249
Normality, Variability and Predictability of Work, Power and Torque Acceleration Energy with Respect to Peak Torque in Isokinetic Muscle Testing P. Kannus Department of Orthopaedics and Rehabilitation, University of Vermont, Burlington, Vermont, USA
Introduction
P. Kannus, Normality, Variability and Pre-
dictability of Work, Power and Torque Acceleration Energy with Respect to Peak Torque in Isokinetic Muscle Testing. mt J Sports Med, Vol 13, No 3, pp 249—256, 1992.
Accepted after revision: October21, 1991 This study evaluated at two different test sessions the normality and variability of the isokinetic peak torque (PT), peak work (PW), peak power (PP) and peak torque acceleration energy (PTAE) data outputs in healthy adult males (n = 10) and females (n = 10). The hamstring and quadriceps muscles were tested at the angular velocities of 60 deg/s (a slow speed test) and 240 deg/s (a high speed test). The predictability of the PW, PP and PTAE from the PT was also assessed. The results showed that the consistency of the PW and PP measurements were equal with that of the PT. This was due to equal (almost normal)
data distribution, equal variability of the outputs (the coefficient of variation (cv) ranged from 14 to 29% in the PWs and PPs versus 16 to 29% in the PTs), and excellent
predictability of the PW and PP from the PT (PTs accounted on an average 85% for the variation seen in the PWs and PPs). In addition, in the regression analyses the standard errors of the estimates (SEEs) were low (< 10%) and the residuals were distributed nonsystematically. In the PTAE measurements, the results were much more inconsistent, especially during the slow speed of the dynamometer. Compared with PT, PW and PP, the PTAE data distribution differed more frequently from normal distribution and the PTAE outputs showed higher variability. In addition, the PTAE outputs could not be acceptably predicted from the PT. In conclusion, the isokinetic PW and PP measurements can be recommended for clinical use, while the PTAE measurements should not be used routinely. Key words
Isokinetics, variability, reproducibility, peak torque _____________________________________________
Int.JSportsMed. 13(1992)249—256 GeorgThiemeVerlagStuttgartNewYork
The development of commercially available isokinetic dynamometers has instigated considerable research on the 'in vivo' characteristics of human muscles. The force or torque generating capabilities of human muscles have been extensively reported in the literature (18, 31). In general, all current dynamometers measure either torque or moment of force at various angular positions and speeds.
The most frequently used isokinetic measurement in clinical and scientific work has been peak torque (PT, newton meters, Nm), referring to the single highest torque output of the joint produced by muscular contraction as the limb moves through the range of motion (4, 12, 15, 23, 27). PT has been shown to be an accurate and highly reproducible variable to measure (3, 19, 20, 22, 23, 33), and its use in isokinetic tests has been accepted in critical reviews (2, 28, 31). PT has become a "gold standard and reference point" in all isokinetic measurements. However, interfacing of microprocessors with isokinetic dynamometers has enabled the rapid quantification of many other, more specific muscle function parameters than PT including peak or total work, peak or average power, and peak torque acceleration energy (15, 23, 27). Peak work (PW, joules, J) is defined as the area under the torque versus angular displacement (time) curve during the best repetition (work = torque x distance) (5, 12, 23); peak power (PP, watts, W) as
the amount of work achieved during the best test repetition divided by the contraction time (5, 23); and peak torque accel-
eration energy (PTAE, joules, J) as the greatest amount of work performed in the first 125 milliseconds of a single torque
production in the test repetitions (5, 23). The latter has been said to be indicative of muscular "explosiveness" giving an estimate about the speed (rate) of the torque production (5).
Recently, the relationships between PT and total work, and PT and average power have been a subject of interest (13, 14, 16, 17, 21, 23). These studies have observed that the predictability of total work and average power from PT is generally good and acceptable. However, no research has evaluated the value and consistency of the outputs of the peak work (PW), peak power (PP) and peak torque acceleration energy (PTAE) measurements, or their relationship to PT. The aims of the present investigation were to characterize the normality and variability of the PW, PP and . PTAE data outputs with respect to PT normality and variabil-
Downloaded by: Universite de Sherbrooke. Copyrighted material.
