The Relationship Between Intermuscular Fat and Physical Performance Is Moderated by Muscle Area in Older Adults

Abstract

Background

Age-related deposition of fat in skeletal muscle is associated with functional limitations. Skeletal muscle fat may be present in people with preserved muscle mass or accompanied by muscle wasting. However, it is not clear if the association between muscle fat deposition and physical performance is moderated by muscle mass.

Objective

To determine whether the association between midthigh intermuscular fat and physical performance is moderated by muscle area.

Methods

We performed a cross-sectional analysis of the Health, Aging, and, Body Composition (ABC) study data collected in 2002–2003 (n = 1897, women: 52.2%). Midthigh muscle cross-sectional area (by computed tomography) and physical performance measures were compared across quartiles of intermuscular fat absolute area. Moderation analysis was performed to determine the conditional effect of intermuscular fat on physical performance as a function of muscle area. Conditional effects were evaluated at three levels of muscle area (mean and ± 1 standard deviation [SD]; 213.2 ± 53.2 cm²).

Results

Simple slope analysis showed that the negative association between intermuscular fat area (cm²) and leg strength (N·m) was of greater magnitude (beta coefficient [b], 95% confidence interval [CI] = −0.288 [−0.427, −0.148]) in participants with greater muscle area (ie, 1 SD above the mean) compared to those with lower muscle area (ie, at mean [b = −0.12 {−0.248, 0.008}] or 1 SD below the mean [b = 0.048 {−0.122, 0.217}]). Similarly, the negative association of intermuscular fat with 400-m walk speed (m/s) and chair stand (seconds) was greater in those with higher muscle areas (p < .001) compared to those with lower muscle areas.

Conclusions

The association between higher intermuscular fat area and impaired physical function in aging is moderated by muscle area.

Age-related declines in muscle strength and physical function are important components of frailty (1,2) and predictors of future risk of mobility disability (3,4), loss of independence, and mortality (5–9). Alterations in muscle composition, which may manifest as a combination of muscle wasting and fat deposition, are among the major contributors to muscular dysfunction in the older adults (10–13). Muscle mass decreases by ~6% per decade after midlife (14) and accelerates to 1% per year in those aged 70 and older (15). Irrespective of changes in the muscle mass or body weight, an age-related increase in muscle fat by 48.5% in men and 29.0% in women was observed over 5 years follow-up in the Health, Aging, and, Body Composition (ABC) study cohort with an average age of 74 years (16). The accumulation of fat in skeletal muscles is a manifestation of the age-related body fat redistribution from subcutaneous sites to ectopic locations (17) and affects intermuscular (ie, visible fat within the fascia and between the muscles) (18) and intramuscular (ie, fat infiltration between muscle fibers) (19) as well as intramyocellular lipid content (ie, fat stored in droplets within muscle cells) (20,21). Previous studies suggest that these fat depots are associated with higher risks of developing muscle dysfunctions and mobility disability in the older adults (22,23).

Fat infiltration of skeletal muscle is a complex pathologic condition and in addition to occurring with aging, it may occur in a variety of diseases, including in cancer cachexia (24), spinal cord injury (25), metabolic dysfunction (eg, obesity and type 2 diabetes mellitus) (26,27), and muscle disuse (28). These diverse pathologic states may be associated with various levels of muscle loss. For instance, both acute (eg, short-term immobility) and chronic (eg, cancer cachexia and spinal cord injury) hypercatabolic conditions are associated with concurrent muscle wasting and muscle fat accretion (29,30). On the other hand, muscle fat accumulation in metabolic diseases is generally not associated with concurrent low muscle mass (26). While the role of age-related muscle wasting in deterioration of muscle function is extensively studied, the impact that muscle size has on the association between muscle fat depots and physical performance has not been addressed to our knowledge.

In the current study, we have examined the interaction between midthigh muscle and intermuscular fat areas with physical performance measures (gait speed, chair stand, standing balance, and leg strength) in participants of the Health ABC study to determine whether the association between intermuscular fat and physical performance varies by the amount of muscle area. We hypothesized that muscle area moderates the association between intermuscular fat and physical performance in community-dwelling older adults. Delineation of the interactions between muscle mass and fat deposition will advance our understanding of the role of muscle composition in the disablement pathway in aging and may ultimately allow for more effective interventions.

