Sex-specific patterns of skeletal muscle aging at the L3 level: a quantitative study based on opportunistic CT

Abstract

Objective

This opportunistic CT-based study aimed to investigate age- and sex-related changes in muscle mass and quality at the L3 vertebral level in men aged ≥ 50 years and postmenopausal women.

Methods

Clinical and abdominal CT data from 292 hospitalized patients (146 men, 146 women) were retrospectively analyzed. At the L3 level, skeletal muscle area (SMA), skeletal muscle index (SMI), skeletal muscle density (SMD), intermuscular adipose tissue (IMAT), and intramuscular adipose tissue content (IMAC) were measured using Image J. Two-way ANOVA was used to test for sex-by-age interactions. Univariate linear regression and LOESS smoothing were additionally applied to quantify and visualize age-related trajectories, respectively.

Results

All CT-derived measurements showed excellent inter-rater reliability (ICC > 0.75). Two-way ANOVA revealed significant sex-by-age interactions for SMA and SMI (p = 0.001 and p = 0.003, respectively), but not for SMD or IMAT. Men exhibited a more rapid age-related decline in muscle mass: each additional year of age was associated with a decrease in SMA of 1.331 cm² in men versus 0.465 cm² in women. Women demonstrated an earlier onset of muscle quality deterioration. Compared with younger postmenopausal women, older postmenopausal women had significantly lower SMD (p < 0.05), and IMAT increased significantly after age 60 (p < 0.05). IMAT was a more sensitive imaging biomarker of age-related myosteatosis than IMAC.

Conclusions

Opportunistic CT-based assessment reveals a distinct sexual dimorphism in skeletal muscle aging: men lose muscle mass more rapidly, whereas women experience an earlier onset of muscle quality decline, driven predominantly by intermuscular fat infiltration. Opportunistic CT effectively captures these sex-specific differences, supporting its application in opportunistic screening and sex‑specific management of sarcopenia.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12891-026-10167-7.

Background

Aging is associated with metabolic, physiological, and functional impairments, partly due to changes in body composition. Even in healthy individuals, muscle mass and function gradually decline with age, a phenomenon termed "sarcopenia" by Rosenberg in 1989 [1]. Sarcopenia is a major health issue in aging societies, linked to conditions such as atherosclerosis, diabetes, and cognitive decline, and is a significant risk factor for falls, fractures, and mortality [2–5]. Therefore, accurate assessment of muscle mass and quality, and early detection of muscle mass and quality decline, are clinically important.

Muscle mass refers to the quantity or size of skeletal muscle, commonly quantified by computed tomography (CT) as the skeletal muscle area (SMA) of the abdominal wall and paravertebral muscle groups at the L3 level, normalized to patient height to generate the skeletal muscle index (SMI) [6]. In contrast, muscle quality is a broader concept encompassing the micro- and macroscopic structure and composition of muscle beyond its mere quantity [7, 8]. Currently, muscle quality is conceptualized within two major frameworks: functional muscle quality, defined as the ratio of muscle strength to muscle mass (i.e., muscle-specific strength), and morphological muscle quality, which refers to the structural and compositional characteristics of muscle tissue, including fat infiltration, fibrosis, and other histological features [9–12]. In imaging-based studies, morphological muscle quality is commonly operationalized via CT-derived variables that reflect myosteatosis (pathological fat infiltration into muscle), such as skeletal muscle density (SMD), intermuscular adipose tissue (IMAT), and intramuscular adipose tissue content (IMAC) [13, 14]. Importantly, these CT metrics are surrogate markers rather than direct measures of muscle quality, and they represent only a subset of the multiple dimensions—including histological, biochemical, and functional assessments—that collectively define this construct [12, 15, 16].

Opportunistic CT has gained attention for assessing muscle mass and quality without additional radiation or cost, using existing diagnostic images [16, 17]. CT-derived variables include SMA, SMI, SMD, as well as IMAT and IMAC.

Decreased muscle mass and quality are key features of skeletal muscle aging [18, 19]. While early research focused primarily on muscle mass [20–22], it is now understood that muscle function often declines more rapidly than mass, implicating muscle quality as a critical factor. Myosteatosis, or fat infiltration into muscle, is a major contributor to declining muscle quality. IMAT and IMAC are established CT variables for assessing myosteatosis, with some studies suggesting IMAT is a more sensitive marker of aging than muscle mass [23, 24]. Sex also plays a significant role. Evidence suggests that women have greater muscle fat infiltration than men, independent of BMI [25, 26]. However, some studies report no interaction between age and sex, suggesting that sex differences may be muscle-specific [26, 27].

