Sarcopenic Obesity

Methods

Study Design and Setting

This was a cross-sectional study of older adults aged 50 years and above recruited from a clinical setting in Ghana. Participants with missing sex data were excluded. The analysis sample comprised individuals for whom anthropometric, bioelectrical impedance, and clinical data were available.

Measures

Sociodemographic and Anthropometric Characteristics

Sociodemographic data included age, sex, marital status, and highest educational attainment. Age was treated as a continuous variable in regression models and additionally grouped into 10-year bands (50–59, 60–69, 70–79, and 80–89 years) for descriptive purposes. Marital status was recorded as single, married, separated, divorced, or widowed, and additionally dichotomised as partnered (married) versus unpartnered (all others) for modelling. Educational level was categorised as no formal education, primary, secondary, or tertiary.

Anthropometric measurements included height (cm), weight (kg), mid-arm circumference (cm), waist circumference (cm), and hip circumference (cm). Body mass index (BMI, kg/m²) was derived from measured height and weight and categorised as underweight, normal, overweight, or obese. Waist-to-hip ratio (WHR) and waist-to-height ratio (WHtR) were computed as derived adiposity indices. Abdominal obesity was defined by any of: waist circumference ≥ 120 cm in men or ≥ 88 cm in women; WHR ≥ 0.90 in men or ≥ 0.85 in women; or WHtR > 0.50.

Exposure: Chronic Disease Multimorbidity Class

Chronic disease status was ascertained from structured instrument columns and a free-text disease-diagnosed field. Acute and infectious conditions (pneumonia, bronchitis, urinary tract infection) were excluded. Five binary indicators were constructed: (i) any diabetes (type 1 or type 2); (ii) hypertension; (iii) chronic respiratory disease (asthma or lung cancer); (iv) physical disabilities (from the neurological disease field); and (v) cardiovascular disease (stroke, coronary heart disease, or heart failure combined). Missing values in condition indicators were treated as absent where no disease was documented.

Latent class analysis (LCA) was performed using the poLCA package in R to identify underlying multimorbidity patterns. Two- to four-class solutions were estimated with 50 random starting sets per model. Model selection was guided by the Bayesian Information Criterion (BIC). The two-class solution was selected and participants were assigned to their modal predicted class. Class 1 (≈85% of participants), characterised by high conditional probabilities for diabetes, chronic respiratory disease, and cardiovascular disease, was labelled Diabetic-Respiratory. Class 2 (≈15%), defined primarily by hypertension, was labelled Hypertensive. Item-conditional response probabilities and model-fit statistics for all solutions are presented in supplementary material.

Outcome Measures

Sarcopenia (EWGSOP2). Sarcopenia was assessed using the European Working Group on Sarcopenia in Older People 2 (EWGSOP2) three-step algorithm. Case-finding used the SARC-F questionnaire, a five-item self-report tool covering difficulty lifting 4.5 kg, assistance with walking, rising from a chair, climbing stairs, and falls in the past year. Each item was scored 0–2, yielding a total score of 0–10; a score ≥ 4 indicated probable sarcopenia. Internal consistency of the SARC-F scale was evaluated using Cronbach’s alpha.

Muscle strength was assessed by handgrip dynamometry with two measurements per hand. Values exceeding 100 kg were treated as implausible and set to missing. Following EWGSOP2 protocol, the maximum valid reading across all measurements was used. Low handgrip strength was defined as < 27 kg in men and < 16 kg in women.

Muscle mass was assessed by whole-body bioelectrical impedance analysis (Omron BF511). The skeletal muscle mass index (SMMI, kg/m²) was computed as skeletal muscle mass in kilograms divided by height squared. Low muscle mass was defined using Janssen et al. (2002) NHANES III reference cut-offs: SMMI < 10.75 kg/m² in men and < 6.75 kg/m² in women. Confirmed sarcopenia required both low handgrip strength and low muscle mass. Severity staging was not possible as no gait speed or physical performance battery was collected; this constitutes a study limitation.

