Human skeletal muscle-specific atrophy with aging: a comprehensive review

Abstract

Age-related skeletal muscle atrophy appears to be a muscle group-specific process, yet only a few specific muscles have been investigated and our understanding in this area is limited. This review provides a comprehensive summary of the available information on age-related skeletal muscle atrophy in a muscle-specific manner, nearly half of which comes from the quadriceps. Decline in muscle-specific size over ∼50 yr of aging was determined from 47 cross-sectional studies of 982 young (∼25 yr) and 1,003 old (∼75 yr) individuals and nine muscle groups: elbow extensors (−20%, −0.39%/yr), elbow flexors (−19%, −0.38%/yr), paraspinals (−24%, −0.47%/yr), psoas (−29%, −0.58%/yr), hip adductors (−13%, −0.27%/yr), hamstrings (−19%, −0.39%/yr), quadriceps (−27%, −0.53%/yr), dorsiflexors (−9%, −0.19%/yr), and triceps surae (−14%, −0.28%/yr). Muscle-specific atrophy rate was also determined for each of the subcomponent muscles in the hamstrings, quadriceps, and triceps surae. Of all the muscles included in this review, there was more than a fivefold difference between the least (−6%, −0.13%/yr, soleus) to the most (−33%, −0.66%/yr, rectus femoris) atrophying muscles. Muscle activity level, muscle fiber type, sex, and timeline of the aging process all appeared to have some influence on muscle-specific atrophy. Given the large range of muscle-specific atrophy and the large number of muscles that have not been investigated, more muscle-specific information could expand our understanding of functional deficits that develop with aging and help guide muscle-specific interventions to improve the quality of life of aging women and men.

INTRODUCTION

A decline in muscle mass is a well-accepted tenet associated with human aging (1), and has been investigated at the whole body, specific body region, specific muscle, and myocellular levels (2–7). Given the large number of skeletal muscles in the human body with distinct functions and contributions to overall daily activities, it is noteworthy that the number of specific muscles that have been investigated is relatively limited. In fact, muscle-specific aging studies have largely focused on the quadriceps or the vastus lateralis subcomponent of this muscle (1, 8–10). This is understandable given the ability to obtain high-quality size (via noninvasive imaging) and functional measurements of the quadriceps, coupled with the relative ease and safety of obtaining a biopsy of the vastus lateralis to perform simultaneous myocellular studies (6, 11–15). Exercise studies focused on mitigating age-related atrophy are also easily designed to focus on the quadriceps (e.g., knee extension resistance exercise or cycling) (10, 16–20). As a result, findings from the quadriceps (and vastus lateralis) investigations are generalized to the other muscles of the body with regard to the aging process and responsiveness to exercise training interventions. However, there is some evidence to suggest that age-related atrophy is muscle specific (2, 21–25) and that the responsiveness to exercise training is also muscle specific (2, 26–28).

Thus, more information is needed about an expanded number of specific muscles of the body with respect to the aging process. Muscle-specific information could improve our understanding of functional deficits that develop with aging and help guide muscle-specific interventions to improve the quality of life of aging women and men. The purpose of this review article was to provide a compilation of the current state of knowledge regarding muscle-specific atrophy in humans associated with the aging process.

LITERATURE REVIEW PARAMETERS AND SUMMARY

Numerous studies have quantified whole muscle atrophy with aging using computed tomography (CT), magnetic resonance imaging (MRI), or ultrasound. Of these, studies with cross-sectional designs that examined the human skeletal muscle cross-sectional area (CSA) or volume of generally healthy young (mean or median age of 20–29 yr) and old (mean or median age of 70–79 yr) individuals were compiled and the rate of muscle atrophy per year was determined for each muscle group. Although several studies reported muscle size comparisons between young and older (≥80 yr) individuals (29–33), these articles were excluded to allow assembling groups with a similar age across muscle groups. Studies that only reported muscle thickness, physiological CSA, or normalized muscle size (e.g., skeletal muscle index: muscle size/height²) were also excluded. Three muscle imaging methods (CT, MRI, and ultrasound) were considered equally in determining the rate of muscle loss, although ultrasound has only been validated against CT or MRI since the 1990s (34). As a result, studies on elbow extensors (anconeus, supinator, and triceps brachii), elbow flexors (biceps brachii, brachialis, brachioradialis, and pronator teres), paraspinals (erector spinae, lumbar multifidus, and quadratus lumborum), psoas (iliopsoas or psoas major), hip adductors (adductor brevis, adductor longus, adductor magnus, gracilis, pectineus, and sartorius), hamstrings (biceps femoris, semimembranosus, and semitendinosus), quadriceps femoris (rectus femoris, vastus lateralis, vastus intermedius, and vastus medialis), dorsiflexors (extensor digitorum longus, extensor hallucis longus, peroneus tertius, and tibialis anterior), and triceps surae (gastrocnemius and soleus) were included. For hamstrings, quadriceps, and triceps surae, studies that also assessed the size of subcomponent muscles within the muscle group were also compiled to determine the rate of atrophy of each muscle.