Many previous studies have assessed relatively healthy individuals undergoing physical examinations [28, 29]. However, these populations often have underlying health issues, and a lack of detailed medical history can introduce errors in muscle assessment [28]. To better understand the sex-specific changes in skeletal muscle with aging, this study used abdominal opportunistic CT data from hospitalized patients with clearly documented medical histories. This approach allows for a cross-sectional exploration of age-related changes in CT-derived metrics of skeletal muscle mass and quality in older adults.

Given the above background, the primary objective of this study was to investigate age- and sex-related changes in CT-derived metrics of skeletal muscle mass and morphological muscle quality at the L3 vertebral level in men aged ≥ 50 years and postmenopausal women. Specifically, we aimed to: (1) characterize the age-related trends of muscle mass and quality decline across different age groups; (2) compare these trends between sexes; and (3) identify which CT-derived variables are most sensitive to age-related changes in men and women.

We proposed the following three hypotheses. First, with advancing age, both muscle mass and muscle quality decline in both sexes, and the extent of this decline varies across age groups. Second, the decline in muscle mass is more rapid in men, whereas the deterioration in muscle quality occurs earlier and is more pronounced in women. Third, among CT-derived measures of muscle adiposity, IMAT is a more sensitive imaging biomarker of age-related myosteatosis than IMAC, and its age-related increase occurs earlier in women than in men.

Methods

Study population

This retrospective study consecutively enrolled hospitalized patients who underwent abdominal CT examinations at Tianjin Hospital, Tianjin University, Tianjin, China, between July 2019 and July 2022. Tianjin is a major metropolitan municipality in northern China, with a mixed population from urban, suburban, and peri-urban areas. During the study period, all patients aged ≥ 50 years (for men) or postmenopausal women who underwent abdominal CT for clinical indications (including trauma evaluation, preoperative assessment, or investigation of gastrointestinal symptoms) were consecutively screened for eligibility. Patients were included if they met the predefined inclusion and exclusion criteria (detailed below) without further selection. All patients were residents of Tianjin municipality and surrounding regions, including both urban and suburban districts.

Inclusion and exclusion criteria

Inclusion criteria: Men ≥ 50 years old, women in postmenopausal state; Clear and usable abdominal CT images without various reasons causing increased or decreased density in the abdominal wall and paravertebral muscles, subcutaneous fat of the abdominal wall, and intra-abdominal fat gap at the L3 level, such as edema, bleeding, artifact, etc. [30, 31].

Exclusion criteria: patients with incomplete medical records or those suffering from conditions known to affect muscle mass or quality, including: (1) musculoskeletal system diseases (e.g., polymyositis, muscular dystrophy); (2) neuromuscular system diseases (e.g., amyotrophic lateral sclerosis, spinal cord injury); (3) malignant tumors (any primary or metastatic cancer diagnosed within the past 5 years); (4) chronic wasting diseases, including severe diabetes (HbA1c > 9.0% or requiring insulin therapy), chronic obstructive pulmonary disease (GOLD stage III or IV), chronic liver disease (Child–Pugh class B or C), chronic kidney disease (stage 4 or 5), or heart failure (NYHA class III or IV) [32–36].

Abdominal CT examination and image reconstruction

A 128-row spiral CT scanner GE revolution CT ES (GE Medical Systems, LLC) was used. Patients lay in a supine position with arms raised above the head, entering head first. Breath-holding during scanning. The scanning range was from the upper edge of the T12 vertebra to the lower edge of the L5 vertebra. Tube voltage 120 kV [19, 37], automatic tube current technology, scanning slice thickness 10 mm, pitch 0.948:1.The original thin axial images were transferred to the GE AW4.7 work station, reconstructed to obtain a single axial image at the L3 vertebral level, parallel to the upper endplate of the L3 vertebra and displaying the longest cross-section of the L3 transverse process as much as possible, reconstruction algorithm standard, reconstruction slice thickness 0.625 mm, window width 350HU, window level 40HU, display field 45 cm (Fig. 1).

Fig. 1.

Methodology for obtaining transverse images at the L3 vertebral level. Coronal and sagittal CT localizer images demonstrating the plane (red line) parallel to the superior endplate of the L3 vertebra, selected to obtain the axial slice with the longest transverse processes for body composition analysis

Measurement of muscle mass and quality in the abdominal wall and paravertebral muscles at the L3 level

The third-party software Image J 1.53e (Wayne Rasband and contributors National Institutes of Health, USA) was used to open the axial abdominal image at the L3 vertebral level. The Polygon selections tool was clicked to outline the overall muscles’ group of the abdominal wall and paravertebral at the L3 level. The Analyze-Measure tool was clicked to calculate the overall muscles’ group SMA (cm²) and SMD (CT attenuation, HU). Then Image-Adjust-Threshold set the intermuscular fat CT attenuation threshold to −190 ~ −30HU [19, 38], calculating IMAT(%) of the overall muscles’ group. The same method was used to outline the overall contour of the subcutaneous adipose tissue (SAT) and measure its CT attenuation (HU). The above variables were measured independently by two radiologists (Reader 1 and Reader 2), both of whom had more than five years of experience in musculoskeletal radiology. To ensure blinding, both readers were unaware of the patients' age, sex, and other clinical information during the measurement process. The readers performed the measurements independently and were not allowed to discuss or compare their results. The average of the two measurements was recorded for each variable.