Sarcopenic Obesity (ESPEN-EASO 2022). Sarcopenic obesity was defined following the ESPEN-EASO 2022 consensus. At the screening step, obesity with sarcopenia risk required obesity (BMI ≥ 30 or abdominal obesity) combined with a SARC-F score ≥ 4. Elevated fat mass for the diagnostic step was defined as BMI ≥ 30 or body fat percentage > 38% in women or > 30% in men (aligned with Gallagher et al. 2000 thresholds). Body fat percentage was derived from BIA total fat mass. Reduced muscle mass or function was assessed using a body-weight-normalised skeletal muscle percentage, as recommended by ESPEN-EASO for use in obese individuals; cut-offs were < 30% in men and < 22% in women. Confirmed sarcopenic obesity required both elevated fat mass and either low handgrip strength or low body-weight-normalised muscle mass.

Statistical Analysis

Descriptive statistics are presented as mean ± standard deviation for continuous variables and as frequency with percentage for categorical variables. Characteristics are reported overall and stratified by probable sarcopenia status and by confirmed sarcopenic obesity status, with group differences assessed by chi-squared tests for categorical variables and Wilcoxon rank-sum tests for continuous variables.

Missing data were imputed using the missForest algorithm, a non-parametric random forest method suitable for mixed-type data.

Unadjusted prevalence ratios (PRs) and 95% confidence intervals for confirmed sarcopenia and confirmed sarcopenic obesity were estimated using quasi-Poisson regression, which was preferred over logistic regression to avoid overestimation of effect sizes for these non-rare outcomes. Adjusted models used logistic regression with the same four sociodemographic covariates in both outcomes — sex, age (continuous), education, and marital status — ensuring comparability across models.

To examine the association between chronic disease multimorbidity class and each outcome, the LCA-derived class was entered as a predictor in separate quasi-Poisson models adjusted for sex, age, marital status, and education. A class × sex interaction was evaluated by F-test in each model. Stratum-specific prevalence ratios were derived using estimated marginal means (emmeans package) where a significant interaction was identified. Results are reported as adjusted prevalence ratios (aPR) with 95% confidence intervals. A two-sided p-value < 0.05 was considered statistically significant. All analyses were conducted in R version 4.5.2.


Results

Sociodemographic Characteristics

There was an equal distribution of participants among the 60–69 years and 70–79 years age groups (42.56% each), and the sample was predominantly female (63.59%). More than half of participants were married, and nearly forty-six per cent had attained some form of formal education (Table 1).

Characteristic
Confirmed Sarcopenia (EWGSOP2)
N = 1951
Age (years) 65.00 (61.00, 70.00)
    Missing 7
Sex
    Female 124 (63.59%)
    Male 71 (36.41%)
Marital Status
    Single 36 (18.46%)
    Married 112 (57.44%)
    Separated 28 (14.36%)
    Divorced 7 (3.59%)
    Widowed 12 (6.15%)
Highest Education
    No formal education 28 (14.43%)
    Primary education 89 (45.88%)
    Secondary 13 (6.70%)
    Tertiary Education 64 (32.99%)
    Missing 1
1 Median (Q1, Q3); n (%)

Anthropometric and Clinical Characteristics

The median height and weight of participants were 170 cm and 70 kg, respectively. The majority of participants were classified as obese (34.02%). The median maximum hand grip strength was 28.4 kg. The median skeletal muscle mass index (SMMI), derived from whole-body bioelectrical impedance analysis (Omron BF511), was 7.67 kg/m². The median total fat mass, body fat percentage, resting metabolism, body age, visceral fat, mid-arm circumference, waist circumference, and hip circumference were 39 kg, 51.5%, 1331 kcal/day, 68 years, 10.0, 31.5 cm, 102 cm, and 107 cm, respectively. Participants had a mean waist-to-hip ratio of 0.96 and a mean waist-to-height ratio of 0.64. More than ninety per cent of participants met criteria for abdominal obesity (Table 2).