Similarly, longitudinal studies that examined human skeletal muscle CSA or volume of generally healthy older individuals were compiled and the rate of muscle atrophy per year was determined for each muscle group. The goal was to include studies near the mean or median age of the older cross-sectional cohort (i.e., 70–79 yr) either at the time of baseline measurement or longitudinal follow-up. Studies with a follow-up period of less than 1 yr were excluded. To provide comparisons across muscle groups with similar starting and longitudinal follow-up age ranges (∼65–75 yr), several studies of the quadriceps were excluded (12, 35–38). As a result, studies on paraspinals, psoas, hamstrings, and quadriceps femoris were included.

With these literature review criteria, data from 47 studies, including an accumulated total of 982 young and 1,003 old individuals were included in this review for cross-sectional comparisons between young and old individuals. Seventeen studies consisting of 717 subjects examined quadriceps, accounting for nearly 40% of the data included in this review. On the contrary, only a limited number of studies were found for each of the other eight muscle groups included in this review. Paraspinals and psoas included over 200 individuals, whereas elbow extensors and flexors only included 58 and 76 individuals, respectively. Paraspinals and psoas included more women (19% men and 17% men, respectively), whereas the other muscle groups had more men included (57%–83% men). In addition, one study for each of the four muscle groups (paraspinals, psoas, hamstrings, and quadriceps) was included for the longitudinal data on aging human skeletal muscle. All the muscle atrophy rate determinations were normalized to the sample size of each study, yet the studies with a small sample size tend to have relatively larger variability. Thus, data from muscle groups with a small total sample size should be interpreted with caution.

Cross-sectional comparisons for skeletal muscle size for the quadriceps and the other muscle groups are summarized in Tables 1 and 2, respectively. In addition, Table 3 summarizes the rate of muscle atrophy in the subcomponent muscles. Atrophy rates across all muscles ranged from −0.13%/yr to −0.66%/yr. Figure 1 presents the annualized rates of atrophy from these tables and percent changes over 50 yr of aging (∼25–75 yr) as most of the studies covered approximately this age range. Because many studies in the aging literature present an overall percent change in muscle mass, this atrophy rate per 50 yr and the rate per year will be discussed later. General muscle fiber type characteristics from separate biopsy and autopsy studies of the muscles included in this review are presented in Table 4. Sex-specific muscle atrophy rates for six muscle groups that had data to allow for this comparison are summarized in Table 5. The longitudinal studies are summarized in Table 6.

Table 1.

Summary of cross-sectional studies of human skeletal muscle aging in quadriceps femoris

Study	Young		Old		ΔAge	Measurement	Method	%Δ	%Δ/yr
	n	Age	n	Age
Young et al. 1984 (39)	25 W	20–29	25 W	71–81	50	CSA	US	−33	−0.66
Young et al. 1985 (40)	12 M	21–28	12 M	70–79	50	CSA	US	−25	−0.50
Overend et al. 1992 (23)	13 M	25 ± 5	12 M	71 ± 5	46	CSA	CT	−22	−0.49
Rutherford and Jones 1992 (41)	31 W	20–29	11 W	70–82	50	CSA	CT	−23	−0.46
Takahashi et al. 2006 (42)	35 W	21 ± 2	35 W	74 ± 3	53	CSA	MRI	−28	−0.52
Haus et al. 2007 (13) and (6, 7)	10 M, 10 W	25 ± 4	10 M, 12 W	78 ± 5	53	CSA/Volume	MRI	−26	−0.49
Kilgour et al. 2013 (43)	19 M	22 ± 2	33 M	70 ± 4	48	CSA	MRI	−27	−0.57
Maden-Wilkinson et al. 2013 (44) and (45–48)	20 M	22 ± 3	25 M	72 ± 5	50	Volume	MRI	−32	−0.64
	18 W	22 ± 2	28 W	72 ± 5	50			−28	−0.56
Nilwik et al. 2013 (49)	25 M	23 ± 5	26 M	71 ± 5	48	CSA	CT	−15	−0.31
Konopka et al. 2014 (50) and (17)	7 M	20 ± 3	6 M	74 ± 7	54	CSA	MRI	−15	−0.28
Ghosh et al. 2014 (51)	6 M, 7 W	26 ± 4	8 M, 4 W	74 ± 7	48	Volume	MRI	−20	−0.40
Rudroff et al. 2014 (52)	6 M	26 ± 6	6 M	77 ± 6	51	Volume	CT	−41	−0.81
Piasecki et al. 2016 (53)	22 M	25 ± 5	20 M	71 ± 6	46	CSA	MRI	−33	−0.71
Yoshiko et al. 2017 (25) and (54, 55)	8 M, 7 W	21 ± 0.4	7 M, 8 W	71 ± 4	50	Volume	MRI	−37	−0.74
Chambers et al. 2020 (2)	10 M	25 ± 3	10 M	75 ± 3	50	CSA/Volume	MRI	−30	−0.60
	10 W	25 ± 3	10 W	75 ± 3	50			−37	−0.73
Ogawa et al. 2021 (24)	20 M	23–27	20 M	67–72	50	CSA	MRI	−24	−0.48
Yagi et al. 2022 (56)	17 M	20–39	18 M	60–87	49	CSA	MRI	−18	−0.36
	15 W	20–38	18 W	62–84	46			−20	−0.43
Total: 17 studies*	353	24	364	73	50				−0.53

Age (yr) is presented as means ± SD or range. Muscle size changes (%Δ and %Δ/yr) were averaged when the study reports both muscle cross-sectional area (CSA) and volume. CT, computed tomography; M, men; MRI, magnetic resonance imaging; US, ultrasound; W, women.