The following formulas were used to calculate the SMI (cm²/m²) and IMAC of the overall muscles’ group of the abdominal wall and paravertebral at the L3 level (Fig. 2).

Fig. 2.

Representative example of body composition analysis at the L3 vertebral level using Image J software. Axial CT image from a 53-year-old female participant (BMI 27.34 kg/m²) showing the segmentation of: A skeletal muscle area (SMA), B intermuscular adipose tissue (IMAT), and C subcutaneous adipose tissue. For this participant, SMA was 101.81 cm.², skeletal muscle density (SMD) was 30.42 HU, IMAT constituted 4.89% of the total muscle group, CT attenuation of subcutaneous adipose tissue (SAT) was −99.88 HU, resulting in IMAC of −0.30

1

2

Statistical analysis

Sample size considerations

This study was based on a convenience sample of hospitalized patients who met the eligibility criteria during the study period. No a priori sample size calculation was performed, as this was an exploratory, opportunistic CT study. The final sample size (n = 292, 146 per sex) is comparable to or larger than previous CT-based studies on age-related muscle changes.

Data presentation and normality

All continuous variables are expressed as mean ± standard deviation (SD). The Shapiro–Wilk test was used to assess data normality, and all variables followed a normal distribution.

Inter-rater reliability

Intra-class correlation coefficients (ICCs) were computed to evaluate the inter-rater reliability of CT-derived indices, including SMA, SMI, SMD, IMAT, and IMAC. ICC reliability was classified as poor (< 0.40), fair (0.40–0.59), good (0.60–0.74), and excellent (≥ 0.75).

Between-group comparisons

Independent-samples t tests were used to compare muscle parameters between men and women within each age group. One-way analysis of variance (ANOVA) with post hoc tests was performed to compare muscle variables across age groups for each sex separately. The least significant difference (LSD) test was applied for homogeneous variances, whereas Dunnett’s T3 test was used for heterogeneous variances.

Primary analysis: two-way ANOVA

Two-way ANOVA was conducted to examine the main and interactive effects of sex and age group (≤ 59, 60–69, ≥ 70 years) on all muscle-related variables. Simple effect analysis was performed when a significant interaction occurred, and Tukey’s honestly significant difference (HSD) test was used for post hoc pairwise comparisons among age groups.

Secondary analysis: univariate linear regression

Given inherent sex-specific differences in muscle composition, Pearson correlation analyses stratified by sex were performed to explore the relationships between age and each CT-derived muscle index. The magnitude of correlation was interpreted as strong (|r|≥ 0.7), moderate (0.4 ≤|r|< 0.7), weak (0.1 ≤|r|< 0.4), and negligible (|r|< 0.1). Sex-stratified univariate linear regression analyses were further performed to quantify the effects of age (independent variable) on each muscle variables (dependent variable).

Exploratory analysis: LOESS smoothing

To complement the linear regression models and visualize potential nonlinear age-related trends in muscle variables, sex-stratified locally estimated scatterplot smoothing (LOESS) curves were plotted with a smoothing span of 0.75. These curves were used for descriptive exploration only and were not intended for hypothesis testing.

Statistical analyses were performed using IBM SPSS Statistics 26.0 (IBM Corp., Armonk, NY, USA).

Results

Clinical characteristics

Figure 3 shows the patient inclusion flowchart. A total of 292 individuals were enrolled, comprising 146 men (mean age 62.25 ± 9.99 years) and 146 women (mean age 69.16 ± 9.95 years), with an overall mean age of 65.71 ± 10.54 years. The cohort included 190 patients hospitalized for traumatic fractures, 77 for cardiovascular diseases (45 NYHA class II, 32 NYHA class I), and 25 for respiratory diseases (16 GOLD stage I, 9 GOLD stage II). Fracture patients underwent abdominal CT to rule out abdominal organ injury, while patients with cardiovascular or respiratory diseases underwent CT to investigate gastrointestinal symptoms (e.g., poor appetite, nausea) or to follow up on existing abdominal conditions (e.g., liver hemangioma, chronic cholecystitis).

Fig. 3.

