuals were categorized according to improvement, maintenance and worsening in frailty status (robust, prefrail and frail in FP; and no frail and frail in FTS5). Mortality Mortality was ascertained through the National Death Index (Ministry of Health, Consumer Affairs and Social Welfare, Spain), phone calls and hospital records up to March 2019. The follow-up was 6.29 (median, range: 0.59–7.47) years. Comorbidity Comorbidity was assessed by the Charlson Comorbidity Index [24]. In addition, due to the relevance of key comorbidities such as hypertension, and type-2 diabetes mellitus, we performed a sensitivity analysis including only these variables. Statistical analysis Characteristics of the subjects at baseline were stratified according to the four sarcopenia-obesity categories: those with neither sarcopenia nor obesity (NS-NO), those with obesity but not sarcopenia (NS-O), those with sarcopenia but not obesity (S-NO), and those meeting criteria for both conditions (SO). Descriptive data were presented as mean (standard deviation, SD) and number (N, %). Differences between sarcopenia and obesity categories were tested using Kruskal Wallis and Chi-square test. The association between sarcopenia–obesity categories and baseline frailty was assessed using binomial logistic regression for FTS5 and multinomial logistic regression for FP. In addition, a descriptive analysis of the transitions between frailty states and death over the first four years was conducted, with results presented as frequencies (percentages) and incident rates. Comparisons of event frequencies were performed using the Chi-square test, while incidence rates were analyzed using a Poisson regression model with the follow-up time included as an offset term, allowing proper evaluation of the rate at which events occurred across groups. The incidence rate ratio between sarcopenia, obesity and SO with the outcomes were assessed using Poisson logistic regression models for frailty status at baseline, and for the improvement, maintenance and worsening in frailty category (robust, prefrail and frail in FP; and no frail and frail in FTS5). Follow-up time was included as an offset term in the Poisson regression models to account for varying observation periods. The goodness-of-fit of the Poisson model was assessed using the Pearson Chi-square test. A p-value close to 1 indicates that the model fits the observed data well. In all the analyses, we used a Model adjusted by Age, Gender, and Charlson Index. We also performed a sensitivity analysis replacing the Charlson Comorbidity Index with Diabetes Mellitus and Cardiovascular Disease as confounders in this Model. Logistic and Poisson regression models were analyzed using the stats package, while Cox proportional hazards models were implemented using the survival package to evaluate the association between sarcopenia–obesity categories and frailty status with mortality, using all available follow-up time. Statistical analyses were performed using R version 4.4.1 for Windows (Cran R Vienna, Austria). Statistical significance was set at a p-value < 0.05. Results Cross-sectional association between SO and frailty Table 1 displays the characteristics of the study participants at baseline. A total of 1538 individuals (838 women; mean age 74.73 ± 5.73 years) were included. Of them, 784 (51%) were classified as NS-NO, 406 (26.4%) as NS-O, 157 (10.2%) as S-NO and 191 (12.4%) as SO. Significant differences in frailty status were observed between categories. Table 1Descriptives according to sarcopenia, obesity and obesityVariableAllNS-NONS-OS-NOSOp-valueSample (%)1538784406157191Age, mean (SD)74.73 (5.73)74.28 (5.84)73.19 (4.77)78.66 (6.02)76.59 (5.00) < 0.001 Gender, male (%)700 (45.51)457 (58.29)183 (45.07)33 (21.02)27 (14.14) < 0.001 Charlson