*Boldface type indicates data are total n size or average values normalized to the n size of each study. For studies listed with more than one reference, the primary reference used to generate the data in the table is listed first, followed by other references that also contain data from the same study.

Table 2.

Summary of cross-sectional studies of human skeletal muscle aging in nonquadriceps muscle groups

Study	Young		Old		ΔAge	Measurement	Method	%Δ	%Δ/yr
	n	Age	n	Age
Elbow extensors
Klein et al. 2002 (57) and (58)	20 M	23 ± 3	10 M	77 ± 1	54	CSAa	MRI	−20	−0.38
Vidt et al. 2012 (59) and (60)	5 M, 5 W	29 ± 5	10 M, 8 W	75 ± 4	47	Volume	MRI	−19	−0.40
Total: 2 studies*	30	26	28	76	50				−0.39
Elbow flexors
Klein et al. 2002 (57) and (58)	20 M	23 ± 3	10 M	77 ± 1	54	CSAb	MRI	−15	−0.28
Vidt et al. 2012 (59) and (60)	5 M, 5 W	29 ± 5	10 M, 8 W	75 ± 4	47	Volume	MRI	−20	−0.43
Smart et al. 2018 (61)	9 M	23 ± 2	9 M	77 ± 5	54	CSAb	US	−26	−0.48
Total: 3 studies*	39	25	37	76	51				−0.38
Paraspinals
Ma et al. 2014 (62)	21 W	29 ± 3	42 W	73 ± 4	44	CSAc	MRI	−19	−0.42
Takayama et al. 2016 (63)	20 M, 20 W	10–29	30 M, 30 W	60–88	55	CSAd	MRI	−31	−0.56
Peng et al. 2020 (64)	69 W	27 ± 2	25 W	74 ± 3	47	CSAd	CT	−20	−0.42
Total: 3 studies*	130	25	127	74	49				−0.47
Psoas
Takahashi et al. 2006 (42)	35 W	21 ± 2	35 W	74 ± 3	53	CSAe	MRI	−37	−0.70
Ma et al. 2014 (62)	21 W	29 ± 3	42 W	73 ± 4	44	CSAe	MRI	−19	−0.43
Yagi et al. 2022 (56)	17 M	20–39	18 M	60–87	49	CSAf	MRI	−28	−0.58
	15 W	20–38	19 W	62–84	47			−28	−0.60
Total: 3 studies*	88	25	114	74	48				−0.58
Hip adductors
Hogrel et al. 2015 (45)	18 M	24 ± 3	19 M	74 ± 3	50	Volume	MRI	−17	−0.33
	16 W	24 ± 3	19 W	75 ± 3	51			−13	−0.26
Yoshiko et al. 2017 (25)	8 M, 7 W	21 ± 0.4	7 M, 8 W	71 ± 4	50	Volume	MRI	−27	−0.55
Ogawa et al. 2021 (24)	20 M	23–27	20 M	67–72	50	CSA	MRI	−14↔	−0.28↔
Total: 3 studies*	69	23	73	73	50				−0.27
Hamstrings
Overend et al. 1992 (23)	13 M	25 ± 5	12 M	71 ± 5	46	CSA	CT	−14↔	−0.30↔
Hogrel et al. 2015 (45) and (44, 46,47, 48)	18 M	24 ± 3	19 M	74 ± 3	50	Volume	MRI	−19	−0.37
	16 W	24 ± 3	19 W	75 ± 3	51			−16	−0.32
Palmer and Thompson 2017 (66)	15 M	25 ± 3	15 M	72 ± 5	51	CSA	US	−17	−0.36
Yoshiko et al. 2017 (25) and (54, 55)	8 M, 7 W	21 ± 0.4	7 M, 8 W	71 ± 4	50	Volume	MRI	−28	−0.56
Ogawa et al. 2021 (24)	20 M	23–27	20 M	67–72	50	CSA	MRI	−30	−0.60
Total: 5 studies*	97	23	100	72	49				−0.39
Dorsiflexors
Kent-Braun et al. 2000 (67)	11 W	29 ± 4	10 W	73 ± 6	44	CSA	MRI	−11	−0.26
Hasson et al. 2011 (68)	6 M	27 ± 3	6 M	73 ± 5	46	Volume	MRI	+0.4↔	+0.01↔
	6W	26 ± 3	6 W	70 ± 5	44			−10	−0.23
Barber et al. 2013 (69)	10 M, 8 W	27 ± 3	9 M, 7 W	70 ± 3	43	Volumeg	MRI	−22	−0.51
Christie et al. 2014 (70)	10 M, 10 W	24 ± 2	9 M, 9 W	73 ± 6	50	CSA	MRI	+3↔	+0.06↔
Power et al. 2014 (71)	12 M	24 ± 3	6 M	79 ± 3	54	CSA/Volumeg	MRI	−17	−0.32
Piasecki et al. 2016 (72)	18 M	26 ± 4	14 M	71 ± 4	45	CSAg	MRI	−11↔	−0.25↔
Total: 6 studies*	91	26	76	73	46				−0.19
Triceps surae
Morse et al. 2004 (73) and (74–77)	14 M	25 ± 5	21 M	74 ± 4	49	CSA/Volume	MRI	−18	−0.36
Hasson et al. 2011 (68)	6 M	27 ± 3	6 M	73 ± 5	46	Volume	MRI	−18	−0.39
	6W	26 ± 3	6 W	70 ± 5	44			−19	−0.43
Barber et al. 2013 (69)	10 M, 8 W	27 ± 3	9 M, 7 W	70 ± 3	43	Volume	MRI	−16	−0.38
Chambers et al. 2020 (2)	10 M	25 ± 3	10 M	75 ± 3	50	Volume	MRI	−13↔	−0.26↔
	10 W	25 ± 3	10 W	75 ± 3	50			−32	−0.63
Pinel et al. 2021 (78)	13 M, 8 W	25 ± 4	9 M, 6 W	70 ± 2	46	Volume	MRI	−1↔	−0.03↔
Total: 5 studies*	85	25	84	72	47				−0.28