Flow diagram of patient inclusion and exclusion

Participant characteristics and inter-rater reliability of CT-derived measurements

Participant characteristics and inter-rater reliability of CT-derived measurements are summarized in Tables 1 and 2. Table 1 presents the demographic characteristics of the study population stratified by age group and sex. Within each age group, no significant differences were observed between men and women in age or body mass index (BMI) (independent-samples t-test, all p > 0.05). As shown in Table 2, all CT-derived muscle variables demonstrated good to excellent inter-rater reliability, with ICCs ranging from 0.793 to 0.917.

Table 1.

Demographic characteristics by age group and sex

Age Group (years)	Sex	n	Age (years)	BMI (kg/m2)
≤ 59	Men	65	53.58 ± 3.47	23.99 ± 3.38
	Women	28	55.25 ± 4.59	24.76 ± 3.37
60–69	Men	50	64.14 ± 2.92	24.48 ± 3.05
	Women	49	65.02 ± 2.72	24.77 ± 4.22
≥ 70	Men	31	77.35 ± 6.17	23.62 ± 3.44
	Women	69	77.75 ± 5.52	23.26 ± 4.06

Table 2.

Inter-rater reliability of key CT measurements at L3 level

Variables	Reader 1	Reader 2	ICC	95% CI	p-value
SMA(cm2)	102.86 ± 24.57	109.41 ± 28.62	0.910	0.820–0.956	< 0.001
SMD(HU)	33.32 ± 8.17	27.94 ± 9.31	0.902	0.804–0.952	< 0.001
IMAT(%)	8.68 ± 3.23	10.28 ± 4.06	0.793	0.610–0.896	< 0.001
SAT attenuation (HU)	−86.39 ± 18.87	−84.35 ± 19.19	0.917	0.833–0.960	< 0.001

Sex‑specific patterns and age‑related changes in CT‑derived muscle variables at the L3 level

Table 3 summarizes the CT‑derived muscle variables at the L3 level stratified by sex and age group, revealing significant sex differences and distinct age‑related trajectories for both muscle mass and quality.

Table 3.

CT‑derived muscle variables at the L3 level: sex‑specific patterns and age‑related changes

Age group (years)	Sex	SMA (cm2)	SMI (cm2/m2)	SMD (HU)	IMAT (%)	IMAC
≤ 59	Men	136.24 ± 24.66	46.37 ± 8.22	37.31 ± 7.16	6.89 ± 3.40	–0.49 ± 0.31
	Women	91.96 ± 12.19***	36.13 ± 4.85***	31.07 ± 8.37***	10.39 ± 4.86***	–0.32 ± 0.11***
60–69	Men	128.73 ± 21.45	44.19 ± 6.69	35.14 ± 8.02	7.89 ± 4.09	–0.45 ± 0.23
	Women	89.70 ± 17.38***	35.28 ± 6.40***	26.15 ± 8.65***	13.70 ± 5.95***	–0.27 ± 0.10***
≥ 70	Men	105.60 ± 25.57†‡	36.68 ± 8.57†‡	29.67 ± 9.10†‡	10.88 ± 4.85†‡	–0.46 ± 0.39
	Women	84.58 ± 18.65***	33.64 ± 6.90	21.66 ± 10.15***	16.06 ± 8.19***†‡	–0.26 ± 0.17*

^* p < 0.05,** p < 0.01,*** p < 0.001 for sex difference (men vs. women) within the same age group

^† p < 0.05 vs. ≤ 59 years group within the same sex

^‡ p < 0.05 vs. 60–69 years group within the same sex

Significant sex differences were observed for all muscle variables—SMA, SMI, SMD,IMAT, and IMAC —within each age group (all p < 0.05). Across all age groups, men exhibited significantly higher muscle mass (SMA, SMI) and muscle quality (SMD) than women (all p < 0.05).

In men, compared with the two younger age groups (≤ 59 years and 60–69 years), those aged ≥ 70 years showed significantly lower SMA, SMI, and SMD (all p < 0.05), accompanied by a significant increase in IMAT (p < 0.05). In contrast, IMAC did not change significantly with age in men (p > 0.05).

In postmenopausal women, SMD decreased progressively and significantly across age groups (p < 0.05). Although SMA and SMI showed a declining trend and IMAC an increasing trend with advancing age, these changes did not reach statistical significance in pairwise comparisons across age groups (p > 0.05). Notably, IMAT increased significantly after the age of 60 years (p < 0.05).

Overall, both muscle mass (SMA, SMI) and muscle quality (SMD) declined with advancing age in both sexes. However, the rate of muscle mass decline was more rapid in men, whereas the increase in IMAT —a key marker of myosteatosis—occurred earlier in women and reached higher absolute levels.

