Participants

We used the freeware tool G*Power (http://www.gpower.hhu.de/) to calculate an a priori power analysis with the test family (t tests) and the respective statistical test (means: difference between two independent means [two groups]) based on a related study that examined between-group differences in running kinetics (i.e., GRF) in injured and uninjured runners [15]. The power analysis was computed with an assumed Type I error of 0.05, a Type II error rate of 0.20 (80% statistical power) and an effect size of 0.80 for active running kinetics. The analysis revealed that 36 children would be sufficient to observe large between group differences. Thirty-six boys aged 9–14 were enrolled in this study. Participants were divided in two groups (genu varus and healthy controls). The genu varus group included 18 boys that were selected during clinical examinations (Table 1). Only patients were enrolled that had a mechanical axis angle (MAA) which was defined as the angle formed by straight lines drawn from the centre of the hip to the centre of the knee and the centre of the knee to the centre of the ankle [16] greater than 1.3° in both knees. This was determined by means of a full-length standing anteroposterior radiograph [17] (Fig 1). Exclusion criteria were signs of joints diseases, diseases of bones, ligament injury, neuromuscular dysfunction, diseases of tendon, history of major trauma or surgery of the lower extremities, chronic joint infection, intra articular corticosteroid injection [5]. All participants were right lower-limb dominant as determined by a kicking ball test [18]. In addition, kinematic data was used to detect footfall patterns (heel strike pattern instead of mid foot or forefoot strike) during running test trials to match our participants in the two experimental groups not only for limb dominance but also for similar footfall characteristics. Table 1 illustrates group characteristics. It was previously postulated that shoe type affects walking kinetics in children [19,20] which is why all participants of the present study were equipped with individually fitted (i.e., size) athletic shoes (New Balance 759, USA) during running.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1.

The view of (A) mechanical axis angle (MAA), and (B) plugin gait marker set. Written informed consent (as outlined in the PLOS consent form) was obtained from the parents of the child in Fig 1. All participants provided their written informed consent/assent.

https://doi.org/10.1371/journal.pone.0185057.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Group-specific characteristics of our participants presented as means and standard deviations.

https://doi.org/10.1371/journal.pone.0185057.t001

All participants and their parents or legal representatives provided written informed consent/assent to participate in this study. This study was approved by the local ethics committee (University of Mohaghegh Ardabili, Iran) and all experiments were conducted in accordance with the latest version of the Helsinki Declaration.

Running analysis

This cross-sectional study included boys aged 9–14 with genu varus and healthy controls. Our participants were recruited from local physical therapy clinics. Three-dimensional running analysis was performed using a Vicon MX system consisting of six T- series cameras (200 Hz) (Vicon Motion Systems, Oxford, UK). Two Force plates (Kistler AG, Winterthur, Switzerland) embedded in the middle of an 18-m walkway simultaneously recorded three dimensional GRF data (Fx, Fy, Fz) at a sampling frequency of 1000 Hz. A plug-in-gait marker set [21] with sixteen reflective markers (14 mm diameter) was used to identify the pelvis, thighs, legs, and feet segments. Marker positioning was identified through palpation and in accordance with guidelines provided by the manufacturer. The markers were directly attached to the skin of the following anatomical landmarks: bilateral anterior and posterior superior iliac spines; bilateral lateral femoral epicondyles; bilateral lateral malleoli; bilateral calcaneus; lateral side of the thighs and shanks, and top of the feet at the base of the second metatarsal.

Thereafter, participants were familiarized with the laboratory environment and walkway area and at least five practice trials were performed to ensure subjects were able to strike the force plate without consciously changing their running cadence. Afterwards, each participant identified with heel strike pattern during running (kinematic analysis) performed five acceptable shod running trials at preferred and comfortable speed. A trial was discarded if both feet did not land on the force plates, if the participant targeted the platforms, lost balance during the trial, ran with mid or forefoot strike pattern, or even fell during running. Kinematic and kinetic data were extracted during the stance phase of running, defined as the interval from ground contact (onset of vertical GRF [Fz] > 10 N) to toe off (vertical GRF [Fz] < 10 N) [22,23]. Kinematic and kinetic data were filtered using a fourth-order low-pass Butterworth filter with a cutoff frequency of 10 and 20 Hz, respectively. Using spline interpolation, all data were normalized to the stance phase with heel contact to toe-off corresponding to 100% [24].