Characteristic N = 1951
Height (cm) 160 (156, 167)
Weight (kg) 70 (63, 83)
    Missing 1
Body Mass Index (kg/m²) 27.5 (23.5, 31.3)
    Missing 1
BMI categories
    Underweight 2 (1.03%)
    Normal 54 (27.84%)
    Overweight 77 (39.69%)
    Obesity 61 (31.44%)
    Missing 1
Maximum Hand Grip Strength (kg) 28 (22, 34)
Total Fat Mass (kg) 39 (26, 45)
    Missing 1
Skeletal Muscle (%) 26.7 (23.6, 32.4)
Skeletal Muscle Mass Index (kg/m²) 7.67 (6.34, 9.12)
    Missing 1
Resting Metabolism 1,331 (1,208, 1,508)
Body Age 68 (65, 77)
Visceral Fat 10.0 (8.0, 13.0)
Mid-arm Circumference (cm) 31.5 (28.0, 34.6)
Waist Circumference (cm) 102 (95, 115)
Hip Circumference (cm) 107 (101, 112)
    Missing 1
Waist-to-Hip Ratio 0.96 (0.90, 1.05)
    Missing 1
Waist-to-Height Ratio 0.64 (0.58, 0.73)
Abdominal Obesity 187 (95.90%)
Low Handgrip Strength (EWGSOP2) 33 (16.92%)
Low Muscle Mass — height-normalised (EWGSOP2/Janssen 2002) 108 (55.38%)
Probable Sarcopenia (SARC-F ≥4) — screening 65 (33.33%)
Confirmed Sarcopenia (EWGSOP2) 24 (12.31%)
Elevated Fat Mass (ESPEN-EASO) 164 (84.10%)
Low Muscle Mass — body-weight-normalised (ESPEN-EASO) 52 (26.67%)
Obesity with Sarcopenia Risk (ESPEN-EASO screening) 63 (32.31%)
Confirmed Sarcopenic Obesity (ESPEN-EASO 2022) 68 (34.87%)
1 Median (Q1, Q3); n (%)

Prevalence of Sarcopenia and Sarcopenic Obesity

Diagnostic definitions followed the EWGSOP2 three-step algorithm for sarcopenia and the ESPEN-EASO 2022 consensus for sarcopenic obesity. Both screening-based and diagnosis-based prevalence estimates are reported to illustrate the gap between case-finding and confirmed diagnosis.

Sarcopenia (EWGSOP2). Using the SARC-F questionnaire (score ≥ 4) as the case-finding step only, 33.3% of participants were classified as having probable sarcopenia. Applying the full EWGSOP2 diagnostic criteria — low maximum handgrip strength (< 27 kg in men, < 16 kg in women) confirmed by low skeletal muscle mass index (< 10.75 kg/m² in men, < 6.75 kg/m² in women; Janssen et al. 2002 BIA reference) — the prevalence of confirmed sarcopenia was 12.3%. This gap (33.3% vs. 12.3%) illustrates the extent to which SARC-F alone overestimates confirmed sarcopenia prevalence in this sample. Low handgrip strength was present in 16.9% of participants and low muscle mass in 55.4%. No gait speed or physical performance measure (SPPB, TUG, or 6MWT) was collected; severity staging (confirmed vs. severe sarcopenia) was therefore not possible and constitutes a study limitation.

A notable sex difference was observed in the direction of effect depending on the case definition. Females had a higher prevalence of probable sarcopenia (38.7% vs. 23.9% in males), likely reflecting the functional nature of SARC-F items. However, males had substantially higher rates of confirmed sarcopenia (25.4% vs. 4.8% in females), low handgrip strength (29.6% vs. 9.7%), and low muscle mass (87.3% vs. 37.1%), consistent with the higher absolute strength thresholds applied in men by EWGSOP2.