Index, mean (SD)1.18 (1.59)1.11 (1.51)1.02 (1.43)1.53 (1.85)1.52 (1.94) < 0.001 Frailty Phenotype Robust, n (%)1058 (68.79)588 (75.00)311 (76.60)72 (45.86)87 (45.55) < 0.001 Prefrail, n (%)436 (28.35)181 (23.09)91 (22.41)74 (47.13)90 (47.12) Frail, n (%)44 (2.86)15 (1.91)4 (0.99)11 (7.01)14 (7.33)Frailty Trait Scale 5 No Frail, n (%)1377 (89.53)765 (97.58)389 (95.81)112 (71.34)111 (58.12) < 0.001 Frail, n (%)161 (10.47)19 (2.42)17 (4.19)45 (28.66)80 (41.88)Data are expressed as mean (SD) and frequency (%)FTS5 Frailty Trait Scale 5, NO No obesity, NS No sarcopenic, O Obesity, S Sarcopenia, SD standard deviationIn bold: p-value < 0.05 Descriptives according to sarcopenia, obesity and obesity Data are expressed as mean (SD) and frequency (%) FTS5 Frailty Trait Scale 5, NO No obesity, NS No sarcopenic, O Obesity, S Sarcopenia, SD standard deviation In bold: p-value < 0.05 Risk of frailty according to the presence of obesity, sarcopenia, and sarcopenic obesity. Table 2 presents the cross-sectional associations between sarcopenia–obesity categories and frailty status. Individuals classified as NS-O exhibited an increased risk of frailty according to the FTS5 criteria, although no significant association was found using the FP definition (p > 0.05 for both prefrailty and frailty). In contrast, participants in the S-NO category demonstrated significantly elevated risks for both prefrailty [Relative Risk Ratio (RRR): 2.69] and frailty [FTS5 OR 9.93; FP RRR 11.57]. Lastly, the presence of SO was associated with a 239% higher likelihood of prefrailty and a frailty risk that was elevated by approximately 11- to 25-fold. Table 2Association between sarcopenia, obesity and sarcopenic obesity and prevalent frailtyFrailty toolSarcopeniaand obesitystateModel 1Model 2OR/RRR (95% CI)p-valueOR/RRR (95% CI)p-valueFTS5NS-NORef.Ref.NS-O1.76 (0.90, 3.42)0.0962.45 (1.22, 4.91) 0.012 S-NO16.18 (9.13, 28.66) < 0.001 9.93 (5.34, 18.48) < 0.001 SO29.02 (16.94, 49.71) < 0.001 25.86 (14.27, 46.84) < 0.001 FrailtyPhenotypePrefrailNS-NORef.Ref.NS-O0.95 (0.71, 1.27)0.7291.15 (0.85, 1.55)0.369S-NO3.34 (2.32, 4.81) < 0.001 2.69 (1.79, 4.04) < 0.001 SO3.36 (2.40, 4.71) < 0.001 3.39 (2.32, 4.96) < 0.001 FrailNS-NORef.Ref.NS-O0.50 (0.17, 1.53)0.2270.91 (0.28, 3.01)0.883S-NO5.99 (2.65, 13.54) < 0.001 7.22 (2.57, 20.30) < 0.001 SO6.31 (2.94, 13.52) < 0.001 11.57 (4.36, 30.73) < 0.001 Model 1: univariate regression model. Model 2: adjusted by Age, Gender and Charlson IndexCI Confidence Interval, NO No obesity, NS No sarcopenic, O Obesity, S Sarcopenia, OR Odds ratio, RRR Relative Risk RatioIn bold: p-value < 0.05 Association between sarcopenia, obesity and sarcopenic obesity and prevalent frailty Frailty Phenotype Model 1: univariate regression model. Model 2: adjusted by Age, Gender and Charlson Index CI Confidence Interval, NO No obesity, NS No sarcopenic, O Obesity, S Sarcopenia, OR Odds ratio, RRR Relative Risk Ratio In bold: p-value < 0.05 Interestingly, when NS-O individuals served as the reference category, the presence of sarcopenia (regardless of the presence of obesity) was strongly associated with higher odds of frailty by both instruments. In this sense, S-NO participants had 4-fold higher odds of FTS5 frailty (OR 4.05, 95% CI 2.12–7.74) and SO participants had more than 10-fold higher odds (OR 10.55, 95% CI 5.81–19.16; both p < 0.001). Similar patterns, although of lower magnitude, were also observed for prefrailty and frailty according to the FP [S-O (RRR) prefrail: 2.96 (1.96, 4.45), < 0.001 and frail 2.34 (1.51, 3.63), < 0.001 and S-NO prefrail: 2.96 (1.96, 4.45), < 0.001 and frail 2.34 (1.51, 3.63)]. Finally, when S-NO individuals were considered as the reference group, in the fully adjusted models, participants with SO exhibited 2.60-fold higher odds of frailty according