Pediatric patients are often referred to physical therapy as a result of deficits in muscle strength that interfere with their ability to achieve motor milestones or perform daily activities. Muscle strength is integral to the performance of functional activities. The ability to assess strength accurately is essential to quantify the severity of impairments. Determination of gains and deficits in muscle strength is paramount in both examining the effectiveness of intervention programs and the development of clinical treatment strategies. While the need for objective measurements of muscle strength in children is recognized, few normative data are available.1–5 In spite of advances in clinical knowledge, the results of strength tests remain difficult to interpret because of the lack of a reference population or normal strength values for comparison.5–10 Muscle strength is typically measured in children using manual muscle testing, a hand-held dynamometer, or an isokinetic dynamometer.2,3,5,11–14
While all three methods have advantages and disadvantages, they are limited in their application by the lack of consensus on a standardized testing protocol or sufficient normative data with which to compare results. Specifically, standardized manual muscle testing (developed on an adult population) relies on the skill and experience of the tester to distinguish between grades (ie, 4 to 5), is not sensitive to small changes in force, and has low reliability in grade ranges below 3.12,15 The hand-held dynamometer has shown higher reliability than manual muscle testing for grades below 3 but has high variability for grades above 3.11 Isokinetic dynamometry is currently regarded as the most clinically valid tool for muscle function testing and is frequently used in research.5,6,8,10,16–18 The primary disadvantages of isokinetic dynamometry cited in the literature are the expense and the lack of a standardized pediatric protocol test making reproducibility of results difficult.1,9,16,19 Merlini et al6 reported the reliability of the isokinetic dynamometer on children six to eight years old. Molnar and Alexander8 conducted a study in 1974 using a Cybex isokinetic dynamometer testing 500 children between the ages of seven and 15 years on quadriceps and hamstrings and biceps and triceps muscles. These researchers concluded that the isokinetic device was useful for muscle strength determination in children and that development of normal muscle strength values would require a large study population of at least 100 children of each gender at each age.
Molnar and Alexander8 indicate that anthropometric variables, particularly height and weight, can affect isokinetic dynamometry measurements of muscle strength. In their study, age and height accounted for more than 50% of the variance in isokinetic strength in normal children between seven and 15 years of age. Researchers also cite training,20 socioeconomic conditions,21 maturation level, limb dominance, activity level, gender, leg length, and growth hormone levels,8,17,18,21–24 as other variables that can affect isokinetic strength. Hormone levels and testosterone show great impacts on strength as prepubertal boys and girls generated similar levels of peak torque while postpubertal differences showed boys to be significantly higher than girls in strength. 10,17,18,20,24–26
The purposes of this study were to (1) develop a standardized isokinetic testing protocol for quadriceps and hamstring muscle groups in children who are typically developing and six through 13 years of age; (2) establish percentile charts and of normal isokinetic muscle strength for quadriceps and hamstrings by gender and age in children in this age range who are typically developing; (3) determine the best predictors of muscle strength or peak torque for children ages six through 13 years of age. These norms should provide a standard by which to compare children who are developing atypically and children who have sustained injuries.
A pilot study was conducted prior to the start of data collection to develop and test a protocol for isokinetic testing of quadriceps and hamstrings muscles in children six through 13 years of age. Seven testers participated in the study, and all were practicing physical therapists with three to 25 years of experience. All therapists were required to take a training course given by Biodex on isokinetic testing and the mechanical setup of the Biodex for testing. Each tester had prior experience performing isokinetic testing of quadriceps and hamstring muscles using not only the Biodex (Shirley, NY), but also other isokinetic devises, ie, Cybex (Phoenix Healthcare Products, Nottingham, UK) and Kin Com (Isokinetic International, Harrison, TN). The testing protocol was developed by therapist's at Texas Children's Hospital and followed closely the Biodex adult protocol.27 Variations were made to the adult protocol specifically for testing children. These variations were based on the experience of these physical therapists testing children and included:
- (1) A warm-up of two minutes at a self-paced walk or on a treadmill to assist the children in preparing their muscles for the test,
- (2) Three submaximal trials were given prior to each testing velocity to assist with familiarization to the testing. In the literature, a familiarization to procedure prior to test is recommended, although its validity has not been confirmed.16
- (3) Three testing velocities were used including 60, 120, and 180 degrees per second with a 90-second rest period between each velocity. These velocities were chosen initially to assess peak torques at slow, medium, and fast speeds. Since a preferred testing velocity in children was not reported in the literature, these velocities were chosen based on the testers' experience. In our experience, the younger children (younger than 10 years of age) were unable to consistently generate torque at speeds greater than 180 degrees per second. At 30 degrees per second, the younger children also frequently became discouraged and refusal to complete repetitions was encountered on a regular basis.