Age (yr) is presented as means ± SD or range. Muscle size changes (%Δ and %Δ/yr) were averaged when the study reports both muscle cross-sectional area (CSA) and volume. CT, computed tomography; M, men; MRI; magnetic resonance imaging; US; ultrasound; W, women.

*Boldface type indicates data are total n size or average values normalized to the n size of each study; ^↔P > 0.05, considered as zero for total muscle atrophy rate determination; ^atriceps brachii only; ^bbiceps brachii only; ^cerector spinae and quadratus lumborum; ^derector spinae and multifidus; ^epsoas major; ^filiopsoas; ^gtibialis anterior only. For studies listed with more than one reference, the primary reference used to generate the data in the table is listed first, followed by other references that also contain data from the same study.

Table 3.

Summary of human skeletal muscle aging in subcomponent muscles

Muscle Group (References)	Young		Old		ΔAge*	Muscles	%Δ/yr*
	n	Age*	n	Age*
Hamstrings (45)	18 M	24	19 M	74	51	Semimembranosus	−0.40
	16 W		19 W			Semitendinosus	−0.39
						Biceps femoris long head	−0.28
						Biceps femoris short head	−0.23
Quadriceps (6, 45, 52)	29 M	24	30 M	76	51	Rectus femoris	−0.66
	21 W		24 W			Vastus lateralis	−0.59
						Vastus intermedius	−0.49
						Vastus medialis	−0.48
Triceps surae (68, 69, 74, 78)	44 M	26	36 M	71	45	Gastrocnemius	−0.41
	22 W		19 W			Soleus	−0.13

Age is presented in years. M, men; W, women.

*Average values normalized to the n size of each study.

Figure 1.

Summary of skeletal muscle-specific atrophy with aging from cross-sectional studies. Muscle atrophy rates (%Δ/yr) are from Tables 1, 2 and 3. Percentage in each bar is the yearly rate multiplied by 50 to equally represent the muscle atrophy over 50 yr of aging from ∼25 yr to ∼75 yr across muscle groups. A: nine muscle groups included in this review. B: subcomponent muscles. Add, hip adductors; BFL, biceps femoris long head; BFS, biceps femoris short head; DF, dorsiflexors; EE, elbow extensors; EF, elbow flexors; Gas, gastrocnemius; Ham, hamstrings; Para, paraspinals; Pso, psoas; Quad, quadriceps; RF, rectus femoris; SM, semimembranosus; Sol, soleus; ST, semitendinosus; TS, triceps surae; VI, vastus intermedius; VL, vastus lateralis; VM, vastus medialis.

Table 4.

Human skeletal muscle fiber type distribution for muscles included in this review

Muscle (References)	Age, yr	Muscle Fiber Type Distribution
		Type I, %	Type II, %
Elbow extensors
Triceps brachii (79–82)	17–40	36 (22–50)	64 (50–78)
Elbow flexors
Biceps brachii (79–81)	17–40	46 (45–46)	54 (53–55)
Brachioradialis (79, 81)	17–40	44 (40–47)	56 (51–60)
Paraspinals
Erector spinae (81, 83–87)	16–44	61 (57–66)	39 (33–47)
Lumbar multifidus (83, 84, 87–89)	17–44	59 (54–65)	41 (36–46)
Psoas
Psoas major/Iliopsoas (81, 90)	17–35	45 (40–49)	55 (51–60)
Hip adductors
Adductor magnus (91)	37–76	55	45
Hamstrings
Biceps femoris short and long heads (79, 81, 92)	17–40	54 (47–67)	46 (33–53)
Semimembranosus (91)	37–76	49	51
Semitendinosus (93)	17–34	48	52
Quadriceps
Rectus femoris (81)	17–30	38	62
Vastus lateralis (80–82, 94–96)	15–30	46 (36–59)	54 (41–64)
Vastus intermedius (91, 97)	21–83	51 (47–54)	49 (46–53)
Vastus medialis (81)	17–30	53	47
Dorsiflexors
Tibialis anterior (79, 81, 98)	17–40	74 (73–76)	26 (24–27)
Triceps surae
Gastrocnemius (79, 81, 99)	17–40	59 (48–70)	41 (30–52)
Soleus (79–82, 94)	17–40	81 (66–91)	19 (9–35)

Data are presented as mean and a range of the mean data when multiple studies are presented. Type II fibers included type IIa and type IIx where applicable. Where possible, references were included that represented young to middle-aged men and women to minimize any potential age influence.