to the FTS5 (OR 2.60, 95% CI 1.58–4.28; p-value < 0.001). This significant association was not observed for the FP, probably due to the limited number of frail participants (prefrailty: RRR 1.26, 95% CI 0.79–2.01; p-value 0.325; frailty RRR 1.60, 95% CI 0.62–4.15; p-value 0.332). Frailty transitions Table 3 shows proportions and incidence rates of transition in frailty status and mortality according to sarcopenia and obesity status at baseline. Table 3Proportions and incidence rates of transition in frailty status according to the presence or absence of sarcopenia and obesity Frailty tool Frailty state at baseline Sarcopenia and Obesity State at baseline Total Number of events (%) Incidence rate per 1000 person-years Frailty state at follow-up Death p -value Frailty state at follow-up Death p -value Robust Prefrail Frail Robust Prefrail Frail Frailty PhenotypeRobustNS-NO559419 (74.96)76 (13.60)6 (1.07)58 (10.38) <0.001 302.7654.924.3441.91 <0.001 NS-O296220 (74.32)45 (15.20)3 (1.01)28 (9.46)286.5658.623.9136.47S-NO6432 (50.00)24 (37.50)3 (4.69)5 (7.81)158.46118.8414.8624.76SO8153 (65.43)21 (25.93)5 (6.17)2 (2.47)219.3686.9220.698.28PrefrailNS-NO17879 (44.38)53 (29.78)6 (3.37)40 (22.47) <0.001 192.58129.2014.6397.51 <0.001 NS-O9044 (48.89)32 (35.56)4 (4.44)10 (11.11)194.41141.3917.6744.18S-NO6916 (23.19)26 (37.68)7 (10.14)20 (28.99)91.33148.4039.95111.15SO8932 (35.96)33 (37.08)11 (12.36)13 (14.61)134.99139.2146.4054.84FrailNS-NO141 (7.14)7 (50.00)2 (14.28)4 (28.58)0.98027.52192.6755.05275.24 <0.001 NS-O41 (25.00)2 (50.00)0 (0.00)1 (25.00)106.37212.730.00106.37S-NO91 (11.11)4 (44.44)2 (22.22)2 (22.22)29.16116.6358.32116.63SO102 (20.00)3 (30.00)2 (20.00)3 (30.00)59.7889.6759.78149.44Frailty toolFrailty state at baselineSarcopeniaand ObesityState at baseline Total No Frail Frail Death p -value No Frail Frail Death p -value Frailty Trait Scale 5No FrailNS-NO745622 (83.49)23 (3.09)100 (13.42) <0.001 346.5712.8255.72 <0.001 NS-O377317 (84.08)23 (6.10)37 (9.81)328.5223.8438.34S-NO10371 (68.93)19 (18.45)13 (12.62)236.7963.3743.36S-O10966 (60.55)31 (20.44)12 (11.01)227.12106.6840.86FrailNS-NO192 (10.53)9 (47.37)8 (42.11) 0.023 55.94251.72223.75 <0.001 NS-O136 (46.15)5 (38.46)2 (15.38)155.74129.7951.91S-NO4211 (26.19)16 (38.09)15 (30.95)98.57143.37152.34SO7821 (26.92)40 (51.28)17 (21.79)94.79180.5676.74NO No obesity, NS No sarcopenic, O Obesity, S SarcopeniaIn bold: p-value < 0.05 Proportions and incidence rates of transition in frailty status according to the presence or absence of sarcopenia and obesity NO No obesity, NS No sarcopenic, O Obesity, S Sarcopenia In bold: p-value < 0.05 211 (FP) and 96 (FTS5) participants worsened in their frailty status. Significant differences were observed between categories. According to the FP, compared with NS-NO individuals, those with SO and S-NO showed higher rates of progression to prefrailty/frailty and lower rates of remaining robust. In prefrail participants, similar patterns were observed, with sarcopenic individuals (S-NO and SO) exhibiting higher mortality and higher risk of progression to frailty than non-sarcopenic participants. According to FTS5, among baseline non-frail participants, S-NO and S-O showed significantly higher progression to frailty and lower likelihood of remaining non-frail than NS-NO and NS-O (p < 0.001). Among baseline frail individuals, NS-O had the highest improvement rate to non-frail and S-O the lowest mortality (both p < 0.008 vs. comparators) (Supplementary Fig. 2). Poisson regression results for frailty transitions according to obesity and sarcopenia categories are presented in Table 4. Individuals who met SO criteria experienced significantly higher elevated rates of worsening in their frailty status when compared to NS-NO, with IRRs of 1.67 in robust and 3.35 in prefrail. When