Sarcopenic obesity (ESPEN-EASO 2022). At the ESPEN-EASO screening step, 32.3% of participants met criteria for obesity with sarcopenia risk (obesity by BMI or abdominal circumference plus a SARC-F score ≥ 4). Applying the full ESPEN-EASO two-step diagnostic model — elevated fat mass (BMI ≥ 30 or body fat percentage > 38% in women / > 30% in men; Gallagher et al. 2000) combined with reduced muscle mass or function normalised to body weight — yielded a confirmed sarcopenic obesity prevalence of 34.9%. Elevated fat mass was present in 84.1% of participants, reflecting the highly obese nature of this clinical sample. Males had markedly higher confirmed sarcopenic obesity prevalence than females (62.0% vs. 19.4%), driven primarily by the greater prevalence of low body-weight-normalised skeletal muscle mass in men.

All prevalence estimates, including the comparison between screening- and diagnosis-based definitions, are presented in Table 2. Characteristics stratified by probable sarcopenia and confirmed sarcopenic obesity status are shown in Table 3.

Prevalence: screening-based vs. diagnosis-based definitions
Characteristic N = 1951
Probable Sarcopenia (SARC-F ≥4) — screening 65 (33.33%)
Low Handgrip Strength (EWGSOP2) 33 (16.92%)
Low Muscle Mass — height-normalised (EWGSOP2/Janssen 2002) 108 (55.38%)
Confirmed Sarcopenia (EWGSOP2) 24 (12.31%)
Obesity with Sarcopenia Risk (ESPEN-EASO screening) 63 (32.31%)
Elevated Fat Mass (ESPEN-EASO) 164 (84.10%)
Low Muscle Mass — body-weight-normalised (ESPEN-EASO) 52 (26.67%)
Confirmed Sarcopenic Obesity (ESPEN-EASO 2022) 68 (34.87%)
1 n (%)
Characteristic
Probable Sarcopenia (SARC-F)
Confirmed Sarcopenic Obesity (ESPEN-EASO)
No
N = 1301
Yes
N = 651
p-value2 No
N = 1271
Yes
N = 681
p-value2
age_con 64 (61, 68) 66 (62, 71) 0.132 64 (61, 70) 66 (61, 71) 0.503
    Unknown 4 3
2 5
Age categories

0.876

0.637
    60-69 54 (41.54%) 29 (44.62%)
54 (42.52%) 29 (42.65%)
    70-79 57 (43.85%) 26 (40.00%)
52 (40.94%) 31 (45.59%)
    80-89 19 (14.62%) 10 (15.38%)
21 (16.54%) 8 (11.76%)
age_group

0.553

0.496
    50-59 12 (9.52%) 3 (4.84%)
10 (8.00%) 5 (7.94%)
    60-69 83 (65.87%) 39 (62.90%)
81 (64.80%) 41 (65.08%)
    70-79 22 (17.46%) 15 (24.19%)
27 (21.60%) 10 (15.87%)
    80-89 9 (7.14%) 5 (8.06%)
7 (5.60%) 7 (11.11%)
    Unknown 4 3
2 5
Sex

0.035

<0.001
    Female 76 (58.46%) 48 (73.85%)
100 (78.74%) 24 (35.29%)
    Male 54 (41.54%) 17 (26.15%)
27 (21.26%) 44 (64.71%)
Marital Status

0.456

0.994
    Single 28 (21.54%) 8 (12.31%)
24 (18.90%) 12 (17.65%)
    Married 71 (54.62%) 41 (63.08%)
71 (55.91%) 41 (60.29%)
    Separated 20 (15.38%) 8 (12.31%)
19 (14.96%) 9 (13.24%)
    Divorced 4 (3.08%) 3 (4.62%)
5 (3.94%) 2 (2.94%)
    Widowed 7 (5.38%) 5 (7.69%)
8 (6.30%) 4 (5.88%)
Highest Education

