IAGS
39:555-561, 1991
The Relationship between Muscle Mass and Muscle Strength in the Elderly Richard L. Reed, MD, MPH,*t Lori Pearlmutter, RPT,§ Kathleen Yochum, RN,GNP,* Keith E. Meredith, PhD,* and Arshag D. Mooradian, MD*$
To determine the extent that muscle mass is predictive of muscle strength in the elderly, anthropomorphic estimates of muscle area and impedance measurements of muscle mass and peak isometric muscle strength were obtained in a relatively healthy older population over 65 years of age (mean age = 71.7; n = 218). Midarm muscle area correlated strongly with upper arm strength (r = 0.68, P < 0.0001) while midthigh muscle area had a much lower correlation with thigh muscle strength (r = 0.29, P < 0.0001). These muscle area calculations also include bone area. Lean body mass calculated by bioelectric impedance correlated highly with cumulative muscle strength measured by summing all muscle groups (r =
< 0.0001). To determine whether aging alters muscle strength per unit of muscle mass, additional middle-aged subjects were included, and three groups, middle-aged (55-64) (n = 78, young-oId (65-74) (n = 161), and old-old (75+) (n = 57), were compared. A significant age-related trend of decreasing muscle strength per unit of lean body mass was noted. It is concfuded that although muscle mass correlates with muscle strength in a healthy older population, use of simple age-independent clinical measurements of body mass should not be used to predict muscle strength. J Am Geriatr SOC39:555561, 1991
adequate muscle strength is a basic requirement for physical functioning and has been shown to decrease with aging in multiple cross-sectional and longitudinal studies.’-’ This age-related loss of muscle strength puts the elderly at high risk for functional disability. The cause of this loss of strength is unknown and may be due to several factors. One possibility is that the loss of muscle strength is attributed totally to the loss of muscle mass associated with aging. Another possibility is that there is a loss of strength per unit of muscle mass (efficiency) associated with aging. Other possibilities include the influence of chronic diseases, decreasing physical activity, inadequate nutrition, other pathological changes in the muscle associated with aging, and decreased motivation in testing among the elderly. Previous studies addressing the relationship between
muscle mass and muscle strength have primarily been 5, These studies have done using grip ~trength.~, shown a variable degree of positive correlations between muscle strength and an index of muscle mass with a range of r values from 0.17 to 0.90. A recent study noted that in addition to an age-associated loss of muscle mass, there is an additional loss of handgrip strength per unit of body mass in males.’ Studies in other muscle groups have shown stronger correlation but have been limited to elbow extensor^,^ plantar flexor^,^ and knee extensors.6 In addition, they have not always used regional body mass measurements for correlation with the muscle group tested.6 The purpose of this study was to determine the extent that muscle mass is predictive of muscle strength in different large muscle groups in the healthy elderly and to determine whether aging is associated with altered muscle “efficiency”(muscle strength per unit muscle mass) in these muscle groups.
A
From the *Arizona Center on Aging and the Departments of tFamily Medicine and $Internal Medicine, University of Arizona College of Medicine, Tucson, Arizona; and SDepartment of Physical Therapy, University Medical Center, Tucson, Arizona. This research was supported in part by The Flinn Foundation and by the Hudson Foundation. Address correspondence to Richard L. Reed, University of Arizona, 1821 East Elm Street, Tucson, AZ 85719.
01991 by the American Geriatrics Society
0.79, P
METHODS Subjects were recruited at a health fair for a retirement community of approximately 17,000 people in southeastern Arizona. Individuals were divided into three age groups; middle-aged (55-64), young-old (650002-861 4/91 /$3.50
556
REED ET AL
74), and old-old (75 and over). In this study, elbow and knee flexor and extensor isometric muscle strength, as well as ankle dorsiflexor strength, was measured to the point of breaking with a hand-held dynamometer (Penny & Giles Transducers Limited, Christ Church, Dorset, U.K.) using the testing technique of Kendall and McCreary.’ Elbow flexor and extensor strength was measured in the sitting position with the forearm in supination with stabilization of the upper arm. Elbow flexor strength was tested with the elbow flexed at 90° and the arm beside the trunk with the dynamometer being applied to the anterior surface of the forearm just proximal to the wrist joint. Elbow extensor strength was measured with the elbow at 90° with the upper arm held perpendicular to the floor and the forearm in supination. The dynamometer was applied to the posterior surface of the forearm just proximal to the wrist joint. Knee flexors (biceps femoris, semitendinous, and semimembranous muscles) were measured in the prone position with the knee at 30° with application of the dynamometer at the ankle. The thigh was manually stabilized. Knee extensors (quadriceps femoris) were measured in the sitting position with the knee and hip flexed to 90°. The dynamometer was applied at the ankle with stabilization at the thigh. Ankle dorsiflexors (peroneus tertius, tibialis anterior) were also measured in the sitting position with the ankle in neutral position (perpendicular to lower leg). The lower limb was stabilized proximal to the ankle. The dynamometer was applied to the plantar surface of the foot just proximal to the metatarsalphalangeal joints. In this study the test-retest correlation for the muscle groups ranged from .92-.98 with the exception of right and left knee flexors (.78-20). There was no significant difference in the strength of each muscle group when test and retest values were compared ( n = 38). The whole study group was tested by only two of us (RLR and LP) with one person (RLR) testing the majority of patients (79%). The interrater reliability ranged in different muscle groups from .78-.90. Arm muscle strength was calculated by summing the forces for left elbow extensors and elbow flexors. Thigh muscle strength was calculated by the addition of left knee flexor and left knee extensor strength. Leg muscle strength was calculated by adding left dorsiflexor strength, left knee flexor, and left knee extensor strength. Cumulative muscle strength was calculated by summation of both the right and the left muscle strength for all 10 muscle groups tested. Anthropomorphic measurements were made on the left arm (midarm circumference, triceps skin fold) and left thigh (midthigh circumference and thigh skin fold). Skin fold thicknesses at these sites were measured with a Lange caliper (Cambridge Instrument Co., Cambridge, MD). Midarm muscle area (MAMA) was cal-
IAGS-]UNE 1991-VOL. 39, NO. 6