- (4) Ten maximal effort repetitions at each speed were chosen at random to determine where fatigue became a limiting factor and whether there was a fairly consistent point where peak torque occurred. Through the pilot study, we found that peak torque occurred throughout the repetitions and we did not find fatigue to be a factor during 10 repetitions.
- (5) Verbal encouragement and the Biodex computer visual display were provided during testing to ensure maximum effort at each velocity and repetition.
- (6) Testing always started with the dominant leg for the knee extensors.
- (7) Simple instructions were provided to help the children understand the task and alleviate fears. These instructions were given to each child when placed on the machine and included “We want you to push and pull as hard and fast as you can 10 times. We will cheer for you and you can watch the screen to see how hard you are pushing. The first 10 are the hardest and then as they get faster, they will be easier. We will do three practice pushes and then will start when the green light comes on the computer.”
Interrater reliability was established by having each therapist test the same subject over a two-week period using the set protocol. The subject was also tested on both systems to establish test reliability between systems II and III. All the testing therapists' measurements were within 10% of each other, which is within the standard error of measurement according to the manufacturer. No significant differences in results were found between the two systems or between the individual testers.
To standardize the testing protocol, a pilot study was undertaken on 88 boys and girls aged six through 13 years at the Texas Children's Sugar Land Health Center. Only therapists who had completed a class on isokinetic testing on Biodex and had established testing reliability using the protocol participated in the pilot study. Adjustments to the protocol were made based on minor problems encountered during a pretest trial of the protocol. These changes consisted of (1) changing the position of the thigh on the chair to two centimeters from knee bend instead of four centimeters (four centimeters was used on adult testing; we found it was difficult to keep the children from sliding forward during maximal pushing when the thigh was not maximally supported on the chair); and (2) the rest period was changed from 90 to 45 seconds. Longer resting periods tended to result in boredom and lack of maximum effort on subsequent testing and shorter resting periods of 30 seconds tended to result in fatigue in younger children. Forty-five seconds was found to be the optimal resting period for children especially at the younger ages.
An Excel spreadsheet was also developed during the pilot study to record peak torque, time to peak torque, and calculate coefficient of variance, and agonist/antagonist ratio at all three speeds for quadriceps and hamstrings, and to record the additional demographic and anthropometric data.
METHODS AND MATERIALS
To determine the number of children required to achieve our primary aim of establishing normal peak torque values in quadriceps and hamstrings in children six through 13 years old, a statistician from Baylor College of Medicine performed a power analysis. Assuming approximate normality, the 95th percentile is estimated as the mean plus 1.645 standard deviation (SD) and the 90th percentile as mean plus 1.281 SD. The distance between these two percentiles is 0.36 SD and half this distance is 0.18 SD. The approximate half width or a 90% confidence interval on the 95th percentile is given by:
Our aim was to determine the sample size that makes the half width ≤0.18 in order to guarantee that the 95th percentile is clearly distinguishable from the 90th percentile. Setting the half width equal to 0.18 and solving for n yielded 196 or approximately 200 subjects. A calculation using a more complicated method that required the lower limit of the 95th percentile confidence interval to equal the upper limit of the 90th percentile confidence interval yielded n = 180. The total number of boys and girls at each age are listed in Table 1. A breakdown of ethnic groups by age is presented in Table 2. In all, 3587 children were included in this study: 1557 males and 2030 females. Children were recruited from school districts in the greater Houston area including Lamar Consolidated Independent School District, Katy Independent School District, Fort Bend Independent School District, and Holy Rosary Catholic School. Institutional review board approval was obtained through Baylor College of Medicine and through individual school districts based on their policies on outside research.