using the FTS5 definition, the estimated increase in the incidence rate reached 6.56. Additionally, among those initially classified as prefrail, SO presented a lower likelihood of transitioning to robustness. Table 4Poisson regression results for worsening, maintenance and improvement in frailty status according to the presence of obesity and sarcopenia considering NS-NO group as referenceFrailty toolFrailty state at baselineSarcopeniaand obesity state at baselineWorseningStaying the sameImprovingAll-cause mortality IRR (95% CI) p -value IRR (95% CI) pv IRR (95%CI) pv HR (95% CI) p -value Frailty PhenotypeRobustNS-NORef.Ref.Ref.NS-O1.16 (0.81, 1.68)0.41320.96 (0.81, 1.13)0.61060.84 (0.45, 1.55)0.5692S-NO1.87 (1.17, 2.98) 0.0093 0.64 (0.45, 0.93) 0.0195 0.81 (0.28, 2.32)0.6925S-O1.67 (1.05, 2.68) 0.0318 0.84 (0.62, 1.13)0.24871.53 (0.69, 3.41)0.2935PrefrailNS-NORef.Ref.Ref.Ref.NS-O1.44 (0.39, 5.38)0.58551.01 (0.65, 1.58)0.95850.86 (0.59, 1.26)0.44410.50 (0.20, 1.24)0.1359S-NO2.28 (0.72, 7.24)0.16301.41 (0.87, 2.29)0.16510.59 (0.34, 1.04)0.06651.15 (0.58, 2.29)0.6912S-O3.35 (1.15, 9.71) 0.0263 1.07 (0.68, 1.68)0.77620.63 (0.41, 0.98)0.03830.83 (0.38, 1.85)0.6520FrailNS-NORef.Ref.Ref.NS-ONANA1.22 (0.31, 4.75)0.7773NANAS-NO1.33 (0.19, 9.48)0.77290.87 (0.28, 2.76)0.81390.50 (0.10, 2.50)0.3980S-O1.33 (0.19, 9.43)0.77840.73 (0.23, 2.32)0.5909NANAFrailty Trait Scale 5No FrailNS-NORef.Ref.Ref.NS-O2.26 (1.25, 4.09) 0.0071 0.93 (0.81, 1.07)0.33370.73 (0.44, 1.22)0.2345S-NO2.80 (1.46, 5.37) 0.0019 0.80 (0.62, 1.03)0.08630.84 (0.42, 1.68)0.6242S-O6.56 (3.68, 11.68) < 0.0001 0.70 (0.54, 0.92) 0.0096 1.34 (0.67, 2.67)0.4104FrailNS-NORef.Ref.Ref.NS-O0.39 (0.12, 1.35)0.13861.74 (0.34, 9.00)0.5072NANAS-NO0.61 (0.25, 1.49)0.27662.14 (0.45, 10.24)0.34033.91 (0.49, 31.07)0.1978S-O0.73 (0.32, 1.63)0.43871.55 (0.34, 7.09)0.56902.52 (0.30, 20.93)0.3920Model adjusted by Age, Gender, and Charlson IndexIRR Incidence Rate Ratio, CI Confidence Interval, HR Hazard Ratio, NA Not available, NO No obesity, NS No sarcopenic, O Obesity, S SarcopeniaIn bold: p-value < 0.05 Poisson regression results for worsening, maintenance and improvement in frailty status according to the presence of obesity and sarcopenia considering NS-NO group as reference Model adjusted by Age, Gender, and Charlson Index IRR Incidence Rate Ratio, CI Confidence Interval, HR Hazard Ratio, NA Not available, NO No obesity, NS No sarcopenic, O Obesity, S Sarcopenia In bold: p-value < 0.05 On the other hand, individuals classified as S-NO also exhibited higher rates of worsening in frailty status among those who were initially robust (IRR 1.87, according to FP) or non-frail (IRR 2.80 based on FTS5), but this association was not observed in participants who were prefrail at baseline (p-value 0.163). Moreover, S-NO individuals showed a significantly lower likelihood of remaining robust according to the Frailty Phenotype (IRR 0.64, 95% CI 0.45–0.93, p-value 0.0195). Meanwhile, NS-O was associated with an increased incidence of frailty worsening only when using the FTS5 criteria (IRR 2.26), though no significant effect was observed under the FP definition. When we have addressed the role of two relevant diseases (i.e., diabetes and hypertension) on the relationship between sarcopenic obesity and frailty, no meaningful findings were observed, except for in the case of prefrail older adults, where the presence of SO was not significantly associated with a reduced likelihood of improving the frailty status (p-value 0.084) (Supplementary Table 3). Poisson regression results for frailty transitions using SO as the reference category are presented in Table 5. According to the Frailty Phenotype, among robust participants at baseline, individuals classified as NS-NO showed a significantly lower incidence of frailty