Table 5.

Summary of sex-specific human skeletal muscle aging

Muscle Group (References)	Sex	Young		Old		ΔAge*	%Δ/yr*
		n	Age*	n	Age*
Psoas (42, 56, 62)	Men	17	27	18	76	49	−0.58
	Women	71	25	96	73	48	−0.58
Hip adductors (24, 45)	Men	38	23	39	73	50	−0.16
	Women	16	24	19	75	51	−0.26
Hamstrings (23, 24, 45, 66)	Men	66	23	66	72	49	−0.37
	Women	16	24	19	75	51	−0.32
Quadriceps (2, 6, 23, 24, 39–44, 49, 50, 52, 53, 56)	Men	171	24	188	73	49	−0.52
	Women	139	23	132	74	50	−0.55
Dorsiflexors (67, 68, 71,72)	Men	36	26	26	74	48	−0.09
	Women	17	28	16	72	44	−0.25
Triceps surae (2, 68, 73)	Men	30	25	37	74	49	−0.26
	Women	16	25	16	73	48	−0.56

*Average values normalized to the n size of each study.

Table 6.

Summary of longitudinal studies of human skeletal muscle aging

Study	n	Baseline Age	Follow-Up Duration, yr	Measurement	Method	%Δ	%Δ/yr
Paraspinals
Murata et al. 2021 (100)	276 M	60–69	10	CSAa	CT	−8	−0.76
	31 W					−10	−1.01
Total: 1 study*	307	65	10				−0.79
Psoas
Murata et al. 2021 (100)	276 M	60–69	10	CSAb	CT	−9	−0.88
	31 W					−4	−0.40
Total: 1 study*	307	65	10				−0.83
Hamstrings
Frontera et al. 2000 (3)	7 M	65 ± 4	12	CSA	CT	−15	−1.22
Total: 1 study	7	65	12				−1.22
Quadriceps
Frontera et al. 2000 (3)	7 M	65 ± 4	12	CSA	CT	−16	−1.32
Total: 1 study	7	65	12				−1.32

Age (yr) is presented as means ± SD, or range. CSA, cross-sectional area; CT, computed tomography; M, men; W, women.

*Data are total n size or average values normalized to the n size of each sex; ^aerector spinae and multifidus; ^bpsoas major.

MUSCLE-SPECIFIC ATROPHY WITH AGING

The main findings of this literature review are 1) nonquadriceps data for human skeletal muscle aging are limited and 2) with the available information, a diverse range of muscle atrophy rates with aging is apparent. Of the nine muscle groups available in the literature and presented here, there was more than a fivefold difference between the least (−6%, soleus) and the most (−33%, rectus femoris) atrophying muscles over the 50 yr of aging (Fig. 1). With the large number of muscles that have not been investigated, it is unknown if there are other specific muscles outside this large range. Given that nearly half of the available literature in this area is from the quadriceps, the large range of muscle-specific atrophy, and the large number of muscles that have not been investigated, it is clear that the area of aging muscle-specific atrophy needs to be expanded.

Thigh

The quadriceps, hamstrings, and hip adductors were included in this review from the thigh region. These three muscle groups cover all functional groups present in the thigh (25, 45) that are involved in different phases of various movements performed on a daily basis such as standing, walking, and stair climbing (101–107), and vigorous movements during exercise such as running, cycling, kicking, jumping, and change of direction tasks (108–114). However, the hamstrings and hip adductors represent less than 30% and 20%, respectively, of the amount of data on the quadriceps in the literature. In addition to the well-documented association between lower quadriceps strength and increased risk of falls in older men and women (115, 116), aging appears to increase the dependence on the hamstring muscles for postural stability and a greater coactivation of hamstrings and quadriceps has been associated with less occurrence of falls (117, 118). Considering the falls that lead to hip fracture occur more commonly in the lateral direction (119, 120), fall incidence is likely related to the age-related decline in hip adductor function as well (121).

Among these thigh muscles, the quadriceps appears to have the greatest rate of atrophy with a hierarchical pattern (Fig. 1): quadriceps (−27%) > hamstrings (−19%) > hip adductors (−13%). The quadriceps atrophy is also different from the loss of total skeletal muscle mass with similar age gaps reported by Janssen et al. (4) with MRI measurement (−11%) and Kyle et al. (5) with dual-energy X-ray absorptiometry measurement (−17%). This discrepancy further confirms that examination of the quadriceps does not reflect the aging process of all muscle groups. The available data on subcomponent muscles provide novel insights into muscle-specific atrophy within the same muscle groups (Table 3). In the quadriceps, the rectus femoris appears to have a greater rate of muscle atrophy than the vasti muscles. In the hamstrings, the medial muscles (semimembranosus and semitendinosus) seem more sensitive to the aging process than the lateral side of the muscle group (biceps femoris). Considering the different contribution of subcomponent muscles during various activities (106, 122, 123), understanding the relative changes in the subcomponent muscles of a muscle group during the aging process need further investigation (6).