Study Participants and Methods

Participants

This study is a cross-sectional analysis of Year-6 data from the Health ABC study cohort collected in 2002–2003 (31). In the original cohort 3075 participants (women: 52% and blacks: 41%) were recruited from Memphis, Tennessee, and Pittsburgh, Pennsylvania. Participants were eligible to enroll if they (i) had no difficulty in walking ¼ mile, climbing up 10 steps, and performing daily living activities; (ii) did not report using a walking aid; and (iii) were free of cancer. Considering that the cohort consisted of healthy and well-functioning participants at baseline, we conducted this analysis using the sixth year follow-up data to ensure relatively high prevalence of intermuscular fat infiltration and functional decline. By Year 6, 1948 participants were still alive and returned for the examination, out of which 44 participants with no midthigh muscle assessment by computed tomography (CT) scan and seven participants with intermuscular fat area ≥ 80 cm² (ie, ≥5 standard deviation [SD]) were excluded from the analysis. The final analysis sample comprised 1897 participants (911 men, 986 women). All participants provided informed written consents. The research protocol was approved by the institutional ethics committees of the Universities of Pittsburgh and Tennessee.

Body Composition

Body composition was assessed by dual-energy X-ray absorptiometry (DXA) (QDR 4500A; Hologic Inc., Waltham, MA) using standard procedures, as previously reported (32). Quality assurance of the DXA measurements was performed by the use of daily and cross-calibration phantoms to ensue scanner reliability and adherence to the study protocol at both clinical centers. Appendicular lean mass was calculated as the sum of nonbone lean mass in arms and legs. Body weight (by calibrated balance-beam scale) and height (using Harpenden Stadiometer, Holtain Ltd, Crosswell, United Kingdom) were determined while participants were wearing a hospital gown and had removed their shoes. Waist circumference was measured with a flexible tape measure at the participant’s largest circumference.

Physical Performance

Average maximum isokinetic muscle torque of the knee extensors was assessed using isokinetic Kin-Com dynamometer (model 125 AP; Chattanooga, TN) on the right leg, unless contraindicated. Participants were excluded from knee extension strength test if they had stroke, uncontrolled hypertension, bilateral knee replacement, or severe bilateral knee pain (13). Lower extremity function was assessed by (i) five repeated timed chair stands (number/second, with those unable assigned a score of 0); (ii) timed standing balance (total time of all stands; 0–90 seconds); (iii) 6-m fast gait speed (m/s); (iv) time to walk 6 m with usual pace (seconds); and (v) long distance (400 m) usual walk speed (m/s), as described previously (33).

Midthigh Muscle Composition

Midthigh muscle cross-sectional area (cm²), intermuscular fat area (cm²), and muscle density (Hounsfield unit) were determined by CT, as described before (22). In brief, a single 10-mm axial image was obtained at the midpoint of the distance between the medial edge of the greater trochanter and the intercondyloid fossa. A development software (RSI Systems, Boulder, CO) was used to determine muscle and adipose tissue areas of the thigh. Intermuscular fat area was determined as the area of fat density within the deep fascial plane surrounding the thigh muscles. Muscle area was quantified as the total area of nonfat and nonbone tissues within the fascial border (16). Muscle density was determined as the skeletal muscle attenuation coefficient by averaging pixel densities in Hounsfield units within the region outlined on the images within the fascial plan (excluding intermuscular and visible intramuscular adipose tissue). A quality review was performed on each participant’s image to confirm the use of proper scan techniques and the appropriate quality of images for analysis.

Other Measures

Demographic variables, including sex, age, race, smoking history, education, and physical activity levels, were collected through questionnaires. Physical activity was assessed by self-report as total kcal/wk spent on walking and stairs (34). The presence of selected medical conditions was determined using algorithms based on physiological measures performed on participants (eg, electrocardiogram, blood pressure, ankle-arm index, spirometry), medication use, and self-report. A comorbidity index (ranging from 0 to 7) was assigned as the sum of the following conditions present in each participant: (i) chronic obstructive pulmonary disease, including chronic bronchitis, emphysema, or asthma; (ii) knee or hip arthritis; (iii) hypertension; (iv) diabetes mellitus: self-reported or physician diagnosis confirmed by the use of diabetic medications or fasting blood glucose ≥ 126 mg/dL; (v) cardiovascular disease, including stroke and coronary heart disease (myocardial infarction or angina), as assessed by the review of medical records; (vi) depression: score ≥ 10 on the modified version of the Center for Epidemiological Studies-Depression scale; (vii) cognitive impairment: modified Mini-Mental State Examination score < 80 (31).