Abstract
P. Kannus
250 mt. J. Sports Med. 13(1992) ity in healthy males and females, and to determine the relationships between PT and the above mentioned parameters in multiple, low (60 deg/s) and high (240 deg/s) speed contractions
of the hamstring and quadriceps muscles. The assumptions
were that if the PWs, PPs and PTAEs are as valuable as the PT measurements, 1) their variability (coefficient of variation = cv) should not be larger than PT variability; 2) men and wom-
en should show almost equal PW, PP and PTAE variation (equal cv values) and these parameters should be equally related to PT values (i. e., PW, PP and PTAE expressed as a percentage of PT should be equal between men and women); 3)
Prior to testing each subject underwent a 5-mm warm-up period of aerobic, low-resistance ergometer cycling.
The subjects were then fixed to the testing device with straps round the chest, pelvis, thigh and ankle. The ankle strap was placed just above the ankle. The axis of the knee was placed in line with the axis of rotation of the dynamometer. Subjects were not allowed to grasp the handles of the bench during testing.
the level of normality in the PW, PP and PTAE data distribution should be equal with that of the corresponding PTs;
Each limb was weighed before testing by the Cybex's automatic limb weighing system to correct for the gravitational effect on torque (GET). Each estimate of the GET was added to the quadriceps torque and subtracted from
and 4) the predictability of the PWs, PPs and PTAEs from the corresponding PTs should be fully acceptable.
the hamstring torque. The subjects were then instructed to perform five to ten submaximal (50%) repetitions of knee exten-
sion and flexion at the two isokinetic testing velocities (60 deg/s and 240 deg/s) to become familiar with the testing dev-
Methods Twenty healthy adults (10 male and 10 female aged 20—40 years) volunteered to participate in the study.
Thereafter, the isokinetic concentric muscle
tional physical activities, but none of them was engaged in regular strength training or competitive sports.
performance test was started. First (at 60 deg/s) two submaximal and two maximal repetitions of knee extension and flexion were performed for practice, and then, after a 30-second rest, five maximal voluntary repetitions of alternating knee extension and flexion occurred. Subjects received consistent verbal encouragement throughout all tests. Special attention was
Table I
paid to the instruction to maximally exert each contraction over the full range of motion; that is, from the very beginning
Characteristics of the male and female subjects are reported in
Table 1. They had healthy knees, used no medication, and were moderately active. They participated in various recrea-
Characteristics of the study subjects (mean±SD).
Age (years) Height (cm)
Weight (kg BMI (kg/rn )
Men
Women
(n=10)
(n=l0)
28.9±4.7 174.0±8.0
30.0±6.1 167.5±4.5 60.0±8.3 21 .3±2.2
75.4± 10.8
25.1±3.0
BMJ = Body Mass Index = weight/height2.
Before the study the subjects were fully informed of the study procedure and design, purposes, and any known risks and gave their informed consent. The study was conducted in conformity with the principles of the Declaration of Helsinki of 1975.
Study Design, Instrumentation and Testing Procedure The study design was to first isokinetically assess the muscle performance of the knee extensors and flex-
ors of both limbs and analyze the normality, variability and predictability of PW, PP and PTAE with respect to PT. After eight weeks, the same test protocol was repeated in order to de-
termine the consistency of the results achieved in the first phase.
For the PT, PW, PP and PTAE measurements of the quadriceps and hamstrings of both legs, Cybex 340 dynamometer was used (Division of Lumex, Inc., Ronkonkoma, NY, 11779). In a random order of the subjects, the limb to be tested first was the dominant or the nondominant limb. The limb dominance was determined according to the limb preference in kicking a ball and taking-off in jumping.
to the very end of the movement.
After a two-minute rest, the same protocol was
repeated at the speed of 240 deg/s. After a five-minute rest (walking and recovery stretching) the same protocol was performed on the other leg.