Muscle-specific activity patterns likely influence muscle-specific aging atrophy, even in sedentary individuals, given the diverse functions across muscle groups. Electromyography (EMG) studies have shown marked differences in the EMG activity pattern, peak, and duration during simple daily movements such as normal gait, stair ambulation, and sit-to-stand movement across the thigh muscles (101–104, 106, 107). Tracking the EMG activities of various muscle groups for extended periods would provide useful information about the normal activation levels of each muscle group (124). In addition, findings from extreme inactivity provide insight into daily activity levels in a muscle-specific manner. Two months of bed rest results in atrophy of ∼17% in the quadriceps, ∼12% in the hamstrings, and ∼7% in the adductors (28, 125, 126). Thus, more than half of the quadriceps atrophy observed with 50 yr of aging occurs in only 2 mo of bed rest (27% vs. 17%), which suggests daily muscle activities significantly slow down the inactivity-induced muscle atrophy. In addition, only a relatively small amount of exercise has potent effects on age- and inactivity-related muscle atrophy (17, 28, 127–133). For example, only a few minutes a week of muscle contraction with resistance exercise has been shown to preserve muscle size in old individuals (132) or during prolonged bed rest (134, 135). Given the muscle-specific responses to exercise (2, 26–28, 134), these exercise training strategies for age-related muscle atrophy warrant investigations of more muscle groups.

Lower Leg

The triceps surae and dorsiflexors were included in this review from the lower leg region. The available information on these two muscle groups (∼20%–30% of the quadriceps data) covers the major muscles that are involved in the key functions of the lower leg for postural stability and ambulation (69, 136–141) and important contributors for running performance (142, 143). However, muscles that reside in the deep (i.e., flexor digitorum longus, flexor hallucis longus, and tibialis posterior) and the lateral lower leg (i.e., peroneus brevis and longus) have no available data, although these muscles also provide vital roles for gait and ankle stability (144, 145). Dysfunction associated with atrophy of the lower leg muscles impairs walking ability, reduces walking speed, and increases the risk of falls (65, 146–148). In addition, due to the complex network of vascular anatomy in this region, contraction of the calf muscles contributes substantially to venous return during exercise (149, 150) and it has been shown that function of the calf muscles is associated with cardiovascular health (146, 151–154).

From the available data in the literature on the lower leg muscles, there seems to be a trend toward a higher rate of muscle atrophy in the gastrocnemius than the other lower leg muscles (Fig. 1). This is supported by greater activities in the soleus and tibialis anterior than gastrocnemius during walking (106). Moreover, it has been suggested that the soleus is more active than the tibialis anterior during normal daily activities (155), which is supported by the findings of greater muscle atrophy in the soleus during bed rest (126, 156–159).

The lower leg muscles appear to be less susceptible to age-related atrophy than the thigh muscles (Fig. 1). It is well accepted that the soleus is an “antigravity muscle” that provides a sustained force for postural stability (160). Because of this, when the muscle is unloaded from gravity with space flight or microgravity simulation, it experiences approximately twice as much atrophy as the quadriceps (130, 161, 162) and this relationship is essentially reversed during the aging process (Fig. 1). In addition, it has been shown that responses to exercise training and nutrition are different between these muscle groups at the whole muscle (2, 28, 130), myocellular (27, 163–165), and molecular levels (94, 166). Thus, these muscle groups likely require different exercise programs or other intervention strategies against the aging process.

Trunk

The paraspinals and psoas were included in this review from the trunk region, both of which only contain three studies each. The paraspinals are responsible for maintaining upright position and stabilizing the spine (167) and support various daily tasks such as standing and picking up an object from the floor (168). The psoas also provides stability of the lumbar spine and femoral head (169–172) and a greater psoas size has been associated with better sprint running performance (109, 173). Although these two muscle groups make up most of the deep trunk muscle mass around the spine, data on muscles connecting the trunk and arms are lacking. For example, the latissimus dorsi and pectoralis muscles provide coordinated movement of the trunk and arms (i.e., pulling and pushing) such as opening doors, moving a shopping cart, and various strokes during swimming (174, 175). One of the most common age-related health concerns in the trunk region is low back pain, which has been reported to occur in more than half of older adults in the United States (176–179). Atrophy of the paraspinal muscles has been shown to accompany chronic low back pain (180–182). In addition, the psoas is commonly used to assess the prognosis of patients in numerous clinical settings due to its availability in various diagnostic imaging (183–185).

Somewhat greater muscle atrophy in the psoas (−29%) than in the paraspinals (−24%) follows the suggestion that muscles with more sustained work atrophy less. During unloading with spaceflight, the paraspinals experience approximately twice as much atrophy as the psoas (161). This indicates that the paraspinals are more chronically activated for spine stabilization during daily life. It has been shown that atrophy of the paraspinals may not occur until the seventh or eighth decade of life (100, 186–188), whereas the psoas has been shown to atrophy as early as the fifth decade (42, 100). This suggests that interventions to age-related muscle mass decline likely need to be muscle-specific and may need to be initiated earlier in the psoas.