Statistical Analyses

Descriptive statistics are shown as the median (25th and 75th percentiles), mean ± SEM, or frequency according to the sex-specific quartiles of intermuscular fat area. Participants’ characteristics were compared between the groups by chi-square (for categorical variables) or Kruskal–Wallis and one-way analysis of covariance (ANCOVA) tests (for continuous variables) to determine the statistical significance.

Next, we determined the influence of midthigh muscle area on the association, expressed as a slope, between intermuscular fat and physical function using the PROCESS macro (version 3.3) developed by Hayes (35). Moderation was tested using regression models in which the outcome (physical performance) was predicted from the predictor (intermuscular fat areas), the moderator (midthigh muscle area), and the interaction of these variables. PROCESS tool centers predictor and moderator at their mean values and shows the moderation effect in two ways; (i) simple slopes analysis and (ii) Johnson–Neyman method. Simple slope analysis shows the regression of physical function on intermuscular fat at 1 SD below, 1 SD above, and at the mean value of muscle area. Additionally, Johnson–Neyman method determines the strength of the associations between intermuscular fat area (predictor) and physical function (outcome) continuously across the values of the muscle area (moderator). The zone of significance indicates the values of the moderator at which the regression slope between predictor and outcome is significant. Statistical analyses were conducted by IBM SPSS Statistics version 26.0 for Windows. p < .05 was considered statistically significant.

Results

Participants’ Characteristics

A total of 1897 participants (age: 73–85 years, women: 52%, white race: 64%, and from Pittsburgh, Pennsylvania: 52%) from the Year-6 follow-up of the Health ABC study were included in the analyses. The characteristics of study participants by quartiles of absolute area of intermuscular fat (cm²) are shown in Table 1. Age and sex were comparable across quartiles of intermuscular fat. Participants with higher intermuscular fat were more likely to be black, current smoker, less educated, and had higher comorbidity scores and sedentary lifestyle.

Table 1.

Descriptive Characteristic by Sex-Specific Quartile (Q) of Intermuscular Fat Area