Measured Variables The measured variables were PT, PW, PP and PTAE. Their definitions and units are reported in the "Introduction". Since the peak value of the torque production is the generally accepted reference parameter in isokinetics, the peak values of the other parameters (work, power and torque acceleration energy) were recorded and used for the analysis, too.
Data Analyses The test and retest data were analysed with the same statistical procedures by an IBM compatible 286 microcomputer, using the 1990 version of the statistical program library of the Biomedical Data Processing (BMDP) Software Inc., Los Angeles, California (8). The measured variables were all interval scaled by nature, and, therefore, the results are re-
ported as mean standard deviation (SD) with 95% confidence interval throughout the study.
The normality of the data distribution within each parameter was tested by the Shapiro and Wilk's W test (29). The coefficient of variation (cv SD/mean x 100%) was used to evaluate the data variability. For quadriceps and hamstrings of both groups, the Pearson product moment correlation coefficient (r) and its square (r2, the coefficient of determination) were calculated between the PT and PW, PT and
Downloaded by: Universite de Sherbrooke. Copyrighted material.
ice.
Subjects
mt. J. Sports Med. 13(1992) 251
Normality, Variability and Predictability of Work, Power and Torque Acceleration Energy
Table 2 Hamstring and quadriceps peak torque (PT), peak work (PW), peak power (PP) and peak torque acceleration energy (PTAE) at two isokinetic velocities (60 and 240 deg/s).
Men*
mean±SD
Women*
W"
p
W"
p
(78—92) (142—166)
.90 .97
.05 .67
mean±SD (95% confidence interval)
(95% confidence interval) PT (Nm)
60 deg/s Hamstrings Quadriceps
240 deg/s Hamstrings Quadriceps
140±36 227±43
(123—157) (206—247)
.93 .89
.15 .03
85±15 154±25
89±26 135±27
( 76—101)
.96 .96
.46 .60
54±10 85±14
(49—58) (78—92)
.91 .97
.05 .82
(53—61) (87—100)
.94 .97
.23 .79
(32—39) (58—62)
.94 .95
.30 .36
(59—68) (99—113)
.97
.72 .84
(122—147)
60 deg/s Hamstrings Quadriceps
97±24
( 86—108) (130—155)
.97 .96
.79 .46
57±9
143±27
240 deg/s Hamstrings Quadriceps
63± 18
(54—71) (82—98)
.97 .95
.72 .34
36±8
90±17
60 deg/s Hamstrings Quadriceps
103±23 153±27
( 92—1 14) (140—166)
.99 .95
.97
64± 10
.41
106± 15
240 deg/s Hamstrings Quadriceps
238±65 346±64
(207—268) (316—376)
.97 .97
.65
138±28
224±38
(125—151) (206—241)
.94 .96
.21
.68
4.5±0.9
(4.0—4.9) (4.5—5.8)
.81
.001
2.8±0.9
.08
4.6 1.6
(2.3—32)
.91
.86 .94
.007 .24
.91
.08 .45
11.4±2.6 16.9±3.9
.91
.06 .97
93± 14
58±10
PP (W)
.97
.59
PTAE (J)
60 deg/s Hamstrings Quadriceps
5.2
1.5
240 deg/s Hamstrings Quadriceps
19.5±5.1 24.6±5.4
(17.0—21.9) (22.0—27.1)
.95
(3.9—5.3)
(10.1—12.6) (15.0—18.7)
.99
* Both legs of the ten men and ten women were tested giving 20 measurements for both groups. * *w statistics tests the normality of the distribution (29). W-value .90 with p .05 means that the distribution of the data does not significantly differ from normal distribution.
for the study (beta level or type II error 0.10). Since there were not significant differences between the results of the right and
PP, and PT and PTAE. The residual error or each correlation was analyzed by a linear regression technique. In the regression equations, the PT values were placed on the x-axis and the
left limbs in any of the measured parameter, the data were
PW, PP and PTAE values on the y-axis, respectively.
pooled and analyzed together.