Recruitment was done through each school's physical education (PE) department. A cover letter, institutional review board–approved permission slip, and a brief student information sheet were sent home with each child for parents to complete. These materials were handed out during the child's PE classes at least one week prior to the scheduled testing week. Exclusion criteria included the following:
- Surgery to lower extremity within six months or treatment for injury to lower extremity within three months that did not receive physician clearance for full activity.
- Heart disease that precluded the child from full participation in all activities
- Myocarditis or pericarditis
- Systemic or pulmonary embolus
- Acute infections
- Chronic infectious disease, eg, mononucleosis, hepatitis, acquired immunodeficiency syndrome
- Uncontrolled diabetes, thyrotoxicosis, or myxedema
- Advanced or complicated pregnancy
A mobile testing unit consisting of a Biodex System II Isokinetic System, metric height stick, metric weight scale, and a mat were taken to each school. Children were tested primarily during their PE class times to avoid disruption in other class schedules. Only those children returning a permission slip were tested. Age was verified by reported birth date.
On the day of testing, the mobile testing unit was set up by the door to the gymnasium. Each child with a signed permission slip and completed information sheet was escorted to the mobile testing unit. Upon arrival at the testing unit the following measurements were taken and recorded on the child's information sheet:
- Height was measured using a metric height stick attached to the wall. Children were asked to remove their shoes, stand up straight, and look straight ahead.
- Weight was obtained using a digital scale and recorded in kilograms. Scales were calibrated according to manufacturer directions and zeroed before each weighing.
- To establish laterality or preferred leg, children were asked to kick a ball centered in front of them. The leg used for kicking was recorded as the preferred leg.
- Leg length was measured with the child lying supine. Each child was asked to do a bridge to ensure that the pelvis was level and a measurement in centimeters of the distance from most prominent point of the anterior superior iliac spine to most prominent point of medial malleolus of the dominant leg was obtained and recorded. Although the reliability of these measures was not examined, this method is commonly used in clinics.
- For children older than 10 years, maturation status was assessed using self-assessed indices by Tanner. This is a rating system for the development of secondary sex characteristics (eg, genital pubic hair).26 Each stage is rated on a five-point scale with stage 1, the prepubertal phase; stage 2, initial development of hair; stages 3 and 4, increasing development of hair; and stage 5, most mature level. Each child was taken aside individually and asked to rate his or her development by pointing to the picture that looked the most like their stage of development.
Once measurements were completed, the child's data were entered into the computer under a coded number. Identifying children only by a subject number ensured confidentiality of all testing results. In addition to the measurement data, the following information was also recorded in each child's profile: date of birth, gender, activity level, and any medical information reported by parents. The child's activity level was rated by parents as (1) school PE only; (2) organized sports such as soccer, baseball, hockey; or (3) unstructured activities in which child engaged more than six hours per week (eg, roller blading, street basketball, bicycling, jogging, swimming).
- A hard cushion setting of 1 was used on System III and Adult Sensitivity level “C” on System II. For children six to seven years of age or who weighed less than 50 pounds, an ankle sensitivity setting of “E” was used. These cushion settings are the default for the Biodex system.
- To decrease extraneous spikes, the window-on option was checked for a presetting of 70% to eliminate any data that were less than 70% of the preset velocity. This is also a default for the Biodex system.
- The back angle of the seat was set in the most upright position, which measured 85 degrees from the horizontal with a goniometer.