culated from triceps skin fold (TSF) and midarm circumference (MAC) measurements using the equation lo MAMA = (MAC in cM - (T X TSF in mM/10))2/4~.9, Arm muscle and bone circumference (AMC) were calculating using the equation AMC = MAC in cM - (T X TSF in cM). Midthigh muscle area (MTMA) and thigh muscle circumference (TMC) were calculated with similar formulas using midthigh circumference and midthigh skinfold thickness. It should be noted that all of the calculations are not corrected for bone area or circumference. Hence, the labels given are somewhat of a misnomer although they follow the previously established convention. Height (Ht) and weight (Wt) were measured using an upright beam scale. The body mass index (BMI) was calculated from these values as Wt/Ht2 (Kg/M2). Lean body mass (LBM) was calculated as the difference between body weight and fat mass as determined by a bioelectric impedance analyzer (Valhalla Co., Model 1990B, San Diego, CA). The analyzer was calibrated prior to each testing session with an internal 499-ohm resister supplied with the unit. Subjects were measured in the supine position, lying on a non-conductive elevated surface. The arms were resting at the sides and the legs were spread apart to prevent contact. One detection electrode was placed between the distal condyles of the ulna and radius on the posterior surface of the right wrist. The other detection electrode was attached between the most prominent portions of the malleoli of the fibula and tibia on the anterior surface of the ankle. The two-source electrodes were placed at the base of the third metacarpal-phalangeal joint and 1-centimeter proximal to the third metatarsal-phalangeal joint. An excitation current was introduced at a frequency of 50 kHz and the resistance was measured. Ambient temperature was relatively constant for all subjects during this study. Lean body mass was calculated with a formula specifically derived for the Valhalla instrument in comparison with underwater weighing.” The equations with height (Ht) expressed in centimeters and resistance (R) in ohms were: LBM = 0.475 X Ht2/R + 5.32 for women and LBM = 0.827 X Ht2/R + 5.49 for men. Statistical analysis was performed using the Statistical Package for the Social Sciences.” A two-way, fixed effects analysis of variance with age group and sex as independent variables was calculated on each of the strength and mass variables. The Pearson productmoment correlation coefficient was used for correlating mass and strength using two-tailed probabilities. Slope was calculated using ordinary least squares regression analysis. One-way fixed effects analysis of variance was done separately for males and females, with the tertile of lean body mass calculated for each sex separately as the independent variable and cumulative muscle strength per kilogram of lean body mass (effi-
557
MUSCLE MASS AND MUSCLE STRENGTH
JAGS-]LINE 1991-VOL. 39, N O . 6
the measurements in males decreased with increasing age while little or no change was noted in females. In individuals over the age of 65, MAMA strongly correlated with arm muscle strength measured as the sum of the force in kilograms produced by elbow flexors and elbow extensors (r = 0.68, P < 0.0001) RESULTS (Figure 1). AMC similarly correlated with arm muscle There were 296 subjects recruited from a total of 326 strength (r = 0.68, P < 0.0001). When divided into individuals participating in the health fair. Excluded males and females the level of correlation between were 30 individuals who refused muscle or body mass MAMA and arm muscle strength was lower (males = testing or were unable to perform all the muscle tests .36, P < 0.0001; females = .23, P < 0.05). A similar due to neuromuscular or joint problems. All the sub- pattern was found for the correlation of AMC with jects were ambulatory and active Caucasians in middle arm muscle strength (r = 0.36, P < 0.0001 for males; Y or upper middle socioeconomic class. Although they = 0.23, P C 0.05 for females). In those 65 and older, MTMA correlated less strongly were not disease free, review of their medical history and medications used suggested that they were a rel- with thigh muscle strength (r = 0.29, P < 0.0001) than atively healthy group. Individuals who were 65 and did MAMA with arm strength (Figure 2). This correlaolder (n = 218; mean age (fSE) = 71.7 f 0.4 years) tion was stronger in males (r = 0.37, P < 0.0001) than were the primary study group. However, for compari- females (r = 0.19, P < 0.05). TMC had a similar level son purposes, measurements were made on an addi- of correlation with thigh muscle strength ( Y = 0.29, P tional 78 middle-aged subjects (mean age (+SE) = 53.1 5 0.0001) that was also stronger in males (r = 0.37, P < 0.0001) than females ( r = 0.19, P C 0.05). & 1.5 years). The correlation of arm muscle strength with leg Muscle strength declined with age in all muscle strength was also strong (r = 0.80, P < 0.0001) and groups (Table 1).As expected, all muscle groups tested had a similar magnitude in both women (r = 0.62, P < were also weaker in women than in men. In addition, many of the anthropomorphic and body composition 0.0001) and men (r = 0.58, P < 0.0001) (Figure 3). values usually declined with age (Table 2). However, Lean body mass calculated by impedance bioelectric AMC, MAMA, thigh skinfold, TMC, percent body fat measurements correlated strongly with cumulative and body mass index were not significantly different muscle strength (r = 0.79, P < 0.0001) (Figure 4). There in the three age groups studied. All anthropomorphic was a weaker correlation when separated by gender (Y and body composition measures differed significantly = 0.41, P 5 0.0001 for males and r = 0.32, P 5 0.0001 by gender with the exception of thigh muscle circum- for females). When the ratio of muscle strength divided ference and thigh muscle area. A statistically significant by lean body mass was analyzed for the elderly (65 interaction between age and sex was found with only and over) by tertiles of lean body mass, a significant two variables, AMC and MAMA. In both situations, difference was noted using one-way analysis of vari-
ciency) as the dependent variable. Two-way fixed effects analysis of variance was also performed with age group and sex as the independent variables and cumulative muscle strength per kilogram of lean body mass (efficiency) as the dependent variable.
TABLE 1. THE MEAN (+SE) OF MUSCLE STRENGTH (Kn)BY AGE GROUP AND GENDER* Middle-Aged Y oung-Old Old-Old Variable
Males n = 33
Females n = 45
Males n = 74
Females n = 87
Upper Extremity 16.7 f 0.4 R elbow flexor 29.7 f 0.6 19.2 f 0.6 26.7 f 0.5 16.0 f 0.4 L elbow flexor 28.5 f 0.6 18.3 f 0.5 26.0 f 0.5 R elbow extensor 12.8 f 0.3 22.6 f 0.7 14.3 f 0.4 19.5 f 0.4 12.4 f 0.3 L elbow extensor 21.8 f 0.7 14.3 f 0.4 18.5 f 0.4 28.4 k 0.6 Combined L arm 50.2 f 0.8 32.7 f 0.8 44.5 f 0.9 Lower Extremity R knee flexor 13.8 f 0.5 26.2 f 1.0 18.0 f 0.8 22.0 f 0.5 L knee flexor 13.8 f 0.4 25.8 f 0.9 17.7 f 0.8 22.2 k 0.5 21.3 f 0.6 R knee extensor 30.0 f 0.6 23.9 f 0.8 27.8 f 0.5 L knee extensor 21.4 f 0.6 30.4 f 0.7 24.0 f 0.8 28.4 f 0.5 Combined L thigh 56.2 f 1.5 41.8 f 1.2 50.5 f 0.8 34.2 f 0.9 R dorsiflexor 21.5 f 0.6 30.2 f 0.6 22.0 f 0.8 28.1 f 0.5 L dorsiflexor 20.8 f 0.6 29.4 f 0.6 22.3 f 0.8 27.9 f 0.5 Combined L leg 48.4 f 1.2 80.9 f 2.0 57.7 f 1.7 72.3 f 1.4 Cumulative Muscle Strength 299.9 f 0.5 211.7 f 4.9 269.9 f 3.8 183.4 f 3.6 * A f f measurements were Significantly different (P < 0.05)by age and sex using analysis of variance.