0.809

0.221
    No formal education 17 (13.08%) 11 (17.19%)
20 (15.75%) 8 (11.94%)
    Primary education 61 (46.92%) 28 (43.75%)
53 (41.73%) 36 (53.73%)
    Secondary 8 (6.15%) 5 (7.81%)
7 (5.51%) 6 (8.96%)
    Tertiary Education 44 (33.85%) 20 (31.25%)
47 (37.01%) 17 (25.37%)
    Unknown 0 1
0 1
Height (cm) 161 (156, 169) 158 (154, 161) 0.016 160 (156, 167) 160 (156, 169) 0.687
Weight (kg) 70 (63, 84) 70 (62, 79) 0.280 70 (62, 82) 70 (63, 84) 0.810
    Unknown 1 0
1 0
BMI categories

0.934

0.022
    Underweight 1 (0.78%) 1 (1.54%)
2 (1.59%) 0 (0.00%)
    Normal 37 (28.68%) 17 (26.15%)
42 (33.33%) 12 (17.65%)
    Overweight 50 (38.76%) 27 (41.54%)
50 (39.68%) 27 (39.71%)
    Obesity 41 (31.78%) 20 (30.77%)
32 (25.40%) 29 (42.65%)
    Unknown 1 0
1 0
Maximum Hand Grip Strength (kg) 31 (26, 35) 23 (17, 31) <0.001 31 (24, 35) 25 (16, 32) <0.001
Total Fat Mass (kg) 38 (25, 45) 40 (27, 45) 0.645 35 (24, 44) 41 (35, 50) 0.002
    Unknown 0 1
1 0
Skeletal Muscle (%) 27.6 (23.6, 34.0) 26.5 (23.6, 32.2) 0.295 30.1 (24.3, 35.1) 25.6 (21.6, 27.6) <0.001
Skeletal Muscle Mass Index (kg/m²) 7.88 (6.60, 9.16) 7.27 (6.22, 8.83) 0.207 8.15 (6.70, 9.37) 6.76 (5.88, 8.24) <0.001
    Unknown 1 0
1 0
Resting Metabolism 1,324 (1,208, 1,482) 1,354 (1,208, 1,511) 0.602 1,324 (1,165, 1,508) 1,343 (1,244, 1,501) 0.319
Body Age 67 (65, 77) 75 (65, 78) 0.154 67 (65, 77) 75 (66, 78) 0.044
Visceral Fat 10.0 (8.0, 13.0) 10.0 (7.0, 12.0) 0.464 10.0 (7.0, 12.0) 10.0 (8.0, 14.0) 0.074
Mid-arm Circumference (cm) 31.5 (27.2, 34.3) 32.5 (29.0, 35.4) 0.097 31.5 (28.0, 33.5) 32.2 (27.6, 36.5) 0.095
Waist Circumference (cm) 104 (94, 118) 102 (97, 111) 0.900 101 (95, 109) 108 (97, 120) 0.027
Hip Circumference (cm) 107 (100, 112) 108 (102, 112) 0.320 105 (101, 111) 109 (101, 113) 0.036
    Unknown 1 0
1 0
Waist-to-Hip Ratio 0.96 (0.90, 1.06) 0.95 (0.90, 1.03) 0.403 0.95 (0.90, 1.05) 0.97 (0.89, 1.04) 0.669
    Unknown 1 0
1 0
Waist-to-Height Ratio 0.63 (0.57, 0.73) 0.65 (0.60, 0.71) 0.371 0.63 (0.58, 0.71) 0.69 (0.58, 0.75) 0.049
Low Handgrip Strength (EWGSOP2) 15 (11.54%) 18 (27.69%) 0.005