Arm

The elbow extensors and flexors were included in this review from the arm region, but the data are composed of the lowest numbers of subjects among the nine muscle groups in this review (∼10% of the quadriceps data each). These two muscle groups make up all the upper arm muscles that are associated with elbow function (59) and they play major roles in cleaning, showering, and lifting heavy objects (189, 190). As the upper extremity muscles allow for performing numerous daily movements, it is not surprising that upper extremity function has been shown to be positively associated with quality of life (191, 192).

The available data suggest that age-related muscle atrophy does not seem to be different between elbow extensors (−19%) and elbow flexors (−20%). What the relative activity patterns of these muscles are over the lifespan is unclear, but the similar levels of atrophy suggest a similar decline. Other literature not included in this review provides some insight into the relative atrophy of the arm muscles. In younger aging men (68 yr) and older aging women (87 yr), measurements of quadriceps and elbow flexor muscle size in the same individuals suggest that the quadriceps atrophies more than the elbow flexors (21, 33). This notion of less atrophy in the elbow flexors is supported by observations of nearly twice as much daily muscle activity duration in the elbow flexors than quadriceps (193), and the relative activity level reduction during aging may also be smaller in the upper arm muscle groups. In addition, muscle adaptations to resistance exercise training have also shown to be different between the quadriceps and upper arm muscle groups (132, 194–196), which may be related to the motor unit number and activation in these muscle groups (197–201).

Muscle Fiber Type

In addition to muscle activity levels, muscle fiber type may also play a role in age-related muscle atrophy (22). It is generally assumed that aging results in a higher distribution of type I skeletal muscle fibers (1, 95). In support of this, four studies on the quadriceps included in this review also reported the muscle fiber type distribution of young and old individuals (from the vastus lateralis) and all showed an increased distribution of type I muscle fibers in old men and women along with whole muscle atrophy (13, 17, 46, 49). In addition, it has been suggested that type II muscle fibers preferentially atrophy with aging (7, 49, 95). Because of these observations, sarcopenia is generally characterized by a loss of type II muscle fiber size and number (95, 96, 202). Thus, muscles with more type II fibers might be more susceptible to aging atrophy. This notion is supported by the relationship between the percentage of type II muscle fibers reported in the literature for all the muscles presented in this review compared with the percentage of age-related skeletal muscle atrophy compiled for this review (Table 4 and Fig. 2).

Figure 2.

Relationship between the percentage of type II muscle fibers and muscle atrophy over ∼50 yr of aging (r = −0.73). Muscle fiber types (%type II) are mean data from Table 4. Muscle atrophy rates (%Δ) are from Fig. 1. For hamstrings, quadriceps, and triceps surae, data for subcomponent muscles are also included as similar, but smaller symbols of the main muscle group. Add, hip adductors; DF, dorsiflexors; EE, elbow extensors; EF, elbow flexors; Ham, hamstrings; Para, paraspinals; Pso, psoas; Quad, quadriceps; TS, triceps surae.

This overall relationship seems to hold within the subcomponent muscles of a muscle group, but not across all muscle groups within a region. The four individual muscles of the quadriceps (rectus femoris, vastus lateralis, vastus intermedius, and vastus medialis), the four muscles of the hamstrings (biceps femoris short and long heads, semimembranosus, and semitendinosus), and the gastrocnemius and soleus of the triceps surae all seem to support this relationship (Fig. 2). However, this relationship does not seem to hold for the elbow extensors and flexors, which had similar rates of atrophy but somewhat different fiber type distribution. Interestingly, unlike the vastus lateralis, alterations in muscle fiber type distribution with age have not been reported in the biceps brachii with cross-sectional (203) and longitudinal (204) examinations. It is also noteworthy that those muscles that are most chronically active with typical activities of daily living [e.g., the soleus and tibialis anterior (106)] have a greater proportion of type I muscle fibers and a lower rate of age-related muscle atrophy. Overall, these data suggest that aging is a muscle-specific process even at the cellular level, and activity and fiber type are interconnected in their influence on muscle-specific atrophy with aging.

Sex Comparisons

Potential sex-related differences in the decline in muscle mass and strength with aging is an important consideration (2, 205, 206). From the currently available data (Table 5), age-related atrophy in the psoas, hamstrings, and quadriceps does not seem to be largely different between men and women. However, the hip adductors, dorsiflexors, and triceps surae do seem to atrophy more in women than men. It is important to note that, except for the quadriceps, these interpretations are based on relatively small numbers of studies and total sample size. Sex differences in skeletal muscle have been suggested to be related to sex hormones, particularly testosterone (207–209), but the muscle-specific influence of sex hormones remains to be determined (210–212). In addition, influence of age on muscle composition (i.e., lipid content) and strength of the lower leg muscle groups appears to be similar between men and women (2, 67, 213, 214). Thus, a plausible explanation for these muscle-specific sex differences is not readily apparent and requires further investigation.