- The hip angle was measured at 100 degrees from the vertical using a goniometer.
- The range of motion was set at 90 degrees of the total range prior to the child's getting on the machine. While peak torque is generated at about 30 and 60 degrees of flexion, at faster velocities, peak torque occurs later. Using a 90-degree arc of motion allowed capturing peak torque at all speeds.
General directions were given to each child prior to placing them on the machine. All tests were performed in a seated position. The trunk, waist, and thigh of the tested leg were stabilized with Velcro straps to prevent any extraneous movement or substitutions that would affect the measurements. Children were asked to cross their arms on their chests during the testing. The resistance pad on the leg attachment was placed three centimeters superior to the most proximal point of the lateral malleolus. The axis of rotation of the dynamometer was aligned with the lateral epicondyle. If a child's overall leg length was less than 70 cm, the pediatric knee attachment was used so that a better alignment could be achieved. Also, a back pad was used to help with correct placement on the chair for smaller children. The testing protocol for each speed (60, 120, and 180 degrees per second) consisted of three submaximal practice repetitions followed by 10 maximal repetitions of knee extension and flexion in isokinetic concentric/concentric mode. A 45-second rest was given between each change in velocity, and dominant leg and extension were always tested first. The nondominant leg was tested with the same setup as the dominant leg. To ensure maximum effort on each trial, verbal encouragement and visual feedback on computer screen were provided throughout the entire cycle of testing.
Peak torque, time to peak torque, and agonist/antagonist ratio was recorded in an Excel spreadsheet along with the anthropometric measures. The coefficient of variance was also checked for each test to ensure that best effort was obtained. Those children with variances of greater than 20% on more than one test were excluded from the study.
Data were screened for incorrect data entry, and influence analysis was conducted to evaluate the possibility of outliers. Issues of multicollinearity were also examined. All possible two-way interactions between height, weight, activity level, laterality, gender, maturation, ethnicity, and age were considered. Since the interactions failed to add any additional value to our overall equation, they were not included in final analysis of data. The Excel spreadsheet was loaded into SPSS v. 12.0. The descriptive statistic package of SPSS was used to establish the percentile charts of strength by age and gender for the quadriceps and hamstring muscle groups at 60, 120, and 180 degrees per second on both the dominant and nondominant lower extremities. A stepwise regression procedure in SPSS was used to determine which anthropometric measures had the greatest influence on variability in peak torque production at each speed. The regression equation for all speeds and both muscle groups was developed.
The peak torque charts of isokinetic muscle strength are presented by age and gender from 6.0 through 13.11 years in the Appendix, in Tables 3 through 18. Values for the quadriceps of the dominant and nondominant legs and values for the hamstrings of the dominant and nondominant legs are reported for the slow (60 degrees per second), medium (120 degrees per second), and a fast (180 degrees per second) speed.
Peak torque results comparing males and females for the 50th percentile are also shown in Figures 1 and 2. Results are similar for males and females until age 12 years when gender differences appear to begin.
Results of the stepwise regression procedures are shown in Table 19 for the quadriceps and hamstring muscles. Height was the best overall predictor of peak torque production for both quadriceps and hamstrings at all speeds. Height alone accounted for 66% to 72% of the variance in isokinetic strength measurements. Weight followed as second best predictor for quadriceps and hamstrings at a speed of 180 degrees per second, but maturation was the second best predictor at 60 and 120 degrees per second for the hamstrings. Table 20 reports the regression coefficients. Note that the regression coefficients demonstrate that children who are taller, heavier, and older produced higher peak torques. The ethnicity vector 4 (African American = 1) was the only vector that added any additional information to the equation. The maturation vector 4 (level 4 = 1) added to the equation for quadriceps but not for hamstrings. Maturation level 5 was the only maturation level that did not take away from the equation for hamstrings. The regression analysis demonstrates that females have lower peak torques. Activity vector 2 (organized sports) added to the equation for both quadriceps and hamstrings. Laterality was generally excluded from the equation.