Low Muscle Mass — height-normalised (EWGSOP2) 74 (56.92%) 34 (52.31%) 0.541


Confirmed Sarcopenia (EWGSOP2) 13 (10.00%) 11 (16.92%) 0.165 1 (0.79%) 23 (33.82%) <0.001
Elevated Fat Mass (ESPEN-EASO)


96 (75.59%) 68 (100.00%) <0.001
Low Muscle Mass — body-weight-normalised (ESPEN-EASO)


3 (2.36%) 49 (72.06%) <0.001
Obesity with Sarcopenia Risk (ESPEN-EASO screening)


36 (28.35%) 27 (39.71%) 0.106
Probable Sarcopenia (SARC-F ≥4)


38 (29.92%) 27 (39.71%) 0.167
1 Median (Q1, Q3); n (%)
2 Wilcoxon rank sum test; Pearson’s Chi-squared test; Fisher’s exact test

Determinants of Sarcopenia and Sarcopenic Obesity

Quasi-Poisson regression was used to estimate unadjusted prevalence ratios (Table 4) across a broad set of sociodemographic and anthropometric predictors. Note that handgrip strength and skeletal muscle percentage are definitional components of confirmed sarcopenia under EWGSOP2 and are included in the unadjusted table for descriptive completeness only. Adjusted models (Table 5) use logistic regression with a common covariate set — sex, age, marital status, and education — applied consistently to both outcomes.

In the unadjusted analysis, male sex was the strongest predictor of confirmed sarcopenia and was also strongly associated with confirmed sarcopenic obesity. Full unadjusted prevalence ratios are shown in Table 4.

In the adjusted analysis, both outcomes were modelled with the same four sociodemographic covariates. For confirmed sarcopenia, male sex was associated with markedly higher odds relative to female sex (OR = 18.5, 95% CI 5.35–84.4, p < 0.001), and each additional year of age was associated with 17% higher odds of confirmed sarcopenia (OR = 1.17, 95% CI 1.09–1.28, p < 0.001). Education and marital status were not independently associated with confirmed sarcopenia after adjustment. For confirmed sarcopenic obesity, male sex similarly conferred substantially higher odds (OR = 9.80, 95% CI 4.66–22.0, p < 0.001), and older age was associated with modestly increased odds (OR = 1.06 per year, 95% CI 1.01–1.12, p = 0.027); marital status and education were not significant. Full adjusted estimates are shown in Table 5.

Characteristic
Confirmed Sarcopenia (EWGSOP2)
Confirmed Sarcopenic Obesity (ESPEN-EASO 2022)
IRR 95% CI p-value IRR 95% CI p-value
Age 1.08 1.02, 1.13 0.003 1.01 0.98, 1.04 0.439
Sex





    Female

    Male 5.24 2.31, 13.5 <0.001 3.20 2.15, 4.84 <0.001
Marital Status





    Single

    Married 1.61 0.56, 6.35 0.430 1.10 0.66, 1.91 0.727
    Separated 1.29 0.26, 6.32 0.746 0.96 0.47, 1.95 0.920
    Divorced 1.71 0.10, 12.1 0.623 0.86 0.20, 2.54 0.805
    Widowed 2.00 0.30, 10.9 0.424 1.00 0.36, 2.39 >0.999
Highest Education





    No formal education

    Primary education 1.18 0.45, 3.83 0.756 1.42 0.78, 2.78 0.278
    Secondary 1.08 0.17, 5.04 0.928 1.62 0.66, 3.82 0.279
    Tertiary Education 0.33 0.07, 1.37 0.125 0.93 0.48, 1.92 0.835
Height (cm) 1.00 0.96, 1.05 0.936 1.01 0.99, 1.03 0.477
Weight (kg) 0.99 0.96, 1.01 0.309 1.00 0.99, 1.01 0.907
BMI 1.01 0.94, 1.08 0.820