Main and interactive effects of sex and age group on CT-derived muscle variables at the L3 level

Table 4 presents the two-way ANOVA results for the main and interactive effects of sex and age group on CT-derived muscle variables at the L3 level. Significant main effects of sex and age group were observed for all muscle variables (p < 0.001), except for the main effect of age group on IMAC (p = 0.394). Notably, significant sex-by-age group interactions were found for SMA (p = 0.001) and SMI (p = 0.003), indicating that the age-related decline in muscle mass differs between men and women. In contrast, no significant interactions were detected for SMD (p = 0.573), IMAT (p = 0.393), or IMAC (p = 0.922), suggesting that the declines in muscle quality (SMD) and fat infiltration (IMAT) with age follow similar trajectories in both sexes, despite absolute differences.

Table 4.

Two-way ANOVA results for CT-derived muscle variables at the L3 level: main effects of sex and age group and their interaction

Variable	Effect	df	F	p value	Partial η2
SMA (cm2)	Sex	1	182.708	< 0.001	0.390
	Age group	2	18.960	< 0.001	0.117
	Sex × Age group	2	7.263	0.001	0.048
SMI (cm2/m2)	Sex	1	69.925	< 0.001	0.196
	Age group	2	16.569	< 0.001	0.104
	Sex × Age group	2	6.090	0.003	0.041
SMD (HU)	Sex	1	52.173	< 0.001	0.154
	Age group	2	20.191	< 0.001	0.124
	Sex × Age group	2	0.557	0.573	0.004
IMAT (%)	Sex	1	47.438	< 0.001	0.142
	Age group	2	14.982	< 0.001	0.095
	Sex × Age group	2	0.938	0.393	0.007
IMAC	Sex	1	37.781	< 0.001	0.117
	Age group	2	0.934	0.394	0.006
	Sex × Age group	2	0.081	0.922	< 0.001

df degrees of freedom, F F-statistic, Partial η² partial eta squared

Age-related correlations and linear regression analyses for CT-derived muscle variables at the L3 level, stratified by sex

Table 5 presents the sex-stratified univariate linear regression results for age as an independent variableof each CT-derived muscle variable at the L3 level, along with estimates for the overall population. At the overall population level (n = 292), age was significantly associated with SMA, SMI, SMD, and IMAT (all p < 0.001), with the strongest effects observed for SMA (β = −0.502, p < 0.001) and SMI (β = −0.474, p < 0.001). For IMAC, age had a weak positive association in the overall population (β = 0.133, p = 0.023).

Table 5.

Univariate linear regression analysis of age on muscle variables at L3 level

Variables	Sex	B	SE	Standardized β	t	p-value
SMA (cm2)	Men	−1.331	0.190	−0.504	−7.001	< 0.001
	Women	−0.465	0.140	−0.267	−3.325	0.001
	Overall	−1.419	0.144	−0.502	−9.888	< 0.001
SMI (cm2/m2)	Men	−0.415	0.063	−0.482	−6.598	< 0.001
	Women	−0.159	0.052	−0.246	−3.047	0.003
	Overall	−0.395	0.043	−0.474	−9.166	< 0.001
SMD (HU)	Men	−0.292	0.065	−0.349	−4.472	< 0.001
	Women	−0.338	0.078	−0.338	−4.306	< 0.001
	Overall	−0.437	0.052	−0.441	−8.358	< 0.001
IMAT (%)	Men	0.131	0.034	0.309	3.895	< 0.001
	Women	0.198	0.058	0.272	3.397	0.001
	Overall	0.242	0.034	0.383	7.061	< 0.001
IMAC	Men	−0.001	0.003	−0.043	−0.515	0.608
	Women	0.002	0.001	0.128	1.554	0.122
	Overall	0.003	0.001	0.133	2.293	0.023

Sex-stratified analysis revealed that the negative effects of age on SMA and SMI were significantly larger in men than in women (SMA: β = −0.504 in men vs. −0.267 in women; SMI: β = −0.482 in men vs. −0.246 in women. In contrast, the age-related decline in SMD was comparable between sexes (β = −0.349 in men vs. −0.338 in women). For IMAT, age had a significant positive effect in both sexes (β = 0.309 in men, p < 0.001; β = 0.272 in women, p = 0.001), with a slightly larger effect size in men. For IMAC, age had a weak positive association in the overall population (β = 0.133, p = 0.023). However, when stratified by sex, no significant age-related effect was detected in either men (β = −0.043, p = 0.608) or women (β = 0.128, p = 0.122). This discrepancy likely reflects the reduced statistical power in sex-stratified analyses due to smaller sample sizes (n = 146 per group), combined with the small effect size of age on IMAC (overall β = 0.133), which falls below the detection threshold in smaller subsamples.