Males n = 26
Females n = 31
24.3 f 0.7 24.2 f 0.8 18.3 f 0.7 18.0 f 0.7 42.2 k 1.3
15.5 f 0.5 15.1 f 0.4 11.6 f 0.4 11.1 f 0.4 26.2 f 0.7
18.7 f 1.2 18.8 f 1.2 25.5 f 1.2 25.4 f 1.1 44.3 f 2.0 26.5 f 0.8 25.9 f 0.9 64.5 f 2.8 244.4 f 7.5
12.6 f 0.6 12.3 f 0.5 19.7 f 0.8 19.5 f 0.8 31.8 f 1.2 18.5 f 0.8 17.8 f 0.8 42.4 f 1.7 166.1 f 4.6 ~~~
~~
558
REED ET AL
IAGS-JUNE 1991-VOL. 39, NO. 6
TABLE 2. THE MEAN (fSE) OF ANTHROPOMORPHIC MEASUREMENTS AND BODY COMPOSITION OF THE STUDY SUBJECTS Middle- Aged Young-Old Old-Old Males n = 33
Females n = 45
Males n = 74
Females n = 87
Males n = 26
32.6 f 0.5
28.3 f 0.2
31.2 f 0.3
28.9 f 0.3
29.8 f 0.5
27.6 f 0.5
18.5 f 1.1 26.8 f 0.4
24.0 f 0.8 20.8 f 0.3
18.3 f 0.8 25.5 k 0.4
24.0 f 0.7 21.4 k 0.2
16.9 f 1.1 24.5 f 0.4
21.2 f 1.1 21.0 f 0.3
57.8 f 1.9
34.9 f 1.0
52.6 f 1.3
36.9 f 0.8
48.3 f 1.6
35.3 k 1.0
55.3 f 1.6 21.8 f 1.9 47.5 f 1.8
55.9 f 0.8 30.5 f 1.0 46.3 f 0.8
52.7 f 0.7 18.8 f 0.9 46.9 f 0.8
56.1 f 0.6 29.8 f 0.9 46.7 f 0.5
51.5 f 0.7 19.7 f 1.1 45.3 f 0.7
53.9 f 1.0 27.1 f 1.4 45.3 f 0.7
187.5 f 7.9
174.2 f 0.5
179.3 f 4.3
172.3 f 3.7
164.3 f 4.8
164.9 f 5.2
22.0 f .01 28.0 f .01 65.3 f 8.5 45.0 f 0.7 1.78 f .02 1.64 f .01 84.5 f 2.6 63.5 f 1.4 26.7 f 0.6 24.2 f 0.8 * Statistically significant difference by age (P< 0.05) ** Statistically significant difference by gender (P < 0.05) Statistically significant interaction of age and sex f P < 0.05)
22.0 f .01 62.1 f 0.6 1.77 f .01 80.0 f 1.0 25.6 f 0.3
30.0 f .01 44.4 f 0.5 1.63 f .01 63.8 f 1.1 24.2 f 0.4
21.0 f .01 59.9 f 1.4 1.75 f .01 75.8 f 2.0 24.9 f 0.6
29.0 f .01 42.9 f 0.8 1.59 f .01 61.0 f 1.6 24.2 & 0.6
Upper Extremity Arm circumference (cm)*,** Triceps skin fold (mm)*,** Arm muscle circumference (cm) (AMC)** Arm muscle area (cm') (MAMA)*, ** Lower Extremity Thigh circumference*, ** Thigh skin fold*, ** Thigh muscle circumference (cm) (TMC) Thigh muscle area (cm') (MTMA)* Total Body Percent fat by impedance** Lean body mass (Kg)*, ** Height (meters)*,** Weight (Kg)*, ** Body mass index**
601
Females n = 31
8olo . ...-. - .
/
7o
50
Male Female
60 0
t
45
S? 40 40 30 -
-
I
I
I
30
40
50
b
20
20 Male
.
I
I
1
60 70 80
I 90
Midarm Muscle Area (Cm2)
10 100
150
200
250
300
Midthigh Muscle Area (Cm2)
FIGURE 1. The relationship between midarm muscle area and upper arm strength in subjects over 65 years of age. The Pearson product moment correlation is .68, P < 0.0001. The linear regression equation (fSE) is y = 0.59 (.04) X +9.51 (1.92). Midarm muscle area also includes bone area.
FIGURE 2. The relationship between midthigh muscle area and thigh strength in subjects over 65 years of age. The Pearson product moment correlation is .29, P < 0.0001. The linear regression equation (fSE) is y = 0.10 (.02) X +24.62 (3.63). Midthigh muscle area also includes bone area.
ance for men and women separately (Figure 5). Individuals who have lower lean body mass are smaller and have less muscle mass. When this ratio of cumulative strength divided by lean body mass (efficiency) was calculated and analyzed for all three age groups using two-way analysis of variance (Figure 6) a statistically significant age effect was noted (P < 0.0001) with a decline with increasing age.
DISCUSSION This study demonstrates that there is a strong correlation between lean body mass and cumulative muscle strength in a relatively healthy community-dwelling elderly population (Figure 4). Regional muscle mass estimates and muscle strength correlations were also strong for the arm (r = 0.68, P 5 0.0001) but had a
MUSCLE MASS AND MUSCLE STRENGTH
IAGS-]LINE 1991-VOL. 39, NO. 6
60 8
5.0
Female Male
0 Male
50
y"
%
0
v
5
Female
4.5
al .0 .c
40
cn C
559
4.0
2
-al
5
0 u)
30
E, a
ga
r'
:. 0.y.
20
2
'
34
0.0
/
10
10
3.5
Lowest Third
1
30
20
40
50
60
70
Leg Strength (Kg) FIGURE 3. The relationship between arm strength and leg strength in subjects over 65 years of age. The Pearson product moment correlation is 30, P < 0.0001. The linear regression equation (2SE) is y = 0.47 (.02) X +8.08 (1.44).
Highest Third
FIGURE 5. Comparison of the means (+SE) of the cumulative muscle strength per kilogram of lean body mass (efficiency) by tertile of lean body mass in elderly males and females. P-value < 0.05 by one-way analysis of variance. Tertiles created for each sex separately.
5.0
350 1 % h
0
y"
al ._
-
v
Middle Third
4.5
r
nMale Female Both Sexes
.-0
5 U J C
4.0
0) -
??
0 m
m
r'
200 150
3.5
0.0 __
8
50
30
40
50
60
70
Female Male 80
90
Lean Body Mass (Kg) FIGURE 4. The relationship between lean body mass and cumulative muscle strength in subjects over 65 years of age. The Pearson product moment correlation is .79,P < 0.0001. The linear regression equation (+SE) is y = 4.16 (.22) + 1.70 (11.73).
Middle-aged Young-old Old-old FIGURE 6. Comparison of the means (+SE) of cumulative muscle strength per kilogram of lean body mass (efficiency) by age group in males and females. Overall P < 0.0001) by two-way analysis of variance. Significant differences were noted by age (P < 0.001) but not sex. No sigruficant interactions were noted.
