Participants
This study was open to healthy children of both sexes between 5 and 17 years of age. Exclusion criteria included muscle disease, treatment possibly influencing the neuromuscular system, practicing high-level sport (more than 5 hours a week in sports clubs) and occurrence of any illness or injury during the preceding month. Subjects were recruited from relatives of the hospital personnel, relatives of patient families, and advertisements displayed in hospitals and in the publications of various patient associations. Informed consent forms were signed by children and parents. This study received the approval of the Local Ethical Committee (CPP Ile-de-France IV).
Auxological assessment
Height was measured to the nearest millimetre using a standard height gauge (SECA 216 Height Rod) and was also expressed in standard deviations (SD) with respect to French population references. Weight was measured to the nearest 0.1 kg using electronic scales (Tanita TBF-543).
Strength measurement
Strength was assessed bilaterally for five muscle functions: handgrip, elbow flexion and extension, and knee flexion and extension. A QMT system was used for the measurements. This system was designed to measure force production in isometric conditions. It included a wall-mounted frame, a load cell, straps for attaching the load cell to the frame and the subject, a grip dynamometer, an examination table and a computer for feedback and force recording (details can be found at http://www.qmasystem.com). The subject was placed in standardized positions on the examination table and the examiner provided appropriate stabilizations during the efforts to avoid artefactual or compensatory movements.
Elbow flexion and extension strength were assessed in the supine position with the elbow at 90° flexion at the side of the trunk, the forearm in a neutral pro-supination position. The evaluator stabilized the subject’s upper limb by holding the anterior part of the shoulder with one hand and the lateral condyle of the elbow with the other hand.
Knee flexion and extension strengths were assessed in the sitting position, with hip and knee at 90° flexion. A flat cushion was installed below the distal part of the working thigh to ensure that the segment was horizontal. For knee extension, the evaluator placed one hand on the lateral part of the subject’s knee and the other hand on the proximal part of the thigh to prevent hip rotation and extension as compensatory movements. For knee flexion, the evaluator maintained the knee with both hands placed on the anterior distal part of the thigh.
For knee and elbow flexion and extension, the lever arm length was measured at each visit to compute the maximal torque produced around the joints. The strap was placed distally on the leg or the forearm segment with the distal edge of the strap at the level of the malleoli or the styloids, respectively. This length was measured as the distance between the rotation axis of the joint and the middle of the strap, to the nearest half-cm using a flexible measuring tape. The rotation axis of the knee was considered at the middle of the lateral part of the femoro-tibial interline, while the rotation axis of the elbow was taken at the epicondyle level.
Handgrip strength was measured while the subject was seated, the elbow at 90° flexion along the side of the trunk. The height of the examination table was adjusted so that the feet were flat on the floor with hips and knees each at a 90° angle. The contralateral hand of the subject was placed on the thigh. The grip handle width was adjusted to hand size. The evaluator supported the subject’s forearm and the device.
The test order was always the same: handgrip right and left, knee extension right and left, knee flexion right and left, elbow flexion right and left, and elbow extension right and left. The measurements were recorded by dedicated software (QMA computer software package).
Trials were carried out with verbal encouragement asking the subjects to provide maximal voluntary isometric contractions (MVIC) during about three seconds with one minute rest between trials. For each muscle function tested, if the difference between the first two measurements was below 10% of the higher value, that higher value was recorded. If not, measurements were repeated until two trials gave values with a difference lower than 10% of the higher value. The maximal value of two reproducible trials was recorded as the MVIC of the function. The force curves were visually checked to ensure that no overshoot or artefact was present.
Particular attention was given to making the subjects feel confident so as to help them to produce their true MVIC. Explanations were adapted to the maturity of the subject and were repeated until the child seemed to understand perfectly what was required of him/her. All measurements were performed by three examiners trained to the same operating procedures (standardized positioning procedures, lever arm measurements, verbal instructions, curve validation and reading…). Reliability between evaluators for both QMT and lever arm measurements was ensured before the study by a preliminary training demonstrating similar results obtained by the different evaluators on the same subjects. Statistical analyses were performed without adjustment on evaluators.
Assessment of learning effects
To ensure that true MVIC was measured, a second measurement session was conducted. This second session allowed to test for learning effects, possibly arising from difficulties in understanding or shyness at first visit; it also allowed assessment of the reliability of the method. The second session took place between two days and three months after the first one.
Data and statistical analysis
Quantitative variables are reported as medians (range), and qualitative variables as frequencies (percentages). The relationship between the various muscle functions was tested using Pearson’s correlation coefficient. Normality was assessed for all variables using the Kolmogorov-Smirnov statistics.
Knee and elbow flexion and extension are expressed as torque in newton.meters (N.m), and handgrip strength is expressed in newtons (N). The dominant hand side was defined as that with which the children wrote.
Paired t-tests were performed for each muscle function measured to detect any learning effect. Reliability was assessed by means of Bland-Altman plots and calculating intraclass correlation coefficients (ICC). Bland-Altman plots represent the differences between the strength values measured during the test and retest sessions against the means of these values. It shows the amount of disagreement between the two measures (via the differences) and how these differences are distributed. ICC2,1 was computed as a single-measure ICC with a two-way random effects model (absolute agreement). We considered coefficients of 0–0.20 as 'slight', 0.21-0.40 'fair', 0.41-0.60 'moderate', 0.61-0.80 'substantial' and >0.80 'almost perfect'. Standard errors of measurement (SEM) were computed as the SD of the differences between test and retest values divided by the square root of two. The SEM is a measure of absolute reliability and is expressed in the actual units. Relative SEM (%) was computed as absolute SEM divided by the mean of the measure.
Reference intervals for muscle functions were estimated by multiple linear regression. Age, height and sex were considered for model building. The SD was estimated as the standard deviation of the residual of the measurement of interest from regression on all parameters. The model fit was assessed by calculating the standard deviation scores (Z-score) as Z = (measurement – mean)/SD. The ordered Z-scores were plotted to provide a graphical check of normality using QQ-plot. The absence of heteroscedasticity was also checked by plotting Z-scores against height and age.
SPSS 15 (SPSS Inc., Chicago, IL) and SAS v9.2 (SAS Institute Inc, Cary, NC, USA) softwares were used for statistical analyses. The limit of statistical significance was set at an alpha risk of error of 0.05.