Quantitatively, with each additional year of age, SMA decreased by 1.331 cm² in men (B = −1.331, p < 0.001) and by 0.465 cm² in women (B = −0.465, p = 0.001). SMI decreased by 0.415 cm²/m² per year in men (B = −0.415, p < 0.001) and by 0.159 cm²/m² per year in women (B = −0.159, p = 0.003). SMD declined annually by 0.292 HU in men (B = −0.292, p < 0.001) and by 0.338 HU in women (B = −0.338, p < 0.001). IMAT increased by 0.131% per year in men (B = 0.131, p < 0.001) and by 0.198% per year in women (B = 0.198, p = 0.001). Collectively, these findings indicate that age exerts a greater negative impact on muscle mass (SMA, SMI) in men than in women.

Conversely, the age-related decline in muscle quality, as reflected by SMD, is similar between sexes. Although IMAT increases significantly with age in both sexes, the effect size is slightly larger in men. For IMAC, no significant age-related effect was detected in either sex after stratification, despite a weak but statistically significant association in the overall population, which may be attributable to reduced statistical power in subsample analyses.

Age-related changes in muscle variables: LOESS-smoothed scatter plots by sex

The LOESS-smoothed scatter plots for the associations between age and individual muscle variables (stratified by sex) are shown in Figs. 4, 5, 6, 7, 8, 9, 10, 11, 12 and 13.

Fig. 4.

Decline in SMA with Age in Women. SMA shows a gradual, relatively linear decline with age, with no obvious inflection point

Fig. 5.

Decline in SMA with Age in Men. SMA shows a steady decline with age

Fig. 6.

Mild Decline in SMI with Age in Women. SMI shows a mild, gradual decline across age groups

Fig.7.

Decline in SMI with Age in Men. SMI declines consistently with age. The decline appears slower before approximately 65 years, with a potentially steeper slope thereafter

Fig. 8.

Moderate decline in SMD with age in women. SMD shows a moderate negative linear trend with age

Fig. 9.

Decline in SMD with Age in Men. SMD shows a consistent decline across the age range

Fig. 10.

Increase in IMAT with Age in Women. IMAT shows an overall increasing trend with age, with substantial inter-individual variability. The curve suggests a potential acceleration after approximately 70 years

Fig. 11.

Increase in IMAT with Age in Men. IMAT increases with age, with marked inter-individual variability, particularly in older age groups

Fig. 12.

No clear age-related trend IMAC in women. The LOESS curve remains nearly flat across the entire age range

Fig. 13.

No Clear age-related trend in IMAC in Men. The LOESS curve shows no apparent linear or nonlinear trend with age

For muscle mass variables, SMA and SMI showed a gradual decline with advancing age in both sexes (Figs. 4, 5, 6 and 7). However, the decline appeared steeper in men than in women, suggesting a more pronounced age-related muscle loss in men. Notably, in men, the decline in SMI appeared relatively slower before approximately 65 years of age, with a potentially steeper slope thereafter.

For muscle density (SMD), a moderate negative linear trend was observed in both sexes, with comparable declining slopes between men and women (Figs. 8 and 9), indicating that age-related myosteatosis progresses at a similar rate regardless of sex.

For IMAT, both sexes showed an overall increasing trend with age (Figs. 10 and 11). Substantial inter-individual variability was observed across all age groups, particularly in older individuals. The LOESS curves suggested a potential acceleration in IMAT accumulation after approximately 70 years of age, though this observation remains descriptive.

For IMAC, the LOESS curves remained nearly flat across the entire age range in both sexes (Figs. 12 and 13), indicating no clear age-related trend.

Overall, age-related declines in muscle mass (SMA and SMI) were more pronounced in men than in women, whereas declines in muscle quality (SMD) were similar between sexes. IMAT showed high inter-individual variability and a potential late-life acceleration, while IMAC appeared insensitive to age.

Discussion

This study leveraged opportunistic abdominal CT to characterize age- and sex-specific changes in muscle mass and quality at the L3 level, testing three pre-specified hypotheses. Our findings confirm that both muscle mass (SMA, SMI) and muscle quality (SMD, IMAT) decline with age in both sexes. Supporting our second hypothesis, men exhibited a more rapid decline in muscle mass, whereas women showed an earlier onset of muscle quality deterioration, particularly in CT markers of myosteatosis — revealing a clear sexual dimorphism. In line with our third hypothesis, IMAT proved to be a more sensitive imaging biomarker of age-related myosteatosis than IMAC, and its age-related increase occurred earlier in women than in men. Collectively, these conclusions are supported by two-way ANOVA, univariate linear regression, and LOESS-smoothed analyses, which together confirmed significant sex-by-age interactions for muscle mass metrics (SMA, SMI) and comparable age-related trajectories for muscle quality metrics (SMD, IMAT) between sexes.