tion in the thigh is that maximal recruitment of muscles, especially in this region, may not have been achieved. We feel that all of the subjects were highly motivated and gave maximal voluntary efforts. It is unknown, weaker correlation for the thigh ( T = 0.29, P 5 0.0001). however, whether an age-related difference in recruitIn another study of subjects over the age of 65, anthro- ment could explain the lower correlation in this study. pomorphic measurements of arm and calf muscle area Alternatively, the measurement of knee flexor and were compared to biceps brachii (elbow flexor) and extensor strength or thigh muscle area may be inaccutriceps surae (dorsiflexor) strength measured using an rate. It is unlikely that the dynamometer used in this isometric dynam~rneter.~ The correlation between study to measure muscle strength would produce lower MAMA and biceps strength ( T = 0.78) was of the same correlations in the thigh due to variations in dynamommagnitude found in the present study. However, the eter measurements in the various muscle groups. This correlation in the calf area ( T = 0.62) was better than instrument has been extensively tested in clinical setthe correlation found in the thigh in this study ( T = tings with high reliabilit~.'~ In addition, measurements 0.29). of the same type have excellent test-retest reliability in The lower thigh correlation, when compared to the several muscle groups including those for knee flexion previously mentioned study of calf ~ t r e n g t h may , ~ be and exten~ion'~ and good interrater reliability when due to the multiple muscle groups present in the thigh studying knee e~tension.'~ These instruments also corarea, not all of which are involved in knee flexion and relate well with the clinical examination when measextension. Another explanation for the lower correla- uring knee extension strength.16 The reliability of the
560
REED ET AL
testing with the dynamometer used in this study was similar to that previously reported.14-16In addition, we have recently demonstrated a high correlation (r > 0.80) between this dynamometer and isokinetic average peak torque and work per repetition in both elbow and knee flexion and extension (unpublished observations). It is also noteworthy that there is a strong correlation between arm strength and leg strength (Figure 3). Although this may be coincidental, it may also be taken as an indication of the precision of the measurement as arm and leg strength would be expected to correlate strongly in most people. The anthropometric calculations also have some limitations as skin fold calculations may not accurately reflect intramuscular fat'7f found more frequently in the elderly. This may be a greater problem in the leg than the arm. In addition, bone area and circumference were not accounted for in these commonly used calculations. Refinements of current estimates are needed, based on ultras~und'~ or computerized tomography" in a large sample of the healthy elderly subjects. When analyzed by gender, the correlation between body composition and muscle strength was weaker. Similar findings are noted in other studies where stronger correlations are found when elderly subjects of both sexes are combined16, and the correlations are weaker when the study population is separated by gender.3 Combining data of male and female subjects enhances the correlations partly by providing a greater range of variability and may inflate the correlation coefficient (two-point phenomenon). However, a greater range of variability is often needed to determine true correlations in biologic systems where significant individual variation is frequently found. As there is no conclusive evidence that there is a difference in muscle function between the sexes, the inferences in this study are primarily made without reference to gender. Nevertheless, caution should be exercised when interpreting combined data of male and female subjects. There were some distinct differences between men and women in regard to the age-related changes observed and the correlation coefficients calculated. Of note is the finding that MAMA correlates with arm muscle strength and MTMA correlates with thigh muscle strength better in men than women. This may be due to inaccuracies of anthropomorphic measurement in women secondary to larger amounts of subcutaneous fat. In all muscle groups tested, strength declined significantly by age and gender. However, MAMA declined significantly in men with age but was stable in women. This may reflect a decrease in upper extremity training with increasing age present in men but not women. An intriguing observation is that with decreasing amounts of lean body mass in those above 65-years
IAGS-lUNE 1991-VOL. 39, NO. 6
old, strength per unit of lean body mass (efficiency) increases (Figure 5). This suggests that individuals with less muscle mass are stronger. However, this is not sufficient to counterbalance the age-related overall decline in strength per unit of lean body mass noted (Figure 6). The largest component of lean body mass is muscle. Muscle is also the major portion of lean body mass which declines with age.20Changes in the proportions of other components of lean body mass (eg bone) could conceivably influence the age-related change in the ratio of strength to unit of lean body mass. However, even when assuming that the change in LBM occurred exclusively either in muscle or in the other compartments of LBM, there would still be a significant component of overall decline in the ratio of strength to muscle in older subjects (calculations not presented). The formulae used for calculation of lean body mass was validated in young adults. Therefore, age related inaccuracies in estimates of lean body mass with bioelectric impedance might also alter the findings of this study. However, when formulae for men and women between 50 and 70 years of age, which were derived at our institution (T. Lohman; unpublished observations), were used for the subgroup of patients in this study who were 50 to 70 years old (n = 161), a statistically significant decrease in muscle efficiency was still noted (data not shown). This loss of strength per unit of muscle (efficiency) with increasing age is similar to that reported recently for male grip strength by the Baltimore Longitudinal Study on Agng.' Although further studies are needed to confirm this finding, the age-related loss suggests that as people get older they should make special efforts to increase their muscle mass through exercise and adequate nutrition. Simply maintaining muscle mass will probably be associated with loss of strength. A potential limitation of this study is its cross-sectional design. However, the Baltimore Longitudinal Study of Aging' noted similar amounts of change in hand-grip strength associated with aging in both crosssectional and longitudinal analyses. Future studies are needed to confirm these findings in longitudinal analyses and to use more accurate measures of regional muscle mass. In addition, it will be useful to do similar studies in the frail elderly to determine if there are preventable disease-related factors that contribute to muscle weakness in this high risk group. Future studies are needed to confirm the findings of this study in longitudinal analyses and to use more accurate measures of regional muscle mass. In addition, it will be useful to do similar studies in the frail elderly to determine if there are preventable disease-related factors that contribute to muscle weakness in this highrisk group.
MUSCLE MASS AND MUSCLE STRENGTH
IAGS-JUNE 1991-VOL. 39, NO. 6
REFERENCES 1 . Clement FJ. Longitudinal and cross-sectional assessments of age changes in physical strength as related to sex, social class, and mental ability. J Gerontol 1974;29:423. 2. Larsson L, Grimby G, Karlsson J. Muscle strength and speed of movement in relation to age and muscle morphology. J Appl Physiol 1979;46:451. 3. MacLennan WJ, Hall MR, Timothy JI, Robinson M. Is weakness in old age due to muscle wasting? Age Ageing 1980;9:188. 4 . Burke WE, Tuttle WW, Thompson CW et al. The relation of gnp strength and grip-strength endurance to age. J Appl Physiol 1988;5:628. 5 . Kallman DA, Plato CC, Tobin JD. The role of muscle loss in agerelated decline of grip strength: Cross-sectional and longitudinal perspectives. J Gerontol 1990;45:M82. 6. Danneskiold-Samsoe, Kofod V, Munter J et al. Muscle strength and functional capacity in 78-81 year-old men and women. Eur J Appl Physiol 1984;52:310. 7. Pearson MB, Bassey EJ, Bendall MJ. Muscle strength and anthropometric indices in elderly men and women. Age Ageing 1985;14:49. 8. Kendall FP, McCreary EK. Muscle Testing and Function. 3rd Ed. Baltimore, MD: Williams and Wilkins, 1983. 9 . Chumlea W, Roche AF, Mukherjee D. Some anthropometric indices of body composition for elderly adults. J Gerontol 1986;41:36. 10. Murray RL. Clinical methods in anthropometry. In: Kreg SH,
11. 12. 13. 14. 15. 16. 17. 18. 19. 20.
561
Murray RL, eds. Dynamics of Nutrition Support: Assessment, Implementation, Evaluation. Norwalk, CT: Appleton-Century Crofts, 1986, pp 99-146. Graves JE, Pollack ML, Colvin AB et al. Comparison of different bioelectric impedance analyzers in the prediction of body composition. Am J Hum Biol 1989;1:603. Norusis M. SPSS/PC+. Baltimore, MD: SPSS Inc., 1986. Hyde SA, Goddard CM. The myometer: The development of a clinical tool. Physiotherapy 1983;69:424. Bohannon RW. Test-retest reliability of hand-held dynamometry. Phys Ther 1986;66:206. Bohannon RW, Andrews AW. lnterrater reliability of hand-held dynamometry. Phys Ther 1987;67:931. Bohannon RW. Manual muscle test scores and dynamometer test scores of knee extension strength. Arch Phys Med Rehabil 1986;67:390. Rice CL, Cunningham DA, Paterson DH et al. Arm and leg composition determined by computed tomography in young and elderly men. Clin Physiol 1989;9:207. Rice CL, Cunningham DA, Peterson DH et al. A comparison of anthropometry with computed tomography in limbs of young and aged men. J Gerontol 1990;45:M175. Young A, Hughes I, Russel P et al. Measurement of quadriceps muscle wasting by ultrasonography. Rheum Rehabil 1980;19:141. Shepard JW. Interrelationship of exercise and nutrition in the elderly. In: Armbrecht HJ. Prendergast JM, Coe RM eds. Nutritional Intervention in the Aging Process. New York: SpringerVerlag, 1984, p. 316.