Parameters that were used for further statistical analyses included peak values in three dimensional GRF and time to peak of the mediolateral (Fx), anteroposterior (Fy), and vertical (Fz) GRF axes. The first and second peaks of the bimodal vertical GRF curve (impact peak [IP] and active peak [AP]) were considered for further analysis (Fig 2). In the mediolateral direction, the peaks of lateral force and medial force were also analyzed (Fig 2). In the anteroposterior direction, the braking peak (BP) and propulsion peak (PP) were analyzed (Fig 2). Forces were normalized to body weight (BW) and the corresponding timing was expressed as a percentage of stance-phase duration (%Stance), to allow between-subject comparison [25].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Time-normalized traces of the ground reaction forces for both groups during the stance phase of a running cycle.

Reaction forces (Fx, Fy, Fz) are demonstrated for the (a) dominant and (b) non-dominant lower limbs. “IP”, “AP”, “BP” and “PP” represent abbreviations that stand for “Impact Peak”, “Active Peak”, “Braking Peak” and “Propulsion Peak”, respectively. The solid lines represent the mean values, whereas the dashed lines indicate standard deviation values.

https://doi.org/10.1371/journal.pone.0185057.g002

Vertical loading rates were computed as the average slope from 20% to 80% of the vertical GRF (Fz) at the point of interest [26]. Impulses were calculated using the trapezoidal integration method (Eq 1) for x, y, and z axes as follows [27]: (Eq 1)

Where, F₁ and F_n are the first and last forces, respectively; Δt is equal to sampling duration and n is the number of force samples.

The FM was calculated by using the following equation [28]: (Eq 2)

Where M_z is the moment around the force plate vertical axis, CoP_x and CoP_y are the locations of the center of pressure (CoP) along the mediolateral and anteroposterior axes, respectively. All FM waveforms were amplitude-normalized to the product of each individual's body weight (N) and height (m) [28]. The first peak of the FM curve (negative; abductor moment) and the second peak (positive; adductor moment) were included in our analyses. Impulses of FM were computed as the net area under the FM curve during stance [29].

Statistical analyses

Data were tested for normal distribution using the Shapiro-Wilk test. Homogeneity of variance was assessed using the Levene’s test and variance ratios. Data were presented in the form of means and standard deviations if normal distribution of data was given. Between-group differences were tested for significance using the independent samples t-test. The significance level was set at p<0.01. In addition, effect sizes (Cohen’s d) were calculated as a ratio of mean difference divided by the mean standard deviation of both groups. Effect sizes were considered small if d < 0.5, medium if 0.5 ≤ d < 0.8 and large with d ≥ 0.8 [30]. The statistical calculations were carried out using the SPSS version 16 (Chicago, IL, USA).

Effects of genu varus on ground reaction forces

Our results are partly in accordance with the relevant literature with regards to the observed GRF values on the dominant and non-dominant limb during running at preferred speed [5,6]. In the present study, individuals with knee genu varus produced significantly higher body weight adjusted GRF values in the lateral direction (Fx) of the dominant limb and significantly higher body weight adjusted GRF values in the lateral and medial directions (Fx) of the non-dominant limb. Furthermore, there is a considerable variation both within and between groups in the pattern of medio-lateral GRF during the stance phase of running (see Fig 2).

Of note, a number of studies reported abnormal control of mediolateral motion in children and/or adolescents with cerebral palsy [31], hearing loss [32], down syndrome [33], and scoliosis [34]. It has been demonstrated that during walking at different velocities (i.e., 0.54; 0.75; 1.15; 1.56 m/s) the abductor, vasti, and plantarflexor muscles) make significantly larger contributions to mediolateral GRF compared to the contributions of passive dynamics (i.e., gravity and velocity-related forces) [35]. On average, the above reported muscles contributed 92% of the total mediolateral GRF over all examined walking speeds [35]. In this context, Liu et al. postulated that the abductors primarily support body weight by contributing a large medial GRF at all walking speeds, the vasti, gastrocnemius, and soleus influence both, forward progression and body weight support [36]. Also, it has been demonstrated that electromyographical activity of the biceps femoris muscle was significantly higher in boys with genu varus compared to healthy controls during the loading response phase of walking [14]. In the present study, we observed higher mediolateral GRF (Fx) in genu varus patients which is primarily caused by varus malalignment and to a lesser extent by muscular disorders [14]. In young individuals with varus malalignment, abnormally high knee internal rotation and hip external rotation moments were detected during walking which may increase the risk of sustaining knee injuries [6]. Apart from that, higher lateral (Fx) GRF were found in the present study in children with genu varus. It has previously been postulated for runners that high lateral GRF result in overpronation during running [37,38] which may again cause overuse syndromes of the leg and the knee joint [39]. If signs of abnormally high lateral GRF are observed in children, it is recommended to conduct preventive training programs (e.g., balance and/or strength training) that focuses on lower limb alignment and control mediolateral motion from an early age on [40]. Therefore, practitioners and therapists are advised to conduct balance and strength training to improve lower limb alignment and mediolateral control during dynamic movements in children with genu varus.