worsening compared with the SO group (IRR 0.60, 95% CI 0.37–0.96), whereas no significant differences were observed for S-NO or NS-O. No significant associations were observed for the likelihood of remaining robust or for mortality within this subgroup. Table 5Poisson regression results for worsening, maintenance and improvement in frailty status according to the presence of obesity and sarcopenia considering S-O group as referenceFrailty toolFrailty state at baselineSarcopeniaand obesity state at baselineWorseningStaying the sameImprovingAll-cause mortality IRR (95% CI) p -value IRR (95% CI) pv IRR (95%CI) pv HR (95% CI) p -value Frailty PhenotypeRobustS-ORef.Ref.Ref.S-NO1.11 (0.65, 1.92)0.69640.77 (0.49, 1.20)0.24710.53 (0.16, 1.75)0.2962NS-O0.70 (0.42, 1.14)0.15211.14 (0.84, 1.57)0.39780.54 (0.22, 1.33)0.1813NS-NO0.60 (0.37, 0.96) 0.0318 1.20 (0.88, 1.62)0.24870.65 (0.29, 1.45)0.2935PrefrailS-ORef.Ref.Ref.Ref.S-NO0.68 (0.26, 1.77)0.42921.32 (0.78, 2.24)0.30290.94 (0.51, 1.72)0.82951.38 (0.59, 3.24)0.4577NS-O0.43 (0.14, 1.37)0.15240.95 (0.57, 1.57)0.83361.36 (0.84, 2.19)0.20570.60 (0.21, 1.73)0.3466NS-NO0.30 (0.10, 0.87) 0.0263 0.94 (0.60, 1.47)0.77621.58 (1.02, 2.43) 0.0383 1.20 (0.54, 2.67)0.6520FrailS-ORef.Ref.Ref.S-NO1.01 (0.14, 7.15)0.99441.20 (0.34, 4.16)0.7768NANANS-ONANA1.67 (0.39, 7.13)0.4858NANANS-NO0.75 (0.11, 5.37)0.77841.38 (0.43, 4.40)0.5909NANAFrailty Trait Scale 5No FrailS-ORef.Ref.Ref.S-NO0.43 (0.23, 0.78) 0.0061 1.13 (0.81, 1.60)0.46910.63 (0.26, 1.53)0.3059NS-O0.34 (0.20, 0.61) 0.0002 1.33 (1.01, 1.75) 0.0439 0.55 (0.25, 1.19)0.1308NS-NO0.15 (0.09, 0.27) < 0.0001 1.42 (1.09, 1.86) 0.0096 0.75 (0.37, 1.49)0.4104FrailS-ORef.Ref.Ref.S-NO0.84 (0.47, 1.50)0.55731.38 (0.64, 2.99)0.41681.55 (0.57, 4.19)0.3884NS-O0.54 (0.19, 1.52)0.24481.12 (0.44, 2.85)0.8100NANANS-NO1.38 (0.61, 3.08)0.43870.64 (0.14, 2.94)0.56900.40 (0.05, 3.29)0.3920Model adjusted by Age, Gender, and Charlson IndexIRR Incidence Rate Ratio, CI Confidence Interval, HR Hazard Ratio, NO No obesity, NS No sarcopenic, O Obesity, S SarcopeniaIn bold: p-value < 0.05 Poisson regression results for worsening, maintenance and improvement in frailty status according to the presence of obesity and sarcopenia considering S-O group as reference Model adjusted by Age, Gender, and Charlson Index IRR Incidence Rate Ratio, CI Confidence Interval, HR Hazard Ratio, NO No obesity, NS No sarcopenic, O Obesity, S Sarcopenia In bold: p-value < 0.05 Among prefrail participants assessed using the Frailty Phenotype, NS-NO individuals exhibited a significantly lower incidence of frailty worsening (IRR 0.30, 95% CI 0.10–0.87) and a higher likelihood of frailty improvement (IRR 1.58, 95% CI 1.02–2.43) compared with the SO reference group. No statistically significant differences were observed for isolated sarcopenia or isolated obesity. Estimates among participants classified as frail at baseline were imprecise due to sparse data, precluding meaningful comparisons. According to the Frailty Trait Scale-5, among participants classified as non-frail at baseline, all phenotypes other than SO were associated with a significantly lower incidence of frailty worsening, with the strongest protective associations observed for NS-NO (IRR 0.15), followed by NS-O (IRR 0.34) and S-NO (IRR 0.43). In addition, NS-O and NS-NO participants showed a significantly higher likelihood of remaining non-frail compared with the S-O group. All Poisson regression models in Tables 4 and 5 and Supplementary Tables 3 and 4 yielded a p-value close to 1 in the Pearson Chi-square test, indicating a good fit of the models to the observed data (data not shown). Regarding mortality, no statistically significant associations were observed between sarcopenia–obesity categories and the risk of death across frailty strata, regardless