39:555-561, 1991
The Relationship between Muscle Mass and Muscle Strength in the Elderly Richard L. Reed, MD, MPH,*t Lori Pearlmutter, RPT,§ Kathleen Yochum, RN,GNP,* Keith E. Meredith, PhD,* and Arshag D. Mooradian, MD*$
To determine the extent that muscle mass is predictive of muscle strength in the elderly, anthropomorphic estimates of muscle area and impedance measurements of muscle mass and peak isometric muscle strength were obtained in a relatively healthy older population over 65 years of age (mean age = 71.7; n = 218). Midarm muscle area correlated strongly with upper arm strength (r = 0.68, P < 0.0001) while midthigh muscle area had a much lower correlation with thigh muscle strength (r = 0.29, P < 0.0001). These muscle area calculations also include bone area. Lean body mass calculated by bioelectric impedance correlated highly with cumulative muscle strength measured by summing all muscle groups (r =
< 0.0001). To determine whether aging alters muscle strength per unit of muscle mass, additional middle-aged subjects were included, and three groups, middle-aged (55-64) (n = 78, young-oId (65-74) (n = 161), and old-old (75+) (n = 57), were compared. A significant age-related trend of decreasing muscle strength per unit of lean body mass was noted. It is concfuded that although muscle mass correlates with muscle strength in a healthy older population, use of simple age-independent clinical measurements of body mass should not be used to predict muscle strength. J Am Geriatr SOC39:555561, 1991
adequate muscle strength is a basic requirement for physical functioning and has been shown to decrease with aging in multiple cross-sectional and longitudinal studies.’-’ This age-related loss of muscle strength puts the elderly at high risk for functional disability. The cause of this loss of strength is unknown and may be due to several factors. One possibility is that the loss of muscle strength is attributed totally to the loss of muscle mass associated with aging. Another possibility is that there is a loss of strength per unit of muscle mass (efficiency) associated with aging. Other possibilities include the influence of chronic diseases, decreasing physical activity, inadequate nutrition, other pathological changes in the muscle associated with aging, and decreased motivation in testing among the elderly. Previous studies addressing the relationship between
muscle mass and muscle strength have primarily been 5, These studies have done using grip ~trength.~, shown a variable degree of positive correlations between muscle strength and an index of muscle mass with a range of r values from 0.17 to 0.90. A recent study noted that in addition to an age-associated loss of muscle mass, there is an additional loss of handgrip strength per unit of body mass in males.’ Studies in other muscle groups have shown stronger correlation but have been limited to elbow extensor^,^ plantar flexor^,^ and knee extensors.6 In addition, they have not always used regional body mass measurements for correlation with the muscle group tested.6 The purpose of this study was to determine the extent that muscle mass is predictive of muscle strength in different large muscle groups in the healthy elderly and to determine whether aging is associated with altered muscle “efficiency”(muscle strength per unit muscle mass) in these muscle groups.
A
From the *Arizona Center on Aging and the Departments of tFamily Medicine and $Internal Medicine, University of Arizona College of Medicine, Tucson, Arizona; and SDepartment of Physical Therapy, University Medical Center, Tucson, Arizona. This research was supported in part by The Flinn Foundation and by the Hudson Foundation. Address correspondence to Richard L. Reed, University of Arizona, 1821 East Elm Street, Tucson, AZ 85719.
01991 by the American Geriatrics Society
0.79, P
METHODS Subjects were recruited at a health fair for a retirement community of approximately 17,000 people in southeastern Arizona. Individuals were divided into three age groups; middle-aged (55-64), young-old (650002-861 4/91 /$3.50
556
REED ET AL
74), and old-old (75 and over). In this study, elbow and knee flexor and extensor isometric muscle strength, as well as ankle dorsiflexor strength, was measured to the point of breaking with a hand-held dynamometer (Penny & Giles Transducers Limited, Christ Church, Dorset, U.K.) using the testing technique of Kendall and McCreary.’ Elbow flexor and extensor strength was measured in the sitting position with the forearm in supination with stabilization of the upper arm. Elbow flexor strength was tested with the elbow flexed at 90° and the arm beside the trunk with the dynamometer being applied to the anterior surface of the forearm just proximal to the wrist joint. Elbow extensor strength was measured with the elbow at 90° with the upper arm held perpendicular to the floor and the forearm in supination. The dynamometer was applied to the posterior surface of the forearm just proximal to the wrist joint. Knee flexors (biceps femoris, semitendinous, and semimembranous muscles) were measured in the prone position with the knee at 30° with application of the dynamometer at the ankle. The thigh was manually stabilized. Knee extensors (quadriceps femoris) were measured in the sitting position with the knee and hip flexed to 90°. The dynamometer was applied at the ankle with stabilization at the thigh. Ankle dorsiflexors (peroneus tertius, tibialis anterior) were also measured in the sitting position with the ankle in neutral position (perpendicular to lower leg). The lower limb was stabilized proximal to the ankle. The dynamometer was applied to the plantar surface of the foot just proximal to the metatarsalphalangeal joints. In this study the test-retest correlation for the muscle groups ranged from .92-.98 with the exception of right and left knee flexors (.78-20). There was no significant difference in the strength of each muscle group when test and retest values were compared ( n = 38). The whole study group was tested by only two of us (RLR and LP) with one person (RLR) testing the majority of patients (79%). The interrater reliability ranged in different muscle groups from .78-.90. Arm muscle strength was calculated by summing the forces for left elbow extensors and elbow flexors. Thigh muscle strength was calculated by the addition of left knee flexor and left knee extensor strength. Leg muscle strength was calculated by adding left dorsiflexor strength, left knee flexor, and left knee extensor strength. Cumulative muscle strength was calculated by summation of both the right and the left muscle strength for all 10 muscle groups tested. Anthropomorphic measurements were made on the left arm (midarm circumference, triceps skin fold) and left thigh (midthigh circumference and thigh skin fold). Skin fold thicknesses at these sites were measured with a Lange caliper (Cambridge Instrument Co., Cambridge, MD). Midarm muscle area (MAMA) was cal-
IAGS-]UNE 1991-VOL. 39, NO. 6