Intermuscular fat (Q), cm2	Q1 (n = 473)	Q2 (n = 475)	Q3 (n = 474)	Q4 (n = 475)	p Value
	Men < 17.2	Men: 17.3–23.6	Men: 23.7–30.6	Men ≥ 30.7
	Women < 15.8	Women: 15.9–21.2	Women: 21.3–28.2	Women: ≥ 28.3
Women, n (%)*	247 (52.2%)	248 (52.2%)	248 (52.3%)	248 (52.2%)	.100
Black, n (%)*	119 (25.2%)	138 (29.1%)	182 (38.4%)	238 (50.1%)	<.001
Never smoker, n (%)*	239 (53.3%)	221 (49.3%)	213 (47.1%)	210 (44.2%)	.020
< High school, n (%)*	63 (13.3%)	97 (20.5%)	105 (22.2%)	132 (27.9%)	<.001
	←‒‒‒‒‒‒‒‒‒Median (25th, 75th)‒‒‒‒‒‒‒‒‒→
Age, y	78 (76, 80)	78 (76, 81)	78 (76, 80)	78 (76, 80)	.150
Comorbidity score, 0–6	1 (1, 2)	2 (1, 3)	2 (1, 3)	2 (1, 3)	<.001
Physical activity, kcal/kg/wk	2.6 (0.3, 7.3)	2.5 (0.5, 7.1)	1.3 (0.1, 5.45)	0.8 (0.0, 4.2)	<.001
Anthropometrics
Height, m	1.6 (1.6, 1.7)	1.6 (1.6, 1.7)	1.7 (1.6, 1.7)	1.7 (1.6, 1.7)	.024
Weight, kg	63.0 (54.7, 70.7)	70.2 (62.3, 77.0)	76.3 (68.4, 84.3)	86.6 (77.2, 96.2)	<.001
BMI, kg/m2	23.2 (21.2, 25.3)	25.6 (23.9, 27.6)	27.6 (25.6, 30.4)	31.2 (28.3, 34.0)	<.001
Waist circumference, cm	93.4 (86.9, 99.3)	100.0 (94.5, 106.8)	105.2 (99.0, 112.7)	114.1 (106.9, 121.6)	<.001
DXA
Fat mass, %	30.5 (26.2, 36.7)	34.3 (29.4, 40.4)	37.1 (31.0, 42.0)	39.4 (34.9, 45.0)	<.001
Appendicular lean mass, kg	16.8 (13.6, 21.1)	17.7 (14.7, 22.0)	19.3 (15.8, 23.1)	21.3 (17.9, 24.8)	<.001
Midthigh CT
Intermuscular fat, cm2	13.5 (11.6, 15.0)	19.3 (17.9, 20.7)	25.8 (24.1, 27.6)	36.7 (32.6, 42.8)	<.001
Muscle CSA, cm2	186.7 (155.3, 233.8)	200.7 (167.3, 243.3)	212.3 (178.7, 251.2)	227.8 (194.5, 266.5)	<.001
Muscle density, HU	40.2 (36.9, 42.9)	36.8 (33.4, 40.3)	34.7 (31.0, 38.6)	30.8 (27.2, 35.0)	<.001
Physical performance
6-m fast walk, m/s	1.2 (1.0, 1.3)	1.1 (1.0, 1.3)	1.1 (0.9, 1.2)	1.0 (0.9, 1.2)	<.001
6-m usual walk, s	5.2 (4.6, 6.0)	5.3 (4.7, 6.0)	5.4 (4.8, 6.3)	5.9 (5.1, 6.8)	<.001
400-m usual walk, m/s	1.3 (1.1, 1.4)	1.2 (1.1, 1.4)	1.2 (1.1, 1.3)	1.1 (1.0, 1.2)	<.001
Chair stand, s	0.4 (0.3, 0.4)	0.4 (0.3, 0.4)	0.3 (0.3, 0.4)	0.3 (0.2, 0.4)	<.001
Standing balance time, 0–90	69.6 (48.3, 90.0)	66.1 (43.5, 79.7)	64.5 (38.3, 76.7)	61.9 (36.4, 70.3)	<.001
SPPB, 0.12	10 (9, 11)	10 (9, 11)	10 (8, 11)	9 (8, 10)	<.001
Maximum torque, N·m	79.0 (62.0, 107.3)	85.0 (65.0, 110.0)	89.0 (67.0, 114.0)	91.5 (70.0, 115.0)	<.001

Notes: p values by Kruskal–Wallis and * by chi-square. BMI = body mass index; CSA = cross-sectional area; HU = Hounsfield unit; N·m = Newton meter; s = second; SPPB = short physical performance battery; wk = week.

Anthropometrics and Muscle Composition

Participants in the higher quartiles of intermuscular fat area had higher body mass index, waist circumference, fat mass, and appendicular lean mass compared to those in the lower quartiles of intermuscular fat (Table 1). We observed a higher midthigh muscle cross-sectional area and a lower muscle density in participants in the higher quartiles of intermuscular fat compared to those in the lower quartiles.

Physical Performance

Table 2 shows mean ± SEM of physical performance by quartiles of intermuscular fat area adjusted for sex, race, education, smoking status, comorbidity score, age, physical activity, height, weight, and waist circumference. Participants in the higher quartiles of intermuscular fat performed worse in all mobility tests (ie, 6-m fast and usual walk and 400-m long walk), chair stand, standing balance test, and had lower leg strength compared to those in the lower quartiles of intermuscular fat both before (data not shown) and after controlling for confounding factors, Table 2.

Table 2.

Adjusted Mean ± Standard Error of Mean of Physical Performance by Quartile of Intermuscular Fat Area

	Q1	Q2	Q3	Q4	p Value
6-m fast walk, m/s	1.11 ± 0.02	1.10 ± 0.01	1.05 ± 0.01	1.00 ± 0.02	<.001
6-m usual walk, s	5.67 ± 0.08	5.68 ± 0.07	5.83 ± 0.07	6.17 ± 0.08	<.001
400-m usual walk, m/s	1.21 ± 0.01	1.18 ± 0.01	1.15 ± 0.01	1.09 ± 0.02	<.001
Chair stand, s	0.33 ± 0.01	0.33 ± 0.01	0.31 ± 0.01	0.28 ± 0.01	<.001
Standing balance time, 0–90	62.21 ± 1.50	59.85 ± 1.38	57.96 ± 1.35	54.06 ± 1.50	.001
SPPB, 0–12	9.55 ± 0.12	9.61 ± 0.11	9.13 ± 0.11	8.84 ± 0.12	<.001
Maximum torque, N·m	92.92 ± 1.48	92.66 ± 1.37	91.01 ± 1.34	85.62 ± 1.50	.001

Notes: p values were derived by one-way analysis of covariance (ANCOVA), adjusted for sex, race, education, smoking status, comorbidity score, age, physical activity, height, weight, waist circumference. m = meter; N·m = Newton meter; Q = quartile; s = second; SPPB = short physical performance battery.