The given significance levels refer to two-tailed
tests. In all tests an alpha level (type I error) less than 0.1 % (p
Normality, Variability and Predictability of Work, Power and Torque Acceleration Energy with Respect to Peak Torque in Isokinetic Muscle Testing P. Kannus Department of Orthopaedics and Rehabilitation, University of Vermont, Burlington, Vermont, USA
Introduction
P. Kannus, Normality, Variability and Pre-
dictability of Work, Power and Torque Acceleration Energy with Respect to Peak Torque in Isokinetic Muscle Testing. mt J Sports Med, Vol 13, No 3, pp 249—256, 1992.
Accepted after revision: October21, 1991 This study evaluated at two different test sessions the normality and variability of the isokinetic peak torque (PT), peak work (PW), peak power (PP) and peak torque acceleration energy (PTAE) data outputs in healthy adult males (n = 10) and females (n = 10). The hamstring and quadriceps muscles were tested at the angular velocities of 60 deg/s (a slow speed test) and 240 deg/s (a high speed test). The predictability of the PW, PP and PTAE from the PT was also assessed. The results showed that the consistency of the PW and PP measurements were equal with that of the PT. This was due to equal (almost normal)
data distribution, equal variability of the outputs (the coefficient of variation (cv) ranged from 14 to 29% in the PWs and PPs versus 16 to 29% in the PTs), and excellent
predictability of the PW and PP from the PT (PTs accounted on an average 85% for the variation seen in the PWs and PPs). In addition, in the regression analyses the standard errors of the estimates (SEEs) were low (< 10%) and the residuals were distributed nonsystematically. In the PTAE measurements, the results were much more inconsistent, especially during the slow speed of the dynamometer. Compared with PT, PW and PP, the PTAE data distribution differed more frequently from normal distribution and the PTAE outputs showed higher variability. In addition, the PTAE outputs could not be acceptably predicted from the PT. In conclusion, the isokinetic PW and PP measurements can be recommended for clinical use, while the PTAE measurements should not be used routinely. Key words
Isokinetics, variability, reproducibility, peak torque _____________________________________________
Int.JSportsMed. 13(1992)249—256 GeorgThiemeVerlagStuttgartNewYork
The development of commercially available isokinetic dynamometers has instigated considerable research on the 'in vivo' characteristics of human muscles. The force or torque generating capabilities of human muscles have been extensively reported in the literature (18, 31). In general, all current dynamometers measure either torque or moment of force at various angular positions and speeds.
The most frequently used isokinetic measurement in clinical and scientific work has been peak torque (PT, newton meters, Nm), referring to the single highest torque output of the joint produced by muscular contraction as the limb moves through the range of motion (4, 12, 15, 23, 27). PT has been shown to be an accurate and highly reproducible variable to measure (3, 19, 20, 22, 23, 33), and its use in isokinetic tests has been accepted in critical reviews (2, 28, 31). PT has become a "gold standard and reference point" in all isokinetic measurements. However, interfacing of microprocessors with isokinetic dynamometers has enabled the rapid quantification of many other, more specific muscle function parameters than PT including peak or total work, peak or average power, and peak torque acceleration energy (15, 23, 27). Peak work (PW, joules, J) is defined as the area under the torque versus angular displacement (time) curve during the best repetition (work = torque x distance) (5, 12, 23); peak power (PP, watts, W) as
the amount of work achieved during the best test repetition divided by the contraction time (5, 23); and peak torque accel-
eration energy (PTAE, joules, J) as the greatest amount of work performed in the first 125 milliseconds of a single torque
production in the test repetitions (5, 23). The latter has been said to be indicative of muscular "explosiveness" giving an estimate about the speed (rate) of the torque production (5).
Recently, the relationships between PT and total work, and PT and average power have been a subject of interest (13, 14, 16, 17, 21, 23). These studies have observed that the predictability of total work and average power from PT is generally good and acceptable. However, no research has evaluated the value and consistency of the outputs of the peak work (PW), peak power (PP) and peak torque acceleration energy (PTAE) measurements, or their relationship to PT. The aims of the present investigation were to characterize the normality and variability of the PW, PP and . PTAE data outputs with respect to PT normality and variabil-
Downloaded by: Universite de Sherbrooke. Copyrighted material.