culated from triceps skin fold (TSF) and midarm circumference (MAC) measurements using the equation lo MAMA = (MAC in cM - (T X TSF in mM/10))2/4~.9, Arm muscle and bone circumference (AMC) were calculating using the equation AMC = MAC in cM - (T X TSF in cM). Midthigh muscle area (MTMA) and thigh muscle circumference (TMC) were calculated with similar formulas using midthigh circumference and midthigh skinfold thickness. It should be noted that all of the calculations are not corrected for bone area or circumference. Hence, the labels given are somewhat of a misnomer although they follow the previously established convention. Height (Ht) and weight (Wt) were measured using an upright beam scale. The body mass index (BMI) was calculated from these values as Wt/Ht2 (Kg/M2). Lean body mass (LBM) was calculated as the difference between body weight and fat mass as determined by a bioelectric impedance analyzer (Valhalla Co., Model 1990B, San Diego, CA). The analyzer was calibrated prior to each testing session with an internal 499-ohm resister supplied with the unit. Subjects were measured in the supine position, lying on a non-conductive elevated surface. The arms were resting at the sides and the legs were spread apart to prevent contact. One detection electrode was placed between the distal condyles of the ulna and radius on the posterior surface of the right wrist. The other detection electrode was attached between the most prominent portions of the malleoli of the fibula and tibia on the anterior surface of the ankle. The two-source electrodes were placed at the base of the third metacarpal-phalangeal joint and 1-centimeter proximal to the third metatarsal-phalangeal joint. An excitation current was introduced at a frequency of 50 kHz and the resistance was measured. Ambient temperature was relatively constant for all subjects during this study. Lean body mass was calculated with a formula specifically derived for the Valhalla instrument in comparison with underwater weighing.” The equations with height (Ht) expressed in centimeters and resistance (R) in ohms were: LBM = 0.475 X Ht2/R + 5.32 for women and LBM = 0.827 X Ht2/R + 5.49 for men. Statistical analysis was performed using the Statistical Package for the Social Sciences.” A two-way, fixed effects analysis of variance with age group and sex as independent variables was calculated on each of the strength and mass variables. The Pearson productmoment correlation coefficient was used for correlating mass and strength using two-tailed probabilities. Slope was calculated using ordinary least squares regression analysis. One-way fixed effects analysis of variance was done separately for males and females, with the tertile of lean body mass calculated for each sex separately as the independent variable and cumulative muscle strength per kilogram of lean body mass (effi-
557
MUSCLE MASS AND MUSCLE STRENGTH
JAGS-]LINE 1991-VOL. 39, N O . 6
the measurements in males decreased with increasing age while little or no change was noted in females. In individuals over the age of 65, MAMA strongly correlated with arm muscle strength measured as the sum of the force in kilograms produced by elbow flexors and elbow extensors (r = 0.68, P < 0.0001) RESULTS (Figure 1). AMC similarly correlated with arm muscle There were 296 subjects recruited from a total of 326 strength (r = 0.68, P < 0.0001). When divided into individuals participating in the health fair. Excluded males and females the level of correlation between were 30 individuals who refused muscle or body mass MAMA and arm muscle strength was lower (males = testing or were unable to perform all the muscle tests .36, P < 0.0001; females = .23, P < 0.05). A similar due to neuromuscular or joint problems. All the sub- pattern was found for the correlation of AMC with jects were ambulatory and active Caucasians in middle arm muscle strength (r = 0.36, P < 0.0001 for males; Y or upper middle socioeconomic class. Although they = 0.23, P C 0.05 for females). In those 65 and older, MTMA correlated less strongly were not disease free, review of their medical history and medications used suggested that they were a rel- with thigh muscle strength (r = 0.29, P < 0.0001) than atively healthy group. Individuals who were 65 and did MAMA with arm strength (Figure 2). This correlaolder (n = 218; mean age (fSE) = 71.7 f 0.4 years) tion was stronger in males (r = 0.37, P < 0.0001) than were the primary study group. However, for compari- females (r = 0.19, P < 0.05). TMC had a similar level son purposes, measurements were made on an addi- of correlation with thigh muscle strength ( Y = 0.29, P tional 78 middle-aged subjects (mean age (+SE) = 53.1 5 0.0001) that was also stronger in males (r = 0.37, P < 0.0001) than females ( r = 0.19, P C 0.05). & 1.5 years). The correlation of arm muscle strength with leg Muscle strength declined with age in all muscle strength was also strong (r = 0.80, P < 0.0001) and groups (Table 1).As expected, all muscle groups tested had a similar magnitude in both women (r = 0.62, P < were also weaker in women than in men. In addition, many of the anthropomorphic and body composition 0.0001) and men (r = 0.58, P < 0.0001) (Figure 3). values usually declined with age (Table 2). However, Lean body mass calculated by impedance bioelectric AMC, MAMA, thigh skinfold, TMC, percent body fat measurements correlated strongly with cumulative and body mass index were not significantly different muscle strength (r = 0.79, P < 0.0001) (Figure 4). There in the three age groups studied. All anthropomorphic was a weaker correlation when separated by gender (Y and body composition measures differed significantly = 0.41, P 5 0.0001 for males and r = 0.32, P 5 0.0001 by gender with the exception of thigh muscle circum- for females). When the ratio of muscle strength divided ference and thigh muscle area. A statistically significant by lean body mass was analyzed for the elderly (65 interaction between age and sex was found with only and over) by tertiles of lean body mass, a significant two variables, AMC and MAMA. In both situations, difference was noted using one-way analysis of vari-
ciency) as the dependent variable. Two-way fixed effects analysis of variance was also performed with age group and sex as the independent variables and cumulative muscle strength per kilogram of lean body mass (efficiency) as the dependent variable.
TABLE 1. THE MEAN (+SE) OF MUSCLE STRENGTH (Kn)BY AGE GROUP AND GENDER* Middle-Aged Y oung-Old Old-Old Variable
Males n = 33
Females n = 45
Males n = 74
Females n = 87
Upper Extremity 16.7 f 0.4 R elbow flexor 29.7 f 0.6 19.2 f 0.6 26.7 f 0.5 16.0 f 0.4 L elbow flexor 28.5 f 0.6 18.3 f 0.5 26.0 f 0.5 R elbow extensor 12.8 f 0.3 22.6 f 0.7 14.3 f 0.4 19.5 f 0.4 12.4 f 0.3 L elbow extensor 21.8 f 0.7 14.3 f 0.4 18.5 f 0.4 28.4 k 0.6 Combined L arm 50.2 f 0.8 32.7 f 0.8 44.5 f 0.9 Lower Extremity R knee flexor 13.8 f 0.5 26.2 f 1.0 18.0 f 0.8 22.0 f 0.5 L knee flexor 13.8 f 0.4 25.8 f 0.9 17.7 f 0.8 22.2 k 0.5 21.3 f 0.6 R knee extensor 30.0 f 0.6 23.9 f 0.8 27.8 f 0.5 L knee extensor 21.4 f 0.6 30.4 f 0.7 24.0 f 0.8 28.4 f 0.5 Combined L thigh 56.2 f 1.5 41.8 f 1.2 50.5 f 0.8 34.2 f 0.9 R dorsiflexor 21.5 f 0.6 30.2 f 0.6 22.0 f 0.8 28.1 f 0.5 L dorsiflexor 20.8 f 0.6 29.4 f 0.6 22.3 f 0.8 27.9 f 0.5 Combined L leg 48.4 f 1.2 80.9 f 2.0 57.7 f 1.7 72.3 f 1.4 Cumulative Muscle Strength 299.9 f 0.5 211.7 f 4.9 269.9 f 3.8 183.4 f 3.6 * A f f measurements were Significantly different (P < 0.05)by age and sex using analysis of variance.