Factors influencing the stability of CT measurements in muscle mass and quality

The stability of CT-derived measurements is influenced by several technical factors. In this study, all patients were examined using the same CT scanner with consistent scanning conditions (120 kV, 10 mm slice thickness, 0.625 mm reconstruction thickness, standard algorithm, window width 350 HU, window level 40 HU), which represents an internal strength. Additionally, inter-rater reliability for all muscle variables was good to excellent (ICC range: 0.793–0.917), further supporting the internal validity of our measurements. However, caution is warranted when comparing our absolute values with those from other studies, as measurement differences have been reported across CT manufacturers [39], tube voltage and slice thickness [40, 41], iodine contrast agents [40], and software packages [42–44]. Since we did not compare different software or scanners, nor perform intra-rater reliability testing, we refrain from overinterpreting these external comparisons.

The impact of age and sex on muscle mass

Muscle mass, reflecting muscle size [45], is a well-established predictor of clinical outcomes [46]. Our study confirms that SMA and SMI at L3 decline with age in both sexes. Consistent with prior research [18, 47], we observed that this decline accelerates after age 70 in men.Two‑way ANOVA revealed significant main effects of sex and age group on SMA and SMI (both p < 0.001), along with significant sex‑by‑age interactions (p = 0.001 for SMA; p = 0.003 for SMI), indicating that the decline in muscle mass with age differs between men and women. Univariate linear regression further quantified this difference: each additional year of age was associated with a decrease in SMA of 1.331 cm² in men versus 0.465 cm² in women, and a decrease in SMI of 0.415 cm²/m² in men versus 0.159 cm²/m² in women. LOESS‑smoothed plots visually confirmed a steeper and more consistent decline in men, particularly after age 65. These findings support the concept that aging men experience a more rapid loss of muscle quantity, potentially linked to a more dramatic decline in anabolic hormones like testosterone [18].

The impact of age and sex on muscle quality

Muscle quality encompasses the micro- and macroscopic structure and composition of muscle [48]. SMD, a composite measure of muscle tissue components [49, 50], serves as a key indicator of quality. Two‑way ANOVA showed significant main effects of sex and age group on SMD (both p < 0.001) but no significant sex‑by‑age interaction (p = 0.573), suggesting that the age-related decline in SMD follows a similar trajectory in both sexes. Linear regression confirmed comparable annual decreases (‑0.292 HU in men vs. ‑0.338 HU in women), and LOESS curves illustrated a moderate negative linear trend in both sexes. These results indicate that while women exhibit an earlier onset of SMD decline (beginning after menopause), the rate of decline once established is similar between sexes. This aligns with findings from Lorbergs [22] and Jung [23], suggesting that myosteatosis is an early and sensitive marker of muscle aging.

IMAT is a more sensitive marker of age-related myosteatosis than IMAC

Myosteatosis, the pathological fat infiltration into muscle, is a primary driver of declining muscle quality [7]. Our study found divergent aging patterns for two CT-derived metrics of fat infiltration [23, 24]. Two‑way ANOVA revealed significant main effects of sex and age group on IMAT (both p < 0.001) but no significant interaction (p = 0.393), indicating that IMAT increases with age at a comparable rate in both sexes. Linear regression supported this, with IMAT increasing annually by 0.131% in men and 0.198% in women. LOESS curves suggested a potential acceleration after age 70 in both sexes. In contrast, two‑way ANOVA showed no significant main effect of age group on IMAC (p = 0.394) and no sex‑by‑age interaction (p = 0.922). Linear regression also failed to detect significant age-related changes in IMAC in either sex after stratification. These findings support that IMAT is a more robust and direct imaging biomarker for tracking age-related myosteatosis than IMAC, consistent with Delmonico et al. [51, 52].

Clinical implications and practical applications

This study has significant clinical and translational value. First, it establishes SMD decline and IMAT increase as sensitive, early imaging markers of deteriorating muscle quality, particularly in postmenopausal women. This enables clinicians to use routine abdominal CT scans for opportunistic screening and risk stratification for sarcopenia and myosteatosis, offering a critical window for early intervention before functional decline becomes apparent.

Second, the observed sexual dimorphism provides a scientific basis for developing sex-specific intervention strategies. For men, interventions should prioritize preserving muscle mass through resistance training and adequate protein intake. For women, strategies should begin early in the postmenopausal period and focus on improving muscle quality and microstructure, which may include a combination of aerobic and resistance exercise and nutritional optimization [53, 54].

Finally, this study systematically validates the utility of "opportunistic CT" for population-level muscle health assessment. This approach maximizes the value of existing diagnostic data without additional cost or radiation, offering a practical pathway for implementing muscle health assessments in large hospital systems and routine health management, thereby advancing the prevention and monitoring of muscle aging-related conditions [16, 17].

Conclusions

In summary, using abdominal CT at the L3 level, this study examined how muscle mass and quality change with age and sex in men aged 50 years and older and in postmenopausal women. Our main findings are as follows.