Moderation Effect of Midthigh Muscle Area on the Association Between Intermuscular Fat and Physical Performance

Table 3 shows the moderation effect of midthigh muscle area on the relationship between intermuscular fat and physical function measures in final models adjusted for age, sex, race, education, physical activity, height, weight, waist circumference, and comorbidity score. Our analysis revealed that muscle area significantly moderated the association of intermuscular fat with leg strength, 400-m usual walk speed, and chair stand (Table 3). However, midthigh muscle area did not moderate the association of intermuscular fat with 6-m fast gait speed, time to walk 6 m with usual pace, and standing balance based on the statistical nonsignificance and small values of the relevant interaction terms in these models.

Table 3.

The Association Between Physical Performance Measures, Intermuscular Fat, and Midthigh Muscle Areas, Based on Moderation Analysis

Predictor*	All
	b (SE)	p Value
Leg strength, N·m
Intermuscular fat, cm2	−0.1201 (0.0652)	.066
Muscle CSA, cm2	0.3716 (0.0205)	<.001
Interaction	−0.0032 (0.0008)	<.001
400-m usual walk, m/s
Intermuscular fat, cm2	−0.0029 (0.0007)	<.001
Muscle CSA, cm2	0.0018 (0.0002)	<.001
Interaction	−2.3 × 10−5 (9 × 10−6)	.008
6-m usual walk, s
Intermuscular fat, cm2	0.018 (0.0038)	<.001
Muscle CSA, cm2	−0.009 (0.0012)	<.001
Interaction	−0.0001 (0.0000)	.158
6-m fast walk, m/s
Intermuscular fat, cm2	−0.0044 (0.0007)	<.001
Muscle CSA, cm2	0.0016 (0.0002)	<.001
Interaction	1.1 × 10−5 (9 × 10−6)	.210
Chair stand, s
Intermuscular fat, cm2	−0.0014 (0.0004)	<.001
Muscle CSA, cm2	0.0017 (0.0001)	<.001
Interaction	1 × 10−5 (5 × 10−6)	.041
Standing balance, 0–90
Intermuscular fat, cm2	−0.2109 (0.0677)	<.001
Muscle CSA, cm2	0.1724 (0.022)	<.001
Interaction	−0.0002 (0.0009)	.800

Notes: b = beta coefficient; CSA = cross-sectional area; N·m = Newton meter; s = second; SE = standard error.

*Models are adjusted for age, sex, race, physical activity, height, weight, waist circumference, education level, comorbidity index, and smoking status.

Figure 1 shows the simple slope analyses of the conditional effect of intermuscular fat on leg strength (top panel), 400-m walking speed (middle panel), and chair stand (lower panel) at three levels (mean value and ± 1 SD) of muscle area, that is, 213.2 ± 53.2 cm². The negative association between intermuscular fat area (cm²) and leg strength (N·m) was of greater magnitude (beta coefficient [b] ± 95% confidence interval [CI] = −0.288 [−0.427, −0.148]) in participants with greater midthigh muscle area (ie, 1 SD above the mean) compared to those with lower muscle area (ie, at the mean value [b = −0.12 {−0.248, 0.008}] or 1 SD below the mean [b = 0.048 {−0.122, 0.217}]). Additionally, the negative association between intermuscular fat area and 400-m usual walk speed (m/s) was of greater magnitude in participants with higher midthigh muscle area (cm²) (ie, 1 SD above the mean [b = −0.004, 95% CI {−0.006 to −0.003}] and at the mean value [b = −0.003, 95% CI {−0.004 to −0.002}]) compared to those with lower muscle area. Likewise, moderation analysis indicated a significant association between intermuscular fat area and chair stand (seconds) in participants with greater midthigh muscle area, that is, 1 SD above the mean (b = −0.002, 95% CI [−0.003 to −0.001]) or at the mean value (b = −0.001, 95% CI [−0.002 to −0.001]), but not in those at 1 SD below the mean value of the muscle area (cm²).