Males n = 26
Females n = 31
24.3 f 0.7 24.2 f 0.8 18.3 f 0.7 18.0 f 0.7 42.2 k 1.3
15.5 f 0.5 15.1 f 0.4 11.6 f 0.4 11.1 f 0.4 26.2 f 0.7
18.7 f 1.2 18.8 f 1.2 25.5 f 1.2 25.4 f 1.1 44.3 f 2.0 26.5 f 0.8 25.9 f 0.9 64.5 f 2.8 244.4 f 7.5
12.6 f 0.6 12.3 f 0.5 19.7 f 0.8 19.5 f 0.8 31.8 f 1.2 18.5 f 0.8 17.8 f 0.8 42.4 f 1.7 166.1 f 4.6 ~~~
~~
558
REED ET AL
IAGS-JUNE 1991-VOL. 39, NO. 6
TABLE 2. THE MEAN (fSE) OF ANTHROPOMORPHIC MEASUREMENTS AND BODY COMPOSITION OF THE STUDY SUBJECTS Middle- Aged Young-Old Old-Old Males n = 33
Females n = 45
Males n = 74
Females n = 87
Males n = 26
32.6 f 0.5
28.3 f 0.2
31.2 f 0.3
28.9 f 0.3
29.8 f 0.5
27.6 f 0.5
18.5 f 1.1 26.8 f 0.4
24.0 f 0.8 20.8 f 0.3
18.3 f 0.8 25.5 k 0.4
24.0 f 0.7 21.4 k 0.2
16.9 f 1.1 24.5 f 0.4
21.2 f 1.1 21.0 f 0.3
57.8 f 1.9
34.9 f 1.0
52.6 f 1.3
36.9 f 0.8
48.3 f 1.6
35.3 k 1.0
55.3 f 1.6 21.8 f 1.9 47.5 f 1.8
55.9 f 0.8 30.5 f 1.0 46.3 f 0.8
52.7 f 0.7 18.8 f 0.9 46.9 f 0.8
56.1 f 0.6 29.8 f 0.9 46.7 f 0.5
51.5 f 0.7 19.7 f 1.1 45.3 f 0.7
53.9 f 1.0 27.1 f 1.4 45.3 f 0.7
187.5 f 7.9
174.2 f 0.5
179.3 f 4.3
172.3 f 3.7
164.3 f 4.8
164.9 f 5.2
22.0 f .01 28.0 f .01 65.3 f 8.5 45.0 f 0.7 1.78 f .02 1.64 f .01 84.5 f 2.6 63.5 f 1.4 26.7 f 0.6 24.2 f 0.8 * Statistically significant difference by age (P< 0.05) ** Statistically significant difference by gender (P < 0.05) Statistically significant interaction of age and sex f P < 0.05)
22.0 f .01 62.1 f 0.6 1.77 f .01 80.0 f 1.0 25.6 f 0.3
30.0 f .01 44.4 f 0.5 1.63 f .01 63.8 f 1.1 24.2 f 0.4
21.0 f .01 59.9 f 1.4 1.75 f .01 75.8 f 2.0 24.9 f 0.6
29.0 f .01 42.9 f 0.8 1.59 f .01 61.0 f 1.6 24.2 & 0.6
Upper Extremity Arm circumference (cm)*,** Triceps skin fold (mm)*,** Arm muscle circumference (cm) (AMC)** Arm muscle area (cm') (MAMA)*, ** Lower Extremity Thigh circumference*, ** Thigh skin fold*, ** Thigh muscle circumference (cm) (TMC) Thigh muscle area (cm') (MTMA)* Total Body Percent fat by impedance** Lean body mass (Kg)*, ** Height (meters)*,** Weight (Kg)*, ** Body mass index**
601
Females n = 31
8olo . ...-. - .
/
7o
50
Male Female
60 0
t
45
S? 40 40 30 -
-
I
I
I
30
40
50
b
20
20 Male
.
I
I
1
60 70 80
I 90
Midarm Muscle Area (Cm2)
10 100
150
200
250
300
Midthigh Muscle Area (Cm2)
FIGURE 1. The relationship between midarm muscle area and upper arm strength in subjects over 65 years of age. The Pearson product moment correlation is .68, P < 0.0001. The linear regression equation (fSE) is y = 0.59 (.04) X +9.51 (1.92). Midarm muscle area also includes bone area.
FIGURE 2. The relationship between midthigh muscle area and thigh strength in subjects over 65 years of age. The Pearson product moment correlation is .29, P < 0.0001. The linear regression equation (fSE) is y = 0.10 (.02) X +24.62 (3.63). Midthigh muscle area also includes bone area.
ance for men and women separately (Figure 5). Individuals who have lower lean body mass are smaller and have less muscle mass. When this ratio of cumulative strength divided by lean body mass (efficiency) was calculated and analyzed for all three age groups using two-way analysis of variance (Figure 6) a statistically significant age effect was noted (P < 0.0001) with a decline with increasing age.
DISCUSSION This study demonstrates that there is a strong correlation between lean body mass and cumulative muscle strength in a relatively healthy community-dwelling elderly population (Figure 4). Regional muscle mass estimates and muscle strength correlations were also strong for the arm (r = 0.68, P 5 0.0001) but had a
MUSCLE MASS AND MUSCLE STRENGTH
IAGS-]LINE 1991-VOL. 39, NO. 6
60 8
5.0
Female Male
0 Male
50
y"
%
0
v
5
Female
4.5
al .0 .c
40
cn C
559
4.0
2
-al
5
0 u)
30
E, a
ga
r'
:. 0.y.
20
2
'
34
0.0
/
10
10
3.5
Lowest Third
1
30
20
40
50
60
70
Leg Strength (Kg) FIGURE 3. The relationship between arm strength and leg strength in subjects over 65 years of age. The Pearson product moment correlation is 30, P < 0.0001. The linear regression equation (2SE) is y = 0.47 (.02) X +8.08 (1.44).
Highest Third
FIGURE 5. Comparison of the means (+SE) of the cumulative muscle strength per kilogram of lean body mass (efficiency) by tertile of lean body mass in elderly males and females. P-value < 0.05 by one-way analysis of variance. Tertiles created for each sex separately.
5.0
350 1 % h
0
y"
al ._
-
v
Middle Third
4.5
r
nMale Female Both Sexes
.-0
5 U J C
4.0
0) -
??
0 m
m
r'
200 150
3.5
0.0 __
8
50
30
40
50
60
70
Female Male 80
90
Lean Body Mass (Kg) FIGURE 4. The relationship between lean body mass and cumulative muscle strength in subjects over 65 years of age. The Pearson product moment correlation is .79,P < 0.0001. The linear regression equation (+SE) is y = 4.16 (.22) + 1.70 (11.73).
Middle-aged Young-old Old-old FIGURE 6. Comparison of the means (+SE) of cumulative muscle strength per kilogram of lean body mass (efficiency) by age group in males and females. Overall P < 0.0001) by two-way analysis of variance. Significant differences were noted by age (P < 0.001) but not sex. No sigruficant interactions were noted.