Effects of genu varus on vertical loading rates

Our results are partly in accordance with the relevant literature with regards to the observed vertical loading rate values on the dominant and non-dominant limb [41,42]. Findings from the present study demonstrated greater body weight adjusted vertical loading rates in both, the dominant and non-dominant lower limbs in children with genu varus compared to healthy controls. Higher vertical loading rates in children with genu varus are most likely due to a shorter time to reach peak GRF (Table 3 and Fig 2).

It has been shown that repetitive impact loading results in subchondral bone microdamage associated with cartilage thinning [43]. Also, the first peak of vertical GRF curve has high frequency domains that are propagated through the lower limb and modulated by the active and passive structure of the lower extremities, such as the muscles, tendons, ligaments, bone, and cartilage. Impairments to the active (muscles) and passive structures (such as ligaments and etc) of the body may be associated with dysfunction in impact attenuation process. This dysfunction results in forces across the articular surfaces of the lower limb joints (i.e., knee) to be increased beyond the tissues' load-bearing capacity [44], especially the cartilage regions that are unaccustomed to such forces [45,46]. It has been suggested that the vertical loading rate is a useful variable for estimating the overload applied on the lower limb musculoskeletal tissues [47,48]. An average vertical loading rate greater than 70, 72, and 100 N/kg/s during running in runners has been associated with a risk of sustaining stress fractures [49,50], patellofemoral pain [41,51], and plantar fasciitis [42]. In the present study, genu varus children demonstrated vertical loading rates that were greater than 80 N/kg/s. Therefore, it can be postulated that children with genu varus are at risk of sustaining stress fractures [49,50] and patellofemoral pain syndrome [41,51].

Effects of genu varus on free moment amplitudes

Our results are not in accordance with the literature in terms of the observed FM values on the dominant and non-dominant limbs [5,28,52]. Our study revealed no statistically significant between-group differences in body weight adjusted FM values during the stance phase of a running cycle. It has previously been demonstrated that knee varus malalignment results in larger endorotation of the foot and greater internal tibia rotation during the stance phase of walking [5]. Further, it has been demonstrated that runners with an injury history (tibial stress fracture and over pronation) showed higher free moment amplitudes compared to healthy (uninjured) runners [28,29,53]. This highlights the importance of assessing FM and thus biomechanical loading of the lower limbs in other populations. Nevertheless, our results did not demonstrate any significant differences in running FM peaks between genu varus group and healthy controls. However, as was illustrated in Fig 5, there was a considerable variation both within and between groups in the pattern of free moment during the stance phase of running. Moreover, as is typical in ensemble curves, the peaks are attenuated relative to the individual curves owing to between-group differences in TTP GRF values. The group average curves (Fig 5) illustrate the overall pattern of FM during stance phase of running, but as can be seen from the large spread demonstrated by the standard deviation in Table 4, there was considerable between subject data variability. The present study provides fundamental information regarding the characteristics of FM in healthy boys as well as boys with genu varus during the stance phase of running. As mentioned in a previous study [52], it seems that FM could be used to classify genu varus children into functional groups since it possesses a considerable inter-subject variability, a relationship to the mechanical demands put on the lower limb joints and potentially to injury risk. This is likely due to the longer support time during walking versus running.

A few limitations of this study warrant discussion. First, we assessed kinetic data only which is why it is not possible to deduce the underlying neuromuscular mechanisms responsible for the observed findings. Thus, further research is needed in this area. Second, we examined boys with genu varus versus age-matched healthy controls. Therefore, our findings should be carefully interpreted in girls and adolescents. More research is needed to confirm our findings for different population groups.