**4. Discussion**

This study found that ankle joint mobilization to improve ankle movements increased ankle ROM and gait function. Moreover, this study demonstrated that ankle joint mobilization is more effective for ankle ROM and gait than sham mobilization. This indicates the importance of ankle joint mobilization in orthopedic managemen<sup>t</sup> for ankle rehabilitation in children with CP. The strength of this study is that it is the first time in children with CP that joint mobilization was applied to ankle joints in a manual therapy technique. A direct comparison is difficult, but it is consistent with other studies showing that additional joint mobilization is more effective in increasing ankle ROM and gait than conventional rehabilitation [19,24].

When compared with studies regarding the change in passive stretching exercises applied to calf muscles on dorsiflexion ROM [18,44], this study confirmed the positive change effects of passive joint mobilization, such as improved dorsiflexion ROM, TUS, and 10-m walk test. Children with CP exhibit spasticity in calf muscles, so passive stretching has been applied to reduce the spasticity, relaxation, and elongation effect on muscles in previous studies [18,44,45].

However, spasticity exacerbates joint contracture and muscles weakness, as well as changes in the muscle contractile properties [46]. Ankle joint mobilization can be applied to reduce the spasticity of the soleus muscles and [33] restore ankle joint flexibility [19,32]. In addition, ankle joint mobilization causes articular reflexogenic effects, increasing dorsiflexor muscle strength [47]. It has been found that for stroke patients, joint mobilization is a way to increase a variety of ankle ranges of motion rather than stretching exercises [19]. Therefore, joint mobilization can be used as an intervention method to increase ankle mobility in children with CP.

The ankle and knee ROM of children with CP is highly correlated with the energy expenditure index, which means gait efficiency [15]. Ankle joint mobilization increased the ankle dorsiflexion ROM and speed of the sit-to-stand performance [32]. Among the variables measured in this study, the timed up and go test included sit-to-stand. The increased ankle ROM can increase the speed of the sit-to-stand performance, so the timed up and go test may be improved. In addition, ankle joint mobilization with movement can improve gait speed [21]. The posterior talar glide can be increased through joint mobilization [24,48]. Increased posterior talar glide increases dorsiflexion before heel-off and time to heel-off during gait movements [24]. Improving gait velocity through ankle joint mobilization can be

considered to a ffect the dorsiflexion increase during gait movements [19,21,24]. Consequently, ankle joint mobilization improved the gait speed by increasing dorsiflexion during gait movement.

Another finding of this study was that there was no change in the standing balance measurements after additional joint mobilization in children with CP. Ankle joint mobilization can reduce COP displacement by improving sensorimotor function and arthrokinematic restrictions [25]. For the elderly, ankle joint mobilization reduces the surface of standing COP excursions [49]. However, children with CP maintain postural control using a body sway rather than ankle strategy in a quiet standing position [8]. Because of poor postural control, children with CP typically have increased static COP displacement compared to developing children [16]. The wearing of hinged ankle–foot orthoses increased ankle strategy contribution but did not improve postural stability in quiet standing [10]. Therefore, it seems that there was no change in standing balance due to the contribution of body transverse rotation and hip strategies [8,16]. In this study, an NDT program consisted of improving trunk control. An NDT-based trunk protocol is beneficial in improving the trunk control and balance in children with spastic diplegic CP [37]. Therefore, improvement of trunk control by an NDT program decreased COP displacements. Since both groups performed the NDT program, the improvement of trunk control and balance by the NDT program may cause no significant di fference between the two groups in COP displacements.

Additionally, the COP displacements in this study evaluated the total COP trajectory. The wearing of hinged ankle–foot orthoses did not change the anterior and posterior displacements but increased the mediolateral displacements [16]. Therefore, to confirm the COP displacements for ankle dorsiflexion ROM increase, it is necessary for future studies to check the mediolateral displacements and anterior–posterior displacements changes, respectively.

Manual therapy applied to growing children can stimulate skeletal growth [30]. Fortunately, no serious or catastrophic adverse events have been reported for children [50,51]. However, due to insu fficient evidence so far with regard to manual therapy and adverse events, caution should still be exercised [30,50].

Spine manipulation reduced wrist muscle spasticity in children with CP [27]. Ankle joint mobilization also reduced ankle muscle spasticity in brain injury or incomplete spinal cord injury patients [33]. In this study, ankle joint mobilization, which is safer than spine manipulation, was applied. Joint mobilization can stop the treatments on its own whenever the patient wants to [28]. Therefore, we hope that joint mobilization is often used in clinical settings because ankle joint mobilization is a safe treatment method.

The limitations of this study are as follows. First, our study has a small sample size and it is thus di fficult to generalize, and also because our recruitment was limited to GMFCS level I or II. Secondly, our walking speed evaluation method is an evaluation frequently used in clinical settings, but ankle kinematic changes during gait and spatiotemporal analysis (motion capture or inertial measurement units) were not confirmed. To support our hypothesis, we need to further refine and systematically evaluate gait analysis. Finally, this study did not compare long-term e ffects. Complementing these limitations, future research should investigate the impact of a greater sample size, multiple assessments, and long-term follow-up.