tion in the thigh is that maximal recruitment of muscles, especially in this region, may not have been achieved. We feel that all of the subjects were highly motivated and gave maximal voluntary efforts. It is unknown, weaker correlation for the thigh ( T = 0.29, P 5 0.0001). however, whether an age-related difference in recruitIn another study of subjects over the age of 65, anthro- ment could explain the lower correlation in this study. pomorphic measurements of arm and calf muscle area Alternatively, the measurement of knee flexor and were compared to biceps brachii (elbow flexor) and extensor strength or thigh muscle area may be inaccutriceps surae (dorsiflexor) strength measured using an rate. It is unlikely that the dynamometer used in this isometric dynam~rneter.~ The correlation between study to measure muscle strength would produce lower MAMA and biceps strength ( T = 0.78) was of the same correlations in the thigh due to variations in dynamommagnitude found in the present study. However, the eter measurements in the various muscle groups. This correlation in the calf area ( T = 0.62) was better than instrument has been extensively tested in clinical setthe correlation found in the thigh in this study ( T = tings with high reliabilit~.'~ In addition, measurements 0.29). of the same type have excellent test-retest reliability in The lower thigh correlation, when compared to the several muscle groups including those for knee flexion previously mentioned study of calf ~ t r e n g t h may , ~ be and exten~ion'~ and good interrater reliability when due to the multiple muscle groups present in the thigh studying knee e~tension.'~ These instruments also corarea, not all of which are involved in knee flexion and relate well with the clinical examination when measextension. Another explanation for the lower correla- uring knee extension strength.16 The reliability of the
560
REED ET AL
testing with the dynamometer used in this study was similar to that previously reported.14-16In addition, we have recently demonstrated a high correlation (r > 0.80) between this dynamometer and isokinetic average peak torque and work per repetition in both elbow and knee flexion and extension (unpublished observations). It is also noteworthy that there is a strong correlation between arm strength and leg strength (Figure 3). Although this may be coincidental, it may also be taken as an indication of the precision of the measurement as arm and leg strength would be expected to correlate strongly in most people. The anthropometric calculations also have some limitations as skin fold calculations may not accurately reflect intramuscular fat'7f found more frequently in the elderly. This may be a greater problem in the leg than the arm. In addition, bone area and circumference were not accounted for in these commonly used calculations. Refinements of current estimates are needed, based on ultras~und'~ or computerized tomography" in a large sample of the healthy elderly subjects. When analyzed by gender, the correlation between body composition and muscle strength was weaker. Similar findings are noted in other studies where stronger correlations are found when elderly subjects of both sexes are combined16, and the correlations are weaker when the study population is separated by gender.3 Combining data of male and female subjects enhances the correlations partly by providing a greater range of variability and may inflate the correlation coefficient (two-point phenomenon). However, a greater range of variability is often needed to determine true correlations in biologic systems where significant individual variation is frequently found. As there is no conclusive evidence that there is a difference in muscle function between the sexes, the inferences in this study are primarily made without reference to gender. Nevertheless, caution should be exercised when interpreting combined data of male and female subjects. There were some distinct differences between men and women in regard to the age-related changes observed and the correlation coefficients calculated. Of note is the finding that MAMA correlates with arm muscle strength and MTMA correlates with thigh muscle strength better in men than women. This may be due to inaccuracies of anthropomorphic measurement in women secondary to larger amounts of subcutaneous fat. In all muscle groups tested, strength declined significantly by age and gender. However, MAMA declined significantly in men with age but was stable in women. This may reflect a decrease in upper extremity training with increasing age present in men but not women. An intriguing observation is that with decreasing amounts of lean body mass in those above 65-years
IAGS-lUNE 1991-VOL. 39, NO. 6
old, strength per unit of lean body mass (efficiency) increases (Figure 5). This suggests that individuals with less muscle mass are stronger. However, this is not sufficient to counterbalance the age-related overall decline in strength per unit of lean body mass noted (Figure 6). The largest component of lean body mass is muscle. Muscle is also the major portion of lean body mass which declines with age.20Changes in the proportions of other components of lean body mass (eg bone) could conceivably influence the age-related change in the ratio of strength to unit of lean body mass. However, even when assuming that the change in LBM occurred exclusively either in muscle or in the other compartments of LBM, there would still be a significant component of overall decline in the ratio of strength to muscle in older subjects (calculations not presented). The formulae used for calculation of lean body mass was validated in young adults. Therefore, age related inaccuracies in estimates of lean body mass with bioelectric impedance might also alter the findings of this study. However, when formulae for men and women between 50 and 70 years of age, which were derived at our institution (T. Lohman; unpublished observations), were used for the subgroup of patients in this study who were 50 to 70 years old (n = 161), a statistically significant decrease in muscle efficiency was still noted (data not shown). This loss of strength per unit of muscle (efficiency) with increasing age is similar to that reported recently for male grip strength by the Baltimore Longitudinal Study on Agng.' Although further studies are needed to confirm this finding, the age-related loss suggests that as people get older they should make special efforts to increase their muscle mass through exercise and adequate nutrition. Simply maintaining muscle mass will probably be associated with loss of strength. A potential limitation of this study is its cross-sectional design. However, the Baltimore Longitudinal Study of Aging' noted similar amounts of change in hand-grip strength associated with aging in both crosssectional and longitudinal analyses. Future studies are needed to confirm these findings in longitudinal analyses and to use more accurate measures of regional muscle mass. In addition, it will be useful to do similar studies in the frail elderly to determine if there are preventable disease-related factors that contribute to muscle weakness in this high risk group. Future studies are needed to confirm the findings of this study in longitudinal analyses and to use more accurate measures of regional muscle mass. In addition, it will be useful to do similar studies in the frail elderly to determine if there are preventable disease-related factors that contribute to muscle weakness in this highrisk group.
MUSCLE MASS AND MUSCLE STRENGTH
IAGS-JUNE 1991-VOL. 39, NO. 6
REFERENCES 1 . Clement FJ. Longitudinal and cross-sectional assessments of age changes in physical strength as related to sex, social class, and mental ability. J Gerontol 1974;29:423. 2. Larsson L, Grimby G, Karlsson J. Muscle strength and speed of movement in relation to age and muscle morphology. J Appl Physiol 1979;46:451. 3. MacLennan WJ, Hall MR, Timothy JI, Robinson M. Is weakness in old age due to muscle wasting? Age Ageing 1980;9:188. 4 . Burke WE, Tuttle WW, Thompson CW et al. The relation of gnp strength and grip-strength endurance to age. J Appl Physiol 1988;5:628. 5 . Kallman DA, Plato CC, Tobin JD. The role of muscle loss in agerelated decline of grip strength: Cross-sectional and longitudinal perspectives. J Gerontol 1990;45:M82. 6. Danneskiold-Samsoe, Kofod V, Munter J et al. Muscle strength and functional capacity in 78-81 year-old men and women. Eur J Appl Physiol 1984;52:310. 7. Pearson MB, Bassey EJ, Bendall MJ. Muscle strength and anthropometric indices in elderly men and women. Age Ageing 1985;14:49. 8. Kendall FP, McCreary EK. Muscle Testing and Function. 3rd Ed. Baltimore, MD: Williams and Wilkins, 1983. 9 . Chumlea W, Roche AF, Mukherjee D. Some anthropometric indices of body composition for elderly adults. J Gerontol 1986;41:36. 10. Murray RL. Clinical methods in anthropometry. In: Kreg SH,
11. 12. 13. 14. 15. 16. 17. 18. 19. 20.
561
Murray RL, eds. Dynamics of Nutrition Support: Assessment, Implementation, Evaluation. Norwalk, CT: Appleton-Century Crofts, 1986, pp 99-146. Graves JE, Pollack ML, Colvin AB et al. Comparison of different bioelectric impedance analyzers in the prediction of body composition. Am J Hum Biol 1989;1:603. Norusis M. SPSS/PC+. Baltimore, MD: SPSS Inc., 1986. Hyde SA, Goddard CM. The myometer: The development of a clinical tool. Physiotherapy 1983;69:424. Bohannon RW. Test-retest reliability of hand-held dynamometry. Phys Ther 1986;66:206. Bohannon RW, Andrews AW. lnterrater reliability of hand-held dynamometry. Phys Ther 1987;67:931. Bohannon RW. Manual muscle test scores and dynamometer test scores of knee extension strength. Arch Phys Med Rehabil 1986;67:390. Rice CL, Cunningham DA, Paterson DH et al. Arm and leg composition determined by computed tomography in young and elderly men. Clin Physiol 1989;9:207. Rice CL, Cunningham DA, Peterson DH et al. A comparison of anthropometry with computed tomography in limbs of young and aged men. J Gerontol 1990;45:M175. Young A, Hughes I, Russel P et al. Measurement of quadriceps muscle wasting by ultrasonography. Rheum Rehabil 1980;19:141. Shepard JW. Interrelationship of exercise and nutrition in the elderly. In: Armbrecht HJ. Prendergast JM, Coe RM eds. Nutritional Intervention in the Aging Process. New York: SpringerVerlag, 1984, p. 316.