Isokinetic dynamometry allows objective assessment of muscle strength throughout the whole range of movement. In children, it is a safe and controlled method to match resistance applied by the subject throughout the range and has the added safe guard of no resistance occurring when movement stops. One of the criticisms of isokinetic testing of muscle strength in children is the lack of a standardized procedure or a normative database for comparison when testing children with impairments. This study established a normative isokinetic peak torque database for quadriceps and hamstrings in children six through 13 years of age. In order to promote comparable data in future studies, an isokinetic testing protocol was developed and standardized during a pilot study and included both isokinetic settings on the Biodex, setup guidelines, preferred speeds, direction guidelines, verbal encouragement, practice, and warm-up prior to testing.
Information on other variables including leg length, height, weight, laterality, maturation level, and activity level was also gathered and used to assess the variability seen in normal muscle strength among and within the age groups. Height, weight, and maturation level have been identified in the literature as accounting for a large percentage of the variations seen in strength in children. However, almost 40% of the variance had not previously been accounted for.19 In our study, we were able to account for 75% to 80% of the variance when all our variables were entered. We attempted to control as many of the variables as we could by including the directions, compensation for a learning effect, and motivators in the protocol.
One limitation of the study is that we did not consider gravity correction in data analysis. The literature suggests that, at least in adults, error levels in isokinetic measurements occur when gravity is uncorrected. Since we could not find any literature on using gravity correction in children, we opted to not correct. According to Jones and Stratton,16 correction using adult procedures is thought to overestimate gravitational torque in children because they do not account for the elastic components of the growing muscle-joint system. Future research should examine the data we obtained without correction and compare them to the gravity correction when a validated procedure for this correction with children is developed.
This study was conducted using Biodex System II and Biodex System III Isokinetic systems. They were chosen and purchased for our clinics because they have a pediatric knee attachment that is not available with other models. This attachment assists in correct positioning and alignment of smaller children on the machine.
These data have wide reaching implications since, for the first time, normal values for a large sample of children are now available for isokinetic peak torque of the quadriceps and hamstrings muscles. These data provide a benchmark for comparison with special populations to determine the presence and severity of muscle weakness, muscle strength profiles, and how special populations might vary from normative populations. Furthermore, the effectiveness of treatments for developing muscle strength and functional changes can also be assessed objectively against this normative database. We can now begin to assess how much strength is necessary to complete various functional activities and design and implement more effective treatment strategies to reach these goals. Since we also have a large database of anthropometric measures, we can begin to assess the effect of these variables both individually and in combination on muscle strength at different stages in development.
Contributors to the study: Faith Anderson, BS, PT, Texas Children's Sugar Land Health Center, Sugar Land, TX; Stacy Bucic, MS, PT, Racelli Caballes, BS, PT, Julie Good, MS, PT, Rachel Jackson, MS, PT, Carla Uria, BS, PT, Emay Yeng, MS, PT, Veronica Victorian, BS, and Ashley Barker, BS, Texas Children's Hospital Physical Medicine & Rehabilitation Department, Houston, TX.
We thank Jennifer Smart in the development office at Texas Children's Hospital for helping us find funds to support this study; Vonco Medical for providing us with their mobile testing unit for the two-year duration of the study; Biodex Isokinetic Systems for providing rental of the equipment for the last year of study; the Physical Medicine and Rehabilitation Department of Texas Children's Hospital for financial assistance, moral support, and training of staff; Fort Bend Independent School District, Katy Independent School District, Lamar Consolidated Independent School District, and Holy Rosary Catholic School for allowing us to test their students during school hours; and to all the children who so enthusiastically participated. We also thank our statistician, O'Brian Smith from Baylor College of Medicine for his support, guidance, and patience during the design of the study. We also thank Les Wiggin for maintaining the truck during the two-year testing period and transporting our generator to and from the testing sites each week.