Total Fat Mass (kg) 1.02 0.99, 1.05 0.187 1.02 1.01, 1.04 0.002
Resting Metabolism 1.00 1.00, 1.00 0.327 1.00 1.00, 1.00 0.485
Body Age 1.04 1.00, 1.09 0.065 1.02 1.00, 1.04 0.062
Visceral Fat 1.06 0.96, 1.18 0.233 1.06 1.00, 1.12 0.035
Mid-arm Circumference 0.97 0.90, 1.05 0.469 1.04 1.00, 1.07 0.060
Waist Circumference 0.99 0.97, 1.02 0.592 1.01 1.00, 1.03 0.028
Hip Circumference 1.02 0.98, 1.05 0.313 1.02 1.00, 1.04 0.022
Waist-to-Hip Ratio 0.26 0.02, 3.38 0.322 1.59 0.44, 5.47 0.469
Waist-to-Height Ratio 0.36 0.01, 14.2 0.590 5.62 0.87, 35.4 0.069
Abbreviations: CI = Confidence Interval, IRR = Incidence Rate Ratio
Characteristic
Confirmed Sarcopenia (EWGSOP2)
Confirmed Sarcopenic Obesity (ESPEN-EASO 2022)
OR 95% CI p-value OR 95% CI p-value
Sex





    Female

    Male 18.5 5.35, 84.4 <0.001 9.80 4.66, 22.0 <0.001
Age (years) 1.17 1.09, 1.28 <0.001 1.06 1.01, 1.12 0.027
Highest Education





    No formal education

    Primary education 1.43 0.34, 7.15 0.635 1.64 0.55, 5.15 0.4
    Secondary 0.94 0.08, 8.26 0.953 2.61 0.52, 13.2 0.2
    Tertiary Education 0.18 0.02, 1.21 0.081 0.58 0.17, 1.95 0.4
Marital Status





    Single

    Married 1.56 0.32, 11.6 0.612 0.95 0.37, 2.56 >0.9
    Separated 2.77 0.36, 26.2 0.333 0.94 0.25, 3.35 >0.9
    Divorced 1.78 0.06, 30.1 0.697 0.54 0.06, 3.86 0.6
    Widowed 0.36 0.02, 5.99 0.477 0.63 0.09, 3.73 0.6
Abbreviations: CI = Confidence Interval, OR = Odds Ratio

Chronic Disease Classes and Association with Sarcopenic Outcomes

Latent class analysis (LCA) was applied to five collapsed chronic-condition indicators (diabetes, hypertension, chronic respiratory disease [asthma/lung cancer], physical disabilities, and cardiovascular disease) to identify multimorbidity patterns. Acute and infectious conditions (pneumonia, bronchitis, UTI) were excluded. A two-class solution was selected on the basis of BIC. Class 1 (≈85% of participants) showed high probabilities of diabetes, chronic respiratory disease, and cardiovascular disease and is labelled Diabetic-Respiratory. Class 2 (≈15%) was characterised almost exclusively by hypertension and is labelled Hypertensive.

Quasi-Poisson regression was used to estimate adjusted prevalence ratios (aPR) for confirmed sarcopenia (EWGSOP2) and confirmed sarcopenic obesity (ESPEN-EASO 2022), adjusting for sex, age, marital status, and education. An F-test confirmed a significant class × sex interaction for sarcopenic obesity (F = 8.32, p = 0.004) but not for sarcopenia (F = 3.98, p = 0.048; interaction not retained given only 3 sarcopenia cases in the Hypertensive stratum). Stratum-specific prevalence ratios for sarcopenic obesity were therefore estimated within each sex using estimated marginal means. Among females, the Diabetic-Respiratory class had significantly lower sarcopenic obesity risk relative to the Hypertensive class (aPR = 0.41, 95% CI 0.20–0.85, p = 0.017), implying that Hypertensive females carry approximately 2.5 times the sarcopenic obesity risk of their Diabetic-Respiratory counterparts. No significant class difference was observed among males (aPR = 2.71, 95% CI 0.80–9.18, p = 0.109). Full adjusted prevalence ratios are shown in Table 6.