Abstract
P. Kannus
250 mt. J. Sports Med. 13(1992) ity in healthy males and females, and to determine the relationships between PT and the above mentioned parameters in multiple, low (60 deg/s) and high (240 deg/s) speed contractions
of the hamstring and quadriceps muscles. The assumptions
were that if the PWs, PPs and PTAEs are as valuable as the PT measurements, 1) their variability (coefficient of variation = cv) should not be larger than PT variability; 2) men and wom-
en should show almost equal PW, PP and PTAE variation (equal cv values) and these parameters should be equally related to PT values (i. e., PW, PP and PTAE expressed as a percentage of PT should be equal between men and women); 3)
Prior to testing each subject underwent a 5-mm warm-up period of aerobic, low-resistance ergometer cycling.
The subjects were then fixed to the testing device with straps round the chest, pelvis, thigh and ankle. The ankle strap was placed just above the ankle. The axis of the knee was placed in line with the axis of rotation of the dynamometer. Subjects were not allowed to grasp the handles of the bench during testing.
the level of normality in the PW, PP and PTAE data distribution should be equal with that of the corresponding PTs;
Each limb was weighed before testing by the Cybex's automatic limb weighing system to correct for the gravitational effect on torque (GET). Each estimate of the GET was added to the quadriceps torque and subtracted from
and 4) the predictability of the PWs, PPs and PTAEs from the corresponding PTs should be fully acceptable.
the hamstring torque. The subjects were then instructed to perform five to ten submaximal (50%) repetitions of knee exten-
sion and flexion at the two isokinetic testing velocities (60 deg/s and 240 deg/s) to become familiar with the testing dev-
Methods Twenty healthy adults (10 male and 10 female aged 20—40 years) volunteered to participate in the study.
Thereafter, the isokinetic concentric muscle
tional physical activities, but none of them was engaged in regular strength training or competitive sports.
performance test was started. First (at 60 deg/s) two submaximal and two maximal repetitions of knee extension and flexion were performed for practice, and then, after a 30-second rest, five maximal voluntary repetitions of alternating knee extension and flexion occurred. Subjects received consistent verbal encouragement throughout all tests. Special attention was
Table I
paid to the instruction to maximally exert each contraction over the full range of motion; that is, from the very beginning
Characteristics of the male and female subjects are reported in
Table 1. They had healthy knees, used no medication, and were moderately active. They participated in various recrea-
Characteristics of the study subjects (mean±SD).
Age (years) Height (cm)
Weight (kg BMI (kg/rn )
Men
Women
(n=10)
(n=l0)
28.9±4.7 174.0±8.0
30.0±6.1 167.5±4.5 60.0±8.3 21 .3±2.2
75.4± 10.8
25.1±3.0
BMJ = Body Mass Index = weight/height2.
Before the study the subjects were fully informed of the study procedure and design, purposes, and any known risks and gave their informed consent. The study was conducted in conformity with the principles of the Declaration of Helsinki of 1975.
Study Design, Instrumentation and Testing Procedure The study design was to first isokinetically assess the muscle performance of the knee extensors and flex-
ors of both limbs and analyze the normality, variability and predictability of PW, PP and PTAE with respect to PT. After eight weeks, the same test protocol was repeated in order to de-
termine the consistency of the results achieved in the first phase.
For the PT, PW, PP and PTAE measurements of the quadriceps and hamstrings of both legs, Cybex 340 dynamometer was used (Division of Lumex, Inc., Ronkonkoma, NY, 11779). In a random order of the subjects, the limb to be tested first was the dominant or the nondominant limb. The limb dominance was determined according to the limb preference in kicking a ball and taking-off in jumping.
to the very end of the movement.
After a two-minute rest, the same protocol was
repeated at the speed of 240 deg/s. After a five-minute rest (walking and recovery stretching) the same protocol was performed on the other leg.