Age-Related Muscle Atrophy Later in the Lifespan

The available data from longitudinal investigations of old individuals allow us to compare age-associated muscle atrophy that occurs from ∼65 yr to 75 yr in the paraspinals, psoas, hamstrings, and quadriceps (Table 6). Similar to the cross-sectional studies that examined 50 yr of aging from ∼25 yr to 75 yr, the atrophy rates are widely different across muscle groups. In addition, the greater rates of muscle atrophy in the last decade of the 50 yr age span studied in the cross-sectional examinations suggest a nonlinear decline in the size of these muscles (25–75 yr vs. 65–75 yr; paraspinals: −0.47%/yr vs. −0.79%/yr, psoas: −0.58%/yr vs. −0.83%/yr, hamstrings: −0.39%/yr vs. −1.22%/yr, quadriceps: −0.53%/yr vs. −1.32%/yr). This has also been shown in cross-sectional studies of the quadriceps that compared middle-aged to old (42, 215, 216) and old to older individuals (42, 216–218). This nonlinearity of atrophy is likely muscle specific as well. In addition, it is noteworthy that there appears to be some sex differences in the muscle atrophy rates in the trunk muscles (Table 6) that are not seen in the young to old comparison (Table 5). Given the possible differences in the onset of age-related atrophy in these muscle groups (42, 100, 186, 188, 216, 219–221), additional longitudinal investigations of older women and men (>75 yr) are warranted.

Future Directions

The existing data in the literature clearly show diverse age-related atrophy rates across muscles and support the need for future studies to expand the knowledge of muscle-specific aging. These studies should consider the time course of the aging process (i.e., the shape of the atrophy curves), exercise and other interventions, and sex while expanding on the existing muscle-specific information and investigating other important muscle groups. It is logical to pursue muscles that are associated with essential daily tasks. These include but are not limited to walking, running, cycling, stair ascending and descending, standing up from a chair, getting up from bed, bathing, dressing, carrying groceries, opening jars, eating, toileting, and lifting things from the floor. In addition, a better understanding of the influence and muscle-specific activity levels of different lifestyles (e.g., sedentary, walker, runner, cyclist, resistance exerciser) and occupations (e.g., office workers, farmers, factory workers) will aid in the future development of exercise countermeasures and other interventions to mitigate the aging process in a muscle-specific manner.

Although adaptation to exercise training is not the focus of this review, there is some evidence to suggest that the effects of aging on whole muscle size changes with an exercise intervention are also muscle specific. Available data have shown that the magnitude of hypertrophy in response to similar exercise training is different in elbow extensors, elbow flexors, hamstrings, quadriceps, and triceps surae (2, 194–196, 222–225). In addition, a few studies have shown that the adaptations to the same exercise intervention in these muscle groups differ between young and old individuals (226, 227).

Current technology for whole muscle imaging allows hundreds of human skeletal muscles to be quantified with proper reliability. Even cost-effective noninvasive imaging with ultrasound allows the examination of over 100 human skeletal muscles (34). Thus, future investigations should not be limited by the noninvasive whole muscle imaging techniques. Clearly, the nonquadriceps muscle groups included in this review require more data to draw definitive conclusions regarding the aging process. In addition, a few example regions of muscles that could be targeted for aging and intervention studies are the neck, shoulder, abdominal, gluteus, forearm, and hand. Neck and shoulder muscles provide complex functions and the musculoskeletal symptoms in this region are commonly attributed to sedentary lifestyle (228). The abdominal muscles provide trunk support and are composed of the rectus abdominis and lateral abdominals, which appear to be most activated for postural stability and balance among all the trunk muscles (229). The gluteal muscles are one of the largest muscle groups in the human body (230–232) and provide hip and pelvic stability and play essential roles in daily movements such as walking, stair climbing, and sit-to-stand motion (102, 103, 106, 107, 233, 234). A change in gluteus muscle composition has been associated with an increased risk of falls (235, 236). Forearm and hand muscles also provide essential functions in daily tasks. Hand grip strength has been used in numerous investigations and associations with various health outcomes have been shown (237–240). However, investigations into muscle-specific age-related atrophy in these muscle groups are limited (241–245). Given the variety of functions that these muscles provide, it is reasonable to hypothesize that there would be numerous muscle-specific changes with aging, which may require muscle-specific considerations for exercise and other interventions.

CONCLUSIONS

This review provides a comprehensive summary of the existing literature on age-related skeletal muscle atrophy in a muscle-specific manner. The large range of atrophy across muscle groups from the available data suggests that a single muscle group or nonmuscle-specific lean mass measurement does not likely provide the necessary information to comprehensively understand human skeletal muscle aging across the entire body. It is acknowledged that the quadriceps is functionally important and studies of this muscle have greatly expanded our understanding (and will continue to) at the whole muscle, myocellular, and molecular level. The availability of current noninvasive imaging technologies that are capable of evaluating some element of muscle composition in addition to muscle size would allow for a broader examination of skeletal muscle health assessment during the aging process. Our understanding of age-related atrophy has expanded substantially over the last several decades and will continue to grow with a more broad-ranging muscle-specific approach to skeletal muscle aging.

GRANTS

This work and our skeletal muscle imaging research have been supported by grants from the NIH under Grants AG020532, AG015833, AG000831, AG038576, AG018409, and AG015486 and National Aeronautics and Space Administration (NASA) under Grants NNJ06HF59G, EC400-NCC9-116, and NNJ04HF72G.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

M.N. and T.A.T. conceived and designed research; M.N. and T.A.T. prepared figures; M.N. and T.A.T. drafted manuscript; M.N., S.T., and T.A.T. edited and revised manuscript; M.N., S.T., and T.A.T. approved final version of manuscript.