Figure 1.

Moderation effects of midthigh muscle area on the association of intermuscular fat with leg strength, 400-m usual walk, and chair stand time. Simple slopes equations of the regression for intermuscular fat as a predictor of physical function at three levels (mean value [213.2 cm²], low and high muscle areas [ie, mean ± 1 SD; 213.2 ± 53.2 cm²]). Regression models are adjusted for age, sex, race, education, physical activity, height, weight, waist circumference, and comorbidity score; b = slope; **p < .001.

Figure 2 shows the regression slope estimate and 95% CI for the relationship between intermuscular fat and physical function parameters (leg strength, 400-m walk and chair stand) across different values of the moderator (ie, midthigh muscle area [cm²]). Vertical lines in each graph represent the boundaries of the significant zones of the regression coefficient (b) between intermuscular fat and physical function tests. The Johnson–Neyman technique revealed that the inverse relationship between intermuscular fat and leg strength (Figure 2, top panel) became greater as the muscle area values increases across the continuum in a coherent fashion, reaching the significance region at ≥219.1 cm² (p < .05). The negative effect of intermuscular fat on leg strength was increasingly stronger at higher muscle areas. Similarly, the magnitude of the negative relationships between intermuscular fat and 400-m walk (Figure 2, middle panel) as well as chair stand (Figure 2, lower panel) continuously increased with higher muscle areas, reaching the significance level at ≥169.3 cm² and ≥166.5 cm² (p < .05), respectively. However, below these transition points, no significant relationships were observed between intermuscular fat and 400-m walk or chair stand.

Figure 2.

Regression slope estimate and 95% confidence interval (CI) for the relationship between intermuscular fat area (X, cm²) and physical function (Y; leg strength [N·m], 400-m walk [m/s], and chair stand [seconds]) at values of muscle area (cm²) by Johnson–Neyman method. All models were adjusted for age, sex, race, education, physical activity, height, weight, waist circumference, and comorbidity score.

Stratified analysis by sex (Supplementary Table 1) showed similar b coefficients for the interactions between muscle area and intermuscular fat with leg strength (b = −0.0027 in both men and women) and chair stand (b = −0.000009 in men and b = −0.000017 in women), although they did not reach the statistical significance levels (p < .1). Race-stratified analysis (Supplementary Table 1) showed stronger interactions in whites compared to blacks between muscle area and intermuscular fat with leg strength (b = −0.0042 vs −0.0019) and 400-m usual walk (b = −0.000048 vs −0.000013). In contrast to intermuscular fat area, midthigh subcutaneous fat area (cm²) was not associated with physical performance and no interaction was observed between subcutaneous fat area and muscle area on physical performance (Supplementary Table 2).

Discussion

The results of the current study in community-dwelling older adults showed that muscle area moderates the association between intermuscular fat area and physical performance. Specifically, the negative association between higher intermuscular fat area with leg strength, 400-m gait speed, and chair stand time was of greater magnitude in participants with greater midthigh muscle area. Although, low midthigh muscle area was independently associated with reduced physical function, this relationship was not affected by the extent of intermuscular fat deposition in people with lower muscle area.

It has been shown that midthigh intermuscular fat area increases with aging in community-living older men and women, irrespective of changes in muscle mass and body weight (16). Age-related increase in intermuscular fat has been shown to be modified by exercise (36,37), diet (38), or both (39,40). The increase in skeletal muscle lipid content has also been observed in various pathological disorders, such as obesity, diabetes, spinal cord injury, and cancer cachexia, regardless of the presence or absence of muscle wasting (41). The increase in the intermuscular fat is associated with muscle weakness and is a significant predictor of mobility disability in aging (41). Whether the association between intermuscular fat and physical function varies in individuals with preserved or lost muscle tissue is not clear.