of the frailty assessment tool used. (Tables 4 and 5). Discussion This study evaluates the role of sarcopenia, obesity and sarcopenic obesity with frailty and mortality in community-dwelling older adults. The main findings of the present study indicate that the coexistence of sarcopenia and obesity might be the phenotype associated with the highest risk of frailty progression, whereas isolated obesity showed only marginal effects. Sarcopenic obesity is closely related to frailty, increasing the incidence of worsening in frailty status. These results confirm the findings of Yang et al. in which the presence of SO was significantly associated with prefrailty and frailty in older adults with a mean age of 67.6 years [16]. In addition, according to our results, frailty is a dynamic entity over time, and this dynamism is highly influenced by the presence of sarcopenic obesity. Those individuals who met the criteria for sarcopenic obesity at baseline worsened in their frailty status in greater proportion than those who did not meet these criteria. This is especially true for the FTS5, and this effect is independent of age, sex or covariates. The presence of SO reduces the probability of maintaining your state of non-frailty (FTS5), as well as the improvement in the frailty status in those categorized as prefrail according to FP. These results reinforce the fact that sarcopenia, and in this case, sarcopenic obesity, could be key in the biological substrate of frailty [6, 25–28]. This manuscript expands the role of sarcopenic obesity in frailty transitions. Previously, our group published a manuscript exploring the role of sarcopenia on frailty transitions using data from the same cohort [6]. In contrast to sarcopenic obesity, prefrail individuals with sarcopenia without obesity did not present a significantly increased probability of worsening in their frailty status according to FP [6]. On the other hand, while a BMI > 29 increased the risk of worsening frailty in robust individuals, only those pre-frail individuals with a BMI > 35 were significantly less likely to recover their robustness [29]. However, the study by Rodriguez-Laso et al. [29] did not explore the presence of sarcopenia according to BMI. Taken together, these results suggest that the risk conferred by sarcopenic obesity on frailty might be influenced to a greater extent by the sarcopenic element rather than by obesity itself, as has also proposed in other study from the TSHA [30]. In our study, we have included two tools to assess frailty status. On the one hand, the instrument proposed by Fried et al. is probably the most widely used scale to assess frailty [31, 32]. To this scale is added the FTS5, created under the same paradigm, which could increase the relationship of the previous one with adverse events [22]. In addition, FTS5 has proved to present a dose-effect relationship with adverse events, and its dynamic nature allows to draw trajectories that would facilitate the monitoring of older adults in routine clinical practice [23]. According to a study conducted in four different clinical settings, the concordance of both scales is 0.52 [33]. This moderate concordance reinforces our findings, as it reinforces that the role of sarcopenic obesity as a risk factor in worsening frailty is not scale-dependent. In addition, our results confirm the association, but independence between frailty, sarcopenia and, in this case, sarcopenic obesity. As can be seen in Tables 1 and 49.69% of frail individuals according to the FTS5 presented sarcopenic obesity, while 50.31% of frail individuals according to this scale did not present this condition. This is important, since sarcopenic obesity, as we have defined it, includes 3 domains assessed in the FTS5 (BMI, gait speed, and grip