Characteristic
Confirmed Sarcopenia (EWGSOP2)
Confirmed Sarcopenic Obesity (ESPEN-EASO 2022)
aPR (95% CI) 95% CI p-value aPR (95% CI) 95% CI p-value
Chronic disease class





    Diabetic-Respiratory

    Hypertensive 0.93 0.31, 2.26 0.875 2.45 1.13, 5.04 0.018
Sex





    Female

    Male 8.34 4.01, 18.8 <0.001 5.20 3.10, 9.09 <0.001
Age (years) 1.11 1.07, 1.16 <0.001 1.03 1.00, 1.07 0.043
Highest Education





    No formal education

    Primary education 1.23 0.54, 3.15 0.638 1.33 0.70, 2.75 0.408
    Secondary 0.95 0.22, 3.52 0.936 1.78 0.68, 4.52 0.228
    Tertiary Education 0.28 0.08, 0.90 0.035 0.71 0.34, 1.56 0.375
Marital Status





    Single

    Married 1.37 0.48, 5.19 0.592 0.99 0.56, 1.84 0.974
    Separated 2.28 0.60, 9.80 0.233 1.08 0.47, 2.40 0.843
    Divorced 1.68 0.20, 9.73 0.580 0.72 0.16, 2.23 0.610
    Widowed 0.49 0.09, 2.63 0.392 0.67 0.21, 1.94 0.482
Class × Sex





    Hypertensive * Male


0.15 0.03, 0.56 0.010
Abbreviations: CI = Confidence Interval, IRR = Incidence Rate Ratio

The class × sex interaction on confirmed sarcopenic obesity is shown in Figure 1. Among males, predicted prevalence was substantially higher in the Diabetic-Respiratory class (64.2%, 95% CI 42.2–97.6%) than in the Hypertensive class (23.7%, 95% CI 6.9–81.2%), whereas among females the pattern reversed: Hypertensive females had higher predicted prevalence (30.2%, 95% CI 15.4–59.5%) than Diabetic-Respiratory females (12.3%, 95% CI 7.1–21.5%).

Predicted prevalence of confirmed sarcopenic obesity by chronic disease class and sex. Points are estimated marginal means from the quasi-Poisson interaction model, adjusted for age, marital status, and education. Error bars represent 95% confidence intervals.

Supplementary Material

LCA Model Selection and Class Profiles

The following supplementary tables document the latent class analysis (LCA) used to derive the chronic disease multimorbidity classes reported in the main text. Table S1 presents fit statistics for 2- to 4-class solutions, supporting the selection of the 2-class model. Table S2 provides the item-conditional response probabilities for the selected model, allowing readers to audit how each class was characterised.

Supplementary Table S1
LCA model fit statistics for 2- to 4-class solutions
Number of classes Log-likelihood No. of parameters AIC BIC G² (df)
2 -554.29 11 1130.58 1166.59 13.9 (20)
3 -550.01 17 1134.01 1189.65 5.33 (14)
4 -549.02 23 1144.04 1219.32 3.36 (8)
Indicators: diabetes, hypertension, chronic respiratory disease (asthma/lung cancer), physical disabilities, cardiovascular disease. Acute conditions (pneumonia, bronchitis, UTI) excluded. Lowest BIC identifies the preferred solution (2-class, bold).
Supplementary Table S2
Item-conditional response probabilities for the 2-class LCA solution
Indicator Class 1 – Diabetic-Respiratory
(n ≈ 170; 87.4%)
Class 2 – Hypertensive
(n ≈ 25; 12.6%)
Diabetes 0.622 0.403
Hypertension 0.120 0.998
Chronic respiratory disease 0.361 0.104
Physical disabilities 0.152 0.210
Cardiovascular disease 0.295 0.067
Values are P(condition present | class). Probabilities ≥ 0.50 indicate a defining feature of that class.