1.Ayalon M, Ben-Sira D, Hutzler Y, et al. Reliability of isokinetic strength measurements in the knee in children with cerebral palsy. Dev Med Child Neurol
2.Griffin JW, McClure MH, Bertorini TE. Sequential isokinetic and manual muscle testing in patients with neuromuscular disease: A pilot study. Phys Ther
3.Kilmer DD, Abresch TR, Fowler WM. Serial manual muscle testing in Duchenne muscular dystrophy. Arch Phys Med Rehabil
4.Mognoni P, Narici MV, Lorenzelli F. Isokinetic torques and kicking: Maximal ball velocity in young soccer players. J Sports Med Phys Fitness
5.National Isometric Muscle Strength Database Consortium. Muscular weakness assessment: Use of normal isometric strength data. Arch Phys Med Rehabil
6.Merlini L, Dell'Acci D, Granata C. Reliability of dynamic strength knee muscle testing in children. J Sports Phys Ther
7.Molnar G, Alexander J. Muscular strength in children: Preliminary report on objective standards. Arch Phys Med Rehabil
8.Molnar G, Alexander J. Development of quantitative standards for muscle strength in children. Arch Phys Med Rehabil
9.Molnar G, Alexander J, Gutfeld N. Reliability of quantitative strength measurements in children. Arch Phys Med Rehabil
10.Sunnegardh J, Bratteby LE, Nordesjo LO, et al. Isometric and isokinetic muscle strength, anthropometry and physical activity in 8 and 13 year old Swedish children. Eur J Appl Physiol.
11.Florence JM, et al. Intrarater reliability of manual muscle test (Medical Research Council scale) grades in Duchenne's muscular dystrophy. Phys Ther
12.Deones VL, Wiley SC, Worrell T. Assessment of quadriceps muscle performance by a hand-held dynamometer and an isokinetic dynamometer. J Sports Phys Ther
13.Miller LC, Michael AF, Baxter Tl, et al. Quantitative muscle testing in childhood dermatomyositis. Arch Phys Med Rehabil
14.Wikholm JB, Bohannon RW. Hand held dynamometer measurements: Tester strength makes a difference. J Sports Phys Ther
15.Hyde SA, Goodard CM. The myometer: The development of a clinical tool. Physiotherapy
16.Jones MA, Stratton G. Muscle function assessment in children. Acta Paediatr
17.Burnie J, Brodie DA. Isokinetic measurement in preadolescent males. Int J Sports Med
18.Van den Berg-Emons RJG, van Baak ME, De Barbanson DC, et al. Reliability of tests to determine peak aerobic power, anaerobic power and isokinetic muscle strength in children with spastic cerebral palsy. Dev Med Child Neurol
19.Kellis S, Kellis E, Manou V, et al. Prediction of knee extensor and flexor isokinetic strength in young male soccer players. J Orthop Sports Phys Ther
20.Guy JA, Micheli LJ. Strength training in children and adolescents. J Am Acad Orthop Surg
21.Henneberg M, Brush G, Harrison GA. Growth of specific muscle strength between 6 and 18 years in contrasting socioeconomic conditions. Am J Phys Anthropol
22.Cuneo RC et al. Growth hormone treatment in growth hormone deficient adults: Effects on muscle mass and strength. Am Physiol Soc
23.Kubo K, Kanehisa H, Kawakami Y, et al. Growth changes in the elastic properties of human tendon structures. Int J Sports Med
24.Rodda C, Jones DA, Round J, et al. Muscle strength in girls with congenital adrenal hyperplasia. Acta Paediatr Scand
25.Seger JY, Thorstensson A. Muscle strength and myoelectric activity in prepubertal and adult males and females. Eur J Appl Physiol.
26.Faulkner R. Maturation. In: Docherty D, ed. Measurements in Pediatric Exercise Science.
Champaign, IL: Human Kinetics; 1996:129–158.
27.Isokinetic Source Book. Shirley, NY: Biodex Medical Systems, 1999.