strength). However, the differences in the way the three entities are categorized means that collinearity is not observed in the regression models. To date, a multitude of ways have been proposed to assess body composition, ranging from objective assessments (such as DEXA, indirect calorimetry or bioimpedance) to indirect estimates (such as waist-hip ratio, arm and ankle circumferences or the BMI used here). Of course, it is true that BMI may be suboptimal marker for adiposity in older adults [34] and might not capture obesity-related detrimental effects sensitively enough, as other variables might. However, we consider that the election of BMI to define obesity is sufficiently validated by literature. Nevertheless, we are aware that other definitions of obesity (using total fat percentage, abdominal fat percentage, waist circumference, waist-to-hip ratio, or even the possible combination of some of the above), could modify the prevalence of obesity and its association (with or without the presence of sarcopenia) with different adverse events. However, it is important to note that in younger populations, where there is an even better correlation between BMI and fat tissue, BMI showed stronger associations with mortality than other body composition endpoints such as BF%, fat mass index (FMI), fat-free mass (FFM) and FFM index (FFMI) [35]. Likewise, although BMI has been shown to increase the likelihood of worsening frailty status, this relationship could be mediated by abdominal adiposity [9], so this aspect requires further interest in the future. With respect to mortality, sarcopenic obesity did not increase the likelihood of mortality. In fact, in the crude model, frail individuals according to the FTS5 without sarcopenic obesity died twice as often as those with the condition. In this same line, previous publications of our group found that sarcopenia decrease the odds of death in frail individuals when compared with frail persons with sarcopenia. Our finding in frail people with sarcopenic obesity may reinforce the idea of different frailty phenotypes [5, 36–38], in which different approaches and interventions should be performed [39]. However, this finding was not consistent when the covariates were included in the analysis. It would be interesting to explore whether in populations with a higher number of mortality events, this data is consistent once confounding factors such as age, sex or comorbidity are included in the analysis. According to our results, the identification of sarcopenic obesity in older people, mainly in those who are frail or prefrail, is of the utmost relevance with at least two relevant purposes: to identify people at the highest risk of suffering undesired transitions in their frailty status but also with the aim of proposing interventions. The TSHA is a population-based observational cohort that investigates the impact of various factors on frailty and frailty transitions over time. Although our work suggests possible hypothetical interventions, it cannot evaluate definitive statements about therapeutic approaches, which should be validated by clinical trials. Furthermore, this cohort does not include possible interventions that physicians and other health care personnel may have applied during follow-up, but these were not systematic or protocol-based, and thus preclude drawing firm conclusions about their effects. Several interventions, especially multimodal approach based on lifestyle such as nutritional and physical exercise, appear to be the most effective in improving muscle mass, strength and performance in frail and sarcopenic older adults [40–42], or at odds of both frailty and sarcopenia [43], and have been shown to protect older obese people from the undesired effects of losing weight, including the impairment in