First, both muscle mass (SMA, SMI) and muscle quality indicators (SMD, IMAT) decline with age, but the patterns differ by sex. Two-way ANOVA showed significant sex-by-age interactions for SMA and SMI (p = 0.001 and p = 0.003, respectively), but not for SMD (p = 0.573) or IMAT (p = 0.393). This suggests that while muscle quality declines at a similar rate in men and women, the loss of muscle mass follows distinct sex-specific trajectories.

Second, men lose muscle mass faster. According to univariate linear regression, each additional year of age corresponds to a decrease in SMA of 1.331 cm² in men (standardized β = −0.504) compared with 0.465 cm² in women (β = −0.267). For SMI, the annual decline is 0.415 cm²/m² in men (β = −0.482) versus 0.159 cm²/m² in women (β = −0.246). LOESS-smoothed curves further indicate that the decline in men steepens after age 65.

Third, women show earlier and more pronounced deterioration in muscle quality, mainly driven by an earlier increase in intermuscular fat (IMAT) and a steady decline in skeletal muscle density (SMD). Although the annual rate of SMD decline is similar between sexes (−0.292 HU in men vs. −0.338 HU in women), women exhibit a significant and progressive SMD decrease across all postmenopausal age groups, whereas men do not show a clear drop until after age 70. IMAT also rises earlier and reaches higher absolute levels in women.

Fourth, IMAT appears to be a more sensitive imaging biomarker of age-related myosteatosis than IMAC. Two-way ANOVA revealed no significant main effect of age group on IMAC (p = 0.394) and no sex-by-age interaction (p = 0.922). Sex-stratified linear regression detected no significant age-related change in IMAC in either men (p = 0.608) or women (p = 0.122).

Taken together, these results highlight a clear sexual dimorphism in muscle aging. They argue for sex-specific approaches in assessing, preventing, and treating sarcopenia: men should prioritize preserving muscle mass, while women may benefit from early interventions aimed at improving muscle quality, particularly countering myosteatosis.

Limitations

This study has several limitations. First, its cross-sectional design can only reveal statistical associations, not causal relationships, and cannot track individual aging trajectories. Longitudinal studies with repeated CT measurements are needed to pinpoint critical transition points and causal factors.

Second, the study population consisted of hospitalized patients with underlying conditions, which may limit the generalizability of our findings to healthy community-dwelling older adults. As noted in the Methods, this selected population enhances the internal validity of the findings by reducing confounding from undiagnosed illnesses, but it does not improve external validity or generalizability to other populations The presence of fractures, cardiovascular, or respiratory diseases could influence muscle variables. Although we strictly excluded direct muscle diseases, the overall health status of inpatients differs from healthy populations. Future studies should include a community-based control group to validate these findings. Furthermore, the sample sizes in some subgroups were relatively small (e.g., women aged ≤ 59 years, men aged ≥ 70 years), which may have reduced the statistical power to detect changes in some variables.

Third, CT cannot distinguish between intramyocellular and extramyocellular lipids, and skeletal muscle density (SMD) reflects a composite of muscle components rather than a pure measure of fatty infiltration. Moreover, this study lacked a correlation with functional measures such as grip strength and gait speed, which are essential for a definitive diagnosis of sarcopenia. Future research should integrate these functional assessments to establish clinically relevant cut-off values for the imaging variables identified here.

Fourth, while scanning variables were standardized, potential influences from different acquisition protocols cannot be completely ruled out. Additionally, this was a single-center study, and multi-center validation is warranted.

Authors’ contributions

D.X., M.H., and G.X.Y. contributed to data acquisition, analysis, and interpretation, and drafted the manuscript. L.R., F.F., Z.H., and G.L. participated in data collection and statistical analysis. W.Y.D. and H.Z. conceived and designed the study, supervised the research, and critically revised the manuscript for important intellectual content. All authors read and approved the final manuscript and agree to be accountable for all aspects of the work.

Funding

This study was supported by the Tianjin Key Medical Discipline (Specialty) Construction Project (Grant No. TJYXZDXK-026A). The funding body had no role in the study design, data collection, analysis, interpretation, or manuscript preparation.

Data availability

The datasets generated and/or analyzed during the current study are not publicly available due to patient privacy and hospital regulations but are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

This study was conducted in accordance with the Declaration of Helsinki and approved by the Ethics Committee of Tianjin Hospital. Due to the retrospective nature of the study and the use of fully anonymized clinical and imaging data, the requirement for written informed consent from participants was waived by the ethics committee.

Consent for publication

Not applicable. This study utilized retrospectively collected, anonymized abdominal CT images and clinical data, and no identifying personal information, photographs, or other clinical details that could compromise patient anonymity are presented in this manuscript. Therefore, informed consent for publication was not required.

Competing interests

The authors declare no competing interests.