The current study demonstrated the presence of an interaction between muscle area and intermuscular fat with physical function. The negative association between intermuscular fat and impaired physical function was of greater magnitude in participants with higher midthigh muscle cross-sectional area. In other words, we observed that as muscle size increased, the relationship between intermuscular fat and physical function became more negative. For instance, older individuals with the same levels of intermuscular fat area (eg, 75th percentile; 49.2 cm²) but different thigh muscle areas; 187.2 cm² (25th percentile) versus 308.2 cm² (75th percentile) have different muscle strengths, 54.0 versus 80 N·m, respectively. This finding is similar to a previous report (16), which showed a negative association between intermuscular fat and muscle strength in participants who were weight-stable during the 5-year follow-up of the Health ABC study. Changes in the local muscle environment, due to the proximity of intermuscular fat to muscle fibers, is a potential mechanism contributing to reduced muscle strength and increased risk of physical limitation (41). The increase in the intermuscular fat may affect muscle quality and cause mechanical changes in the muscle, such as changes in the orientation of muscle fibers, leading to difficulties in functional activities (42). Additionally, intermuscular fat, similar to visceral adipose tissue, is related to an increase in proinflammatory cytokines (43,44), which are inversely associated with muscle function (45).

A novel aspect of our study is that intermuscular fat accumulation does not significantly affect physical performance in individuals with low muscle area, in whom muscle area is the main determinant of physical function. Low muscle mass is significantly associated with functional impairment (46) and is a predictor of physical disability (47) in older adults. It has been shown that age-related decline in muscle strength is predominantly related to the lower muscle mass, rather than fat mass as determined by DXA, in the Health ABC study (13). However, it is noteworthy that the relationship between skeletal muscle mass alone and physical function is complex, as there are other important factors such as low muscle strength (48) and power (49) that contribute to low physical function in aging. The increased intermuscular fat accompanied by muscle wasting is an aberrant adipogenic reparative response that is observed in degenerative conditions (eg, muscular dystrophies) (50), where adipose tissue and fibrosis “replace” the attenuated muscle. Conversely, increased intermuscular fat in metabolic diseases (eg, obesity) is generally not associated with concurrent low muscle mass, which represents adipose tissue “infiltration” of intact muscles (26). Age-related changes in muscle composition is complex and the contribution of these two mechanisms in the accumulation of muscle fat depots in older individuals remains to be fully understood.

Strength and Limitations

Recruiting of a large number of community-dwelling older adults from different sexes and racial groups increases the power and generalizability of our study. However, the inclusion of relatively healthy and well-functioning older adults limits our ability to generalize our conclusions to frail or institutionalized older individuals. Another limitation of this study is that fat accumulation and muscle area were measured at a single level of midthigh and may not incompletely reflect the whole thigh muscle composition. Nevertheless, one of the major strengths of this study is addressing the unique interconnection between muscle fat deposition and muscle mass.

In summary, this study extends the scope of the previous reports by showing that muscle area moderates the association between greater intermuscular fat levels and impaired physical performance. Further, low muscle area was associated with low physical function, and this relationship was not affected by the amount of intermuscular fat deposition in participants with low muscle area. This finding may have implications with respect to the reversibility of intermuscular fat deposition and its functional consequences in people with persevered or wasted muscle. It may be speculated that individuals with high intermuscular fat area in the presence of low muscle mass area may experience a faster and less reversible decline in physical function compared to those with higher muscle area. Understanding the key biological alterations in the muscle composition during aging will help improve targeted approaches to prevent or reverse muscle fat retention and muscle dysfunction.

Funding

This work was supported by National Institute on Aging (NIA) contracts N01-AG-6-2101; N01-AG-6-2103; N01-AG-6-2106; NIA grant R01-AG028050, and National Institute of Nursing Research grant R01-NR012459. This research was funded in part by the Intramural Research Program of the National Institutes of Health (NIH), National Institute on Aging. A.J.S. was supported by a career development award from the NIH/NIA (K01 AG057726) and the University of Pittsburgh Claude D. Pepper Older Americans Independence Center (5P30 AG024827). S.F. is supported by the Epidemiology of Aging training grant at the University of Pittsburgh (NIA T32-AG0001810).

Author Contributions

The authors’ contributions were as follows—S.F. and A.B.N.: designed the research; A.B.N., S.B.K., and B.H.G. designed and conducted the Health ABC study; S.F.: analyzed the data and wrote the manuscript; R.M.B. and A.J.S.: provided statistical recommendation; S.F. and A.B.N. had primary responsibility for the final content. All authors: read and approved the final manuscript.

Conflict of Interest

None declared.