function [44, 45]. This study presents several strengths such as the large community-based population, the standardization of frailty tools and sarcopenia, and the excellent ascertainment of mortality. Moreover, ALM has been assessed by DEXA, considered the gold-standard for evaluating body composition. In addition, the inclusion of two frailty tools reinforces our findings. On the other hand, some limitations should be mentioned, as the low prevalence of frailty according to the FP, that is lower in this study than in the whole TSHA study due to the fact that some individuals could not attend the DEXA assessment. This fact may impact on the role of evaluating SO categories on frailty transitions in those individuals who were frail at the beginning of the study, and to assess its association with mortality among this population. However, the inclusion of a second frailty scale based on the FP strengthens our results. In addition, the FTS5 [22] could provide better management of older adults, allowing to assess trajectories (both improvement and worsening) in older adults that are associated with differences in the odds of different adverse events (such as hospitalization, disability, and mortality) [23]. Another limitation that should be highlighted is that as since we evaluate frailty transitions at 3 years, there may be some transitions in shorter time that should be captured in the individuals. At last, dichotomize the participants according to the BMI limits to draw conclusions according to the gradient in BMI. Conclusions Sarcopenic obesity and sarcopenia are tightly associated with the prevalence of frailty. Moreover, SO is a powerful modulator of frailty transitions over time, but does not seem to modify the odds of mortality ascribed to each frailty state. Early identification and management of SO should be a priority to prevent frailty progression in older adults at the highest odds of developing frailty or to impair their previous frailty condition in community-dwelling older adults. Supplementary Information Supplementary Material 1. Acknowledgements We would like to thank the participants, cohort members and team researcher members. Authors' contributions AAB, JAC, WSL, and LRM designed the study, methodology and investigation. JAC contributed to data analysis and formal analysis. FJGG and LRM provided resources such as research equipment and trained clinicians in assessments and intervention protocols. AAB, WSL, and LRM wrote and reviewed the manuscript. All authors read and approved the final manuscript. Funding The present work was funded by grants from the CIBERFES (CB16/10/00464) and Instituto de Salud Carlos III (PI20/00977). The funders had no role in the design of the study; in the collection, analysis and interpretation of the data; in the writing of the report; and in the decision to submit the article for publication. Data availability Data will be available upon reasonable request to the corresponding author. Declarations Ethics approval and consent to participate Participant data were taken from the Toledo Study of Healthy Ageing (TSHA), a longitudinal cohort-based study that was designed to analyze different models of frailty and ageing and to measure the impact that frailty, comorbidity and disability may have on health system by assessing rural and urban community dwelling older adults aged 65 years or older. TSHA was performed in concordance with the Helsinki Declaration of 1975 and was approved by the Clinical Research Ethics Committee of the Toledo Hospital in Spain. All the participants signed the Informed Consent. Consent for publication Not applicable. 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. Data Availability Statement Data will be available upon reasonable request to the corresponding author.