Measured Variables The measured variables were PT, PW, PP and PTAE. Their definitions and units are reported in the "Introduction". Since the peak value of the torque production is the generally accepted reference parameter in isokinetics, the peak values of the other parameters (work, power and torque acceleration energy) were recorded and used for the analysis, too.
Data Analyses The test and retest data were analysed with the same statistical procedures by an IBM compatible 286 microcomputer, using the 1990 version of the statistical program library of the Biomedical Data Processing (BMDP) Software Inc., Los Angeles, California (8). The measured variables were all interval scaled by nature, and, therefore, the results are re-
ported as mean standard deviation (SD) with 95% confidence interval throughout the study.
The normality of the data distribution within each parameter was tested by the Shapiro and Wilk's W test (29). The coefficient of variation (cv SD/mean x 100%) was used to evaluate the data variability. For quadriceps and hamstrings of both groups, the Pearson product moment correlation coefficient (r) and its square (r2, the coefficient of determination) were calculated between the PT and PW, PT and
Downloaded by: Universite de Sherbrooke. Copyrighted material.
ice.
Subjects
mt. J. Sports Med. 13(1992) 251
Normality, Variability and Predictability of Work, Power and Torque Acceleration Energy
Table 2 Hamstring and quadriceps peak torque (PT), peak work (PW), peak power (PP) and peak torque acceleration energy (PTAE) at two isokinetic velocities (60 and 240 deg/s).
Men*
mean±SD
Women*
W"
p
W"
p
(78—92) (142—166)
.90 .97
.05 .67
mean±SD (95% confidence interval)
(95% confidence interval) PT (Nm)
60 deg/s Hamstrings Quadriceps
240 deg/s Hamstrings Quadriceps
140±36 227±43
(123—157) (206—247)
.93 .89
.15 .03
85±15 154±25
89±26 135±27
( 76—101)
.96 .96
.46 .60
54±10 85±14
(49—58) (78—92)
.91 .97
.05 .82
(53—61) (87—100)
.94 .97
.23 .79
(32—39) (58—62)
.94 .95
.30 .36
(59—68) (99—113)
.97
.72 .84
(122—147)
60 deg/s Hamstrings Quadriceps
97±24
( 86—108) (130—155)
.97 .96
.79 .46
57±9
143±27
240 deg/s Hamstrings Quadriceps
63± 18
(54—71) (82—98)
.97 .95
.72 .34
36±8
90±17
60 deg/s Hamstrings Quadriceps
103±23 153±27
( 92—1 14) (140—166)
.99 .95
.97
64± 10
.41
106± 15
240 deg/s Hamstrings Quadriceps
238±65 346±64
(207—268) (316—376)
.97 .97
.65
138±28
224±38
(125—151) (206—241)
.94 .96
.21
.68
4.5±0.9
(4.0—4.9) (4.5—5.8)
.81
.001
2.8±0.9
.08
4.6 1.6
(2.3—32)
.91
.86 .94
.007 .24
.91
.08 .45
11.4±2.6 16.9±3.9
.91
.06 .97
93± 14
58±10
PP (W)
.97
.59
PTAE (J)
60 deg/s Hamstrings Quadriceps
5.2
1.5
240 deg/s Hamstrings Quadriceps
19.5±5.1 24.6±5.4
(17.0—21.9) (22.0—27.1)
.95
(3.9—5.3)
(10.1—12.6) (15.0—18.7)
.99
* Both legs of the ten men and ten women were tested giving 20 measurements for both groups. * *w statistics tests the normality of the distribution (29). W-value .90 with p .05 means that the distribution of the data does not significantly differ from normal distribution.
for the study (beta level or type II error 0.10). Since there were not significant differences between the results of the right and
PP, and PT and PTAE. The residual error or each correlation was analyzed by a linear regression technique. In the regression equations, the PT values were placed on the x-axis and the
left limbs in any of the measured parameter, the data were
PW, PP and PTAE values on the y-axis, respectively.
pooled and analyzed together.
The given significance levels refer to two-tailed
tests. In all tests an alpha level (type I error) less than 0.1 % (p
