**4. Discussion**

Gait impairments are common in children and adolescents with CP. In spastic hemiplegic CP, decreased walking speed, increased stance phase on non-a ffected leg, as well as longer gait stride are observed. The stride is usually shorter on a ffected leg and cadence is increased to maintain gait speed [13]. Gait pattern can be impaired due to motor coordination, balance, and stability problems during walking. Additionally, there are significant di fferences in kinematic parameters of the hip, knee, and ankle joints, compared to healthy children [14]. For the purpose of the improvement of the dynamic gait e fficiency, di fferent types of AFO are recommended for patients with spastic hemiplegic CP [15].

Step length and gait speed are commonly used to assess patient's gait quality [16]. Hayek et al. [17] compared gait of the children with CP, while walking barefoot and with AFOs. In hemiplegic group, stride length increased by 11.7% with AFOs, in both a ffected (10.2%) and non-a ffected (12.4%) lower leg. The gait cadence decreased by 9.7%, but walking speed was not improved. Additionally, the authors of this study assessed the impact of using AFOs on the kinematic gait parameters. The ankle dorsiflexion at initial contact increased by 9.4◦ on the a ffected side and by 5.87◦ on the non-a ffected side, on average. The ankle dorsiflexion at swing was improved about 9.9◦, while the knee flexion at initial contact on the a ffected side was decreased by 8.5◦ [17]. In their study, Wren et al. compared the gait of children with CP (five patients with spastic diplegia and five with spastic hemiplegia), while walking barefoot and with two types of orthotic devices (dynamic ankle–foot orthosis—DAFO and adjustable dynamic response—ADR AFO). In both types of orthoses children improved stride length, hip extension, and dorsiflexion in swing phase. Using of ADR-AFOs produced better push-o ff power and knee extension, but the level of activity and patient's satisfaction were higher for DAFOs [18]. Liu et al. evaluated the changes of foot and ankle motion in long follow-up studies (18 months) in the group of 23 children with spastic CP (7 hemiplegic, 16 diplegic), during which patients were using solid ankle–foot orthoses (SAFOs), hinged ankle–foot orthoses (HAFOs), or supramalleolar orthoses (SMO). They observed that the long term use of AFOs can lead to maintaining or improving foot and ankle motion and function [15]. In our study, the step length was longer while the subjects were walking with AFO, comparing to walking barefoot, the gait speed has also improved significantly and the gait cadence was reduced. We have also observed improvement of kinematic parameters, including an increase of the ankle dorsiflexion in stance and in swing phase in both legs. Moreover, the knee flexion at the initial contact decreased significantly only in the a ffected leg. We did not observe significant changes at the knee flexion during the swing phase, but, both hip and knee flexion range of motion were improved.

In our study, we also evaluated whether the GDI and the GVI can reflect changes in spatio-temporal and kinematic parameters in spastic hemiplegic CP children wearing AFO. GDI and GVI indexes were developed to present how far an individual's gait di ffers from a normal gait pattern [7,9]. The GDI has been used to assess the gait of children with CP [5,10,16,19,20]. However, we did not find any scientific researches concerning using GVI in this group of patients. This fact encouraged us to check if GVI will be useful to assess changes in the gait of children with hemiplegic CP.

In our study, as it was expected, the value of both gait indices increased, along with the improvement of the selected spatio-temporal and kinematic gait parameters, when the children were wearing AFO. Bickley et al. evaluated two outcome measures: technical surgical goals system (TAG) and GDI and how they reflect changes in postoperative status in children with spastic CP. The TAGs goals have been developed at Shriners Hospital for Children in Huston in 1994 to reflect expected clinical examination and kinematic changes after the surgical treatment. In their study, they observed that using both GDI and TAGs system can improve postoperative assessment in patients with CP at di fferent level of Gross Motor Function Classification System (GMFCS) (I–III). They did not observe any significant di fferences concerning changes of the GDI results after surgical treatment between di fferent levels (I–III) of GMFCS [19]. Reis et al. showed, that wearing of AFO significantly improves step length, gait speed, and GDI in children with diplegic CP, but only changes in step length was clinically important [16]. A few studies show correlation between GDI and GMFCS. Malt et al. concluded

that GDI can be useful to evaluate and present walking impairments in children with spastic CP [6]. Molloy et al. sugges<sup>t</sup> that gait problems may be not well enough recognized when only GMFCS is used; while GDI, as a specific tool for gait assessment, can better reflect functional gait aspects and components [20].

On the other hand, Massaad et al. pointed out that the main clinical utility of GDI is the assessment of the global changes in patient's gait after the intervention, expressed as a single, numerical value, but without specific information about the source or nature of these observed changes [5]. Additionally, Domino et al. showed that the improvement in spatio-temporal and kinematic parameters of gait in children with diplegic CP did not correspond with changes in GDI, Gillette gait index (GGI), or gait profile score (GPS) [10]. In contrast to our study, which has shown that the changes in spatio-temporal parameters and gait kinematics were reflected in GDI and GVI results changes. Galli et al. evaluated the gait of hemiplegic and diplegic children with CP using other gait indices: GPS and gait variable score (GVS), which are also based on kinematic gait parameters. The children were assessed barefoot and while wearing AFO. They concluded that the gait indexes can be useful in evaluation of immediate effects of using the AFOs in hemiplegic and diplegic CP patients, but it was observed only in some GVS parameters—pelvic tilt and ankle dorsiflexion—but not in GPS [21].

The GVI, based on nine spatiotemporal gait parameters was developed and by Gouelle et al. in 31 patients with Friedreich's Ataxia. They showed significant decrease of GVI in patients with Friedreich's Ataxia (70.4 ± 7.9) compared to the healthy subjects (100.3 ± 8.6) [9]. Guzik et al. used GVI for the gait assessment in patients after ischemic stroke. In the control group of the healthy-matched subjects, the GVI scores were 98.34 ± 6.83 for the right lower limb and 96.3 ± 7.19 for the left lower limb and there were significantly di fferent, comparing to the patients after stroke (76.32 ± 7.98 a ffected leg, 80.74 ± 4.68 non-a ffected leg) [22]. In the validation study of GVI in patients after stroke in a chronic stage of recovery, the same authors showed that the GVI for the a ffected and una ffected leg were significantly correlated with the results of clinical functional assessment [23]. Balasubramanian et al. evaluated changes in spatio-temporal gait parameters in two groups of old people, younger and older adults (age < 65 and age ≥ 65, respectively) with GVI. They observed significantly reduced GVI in older adults (91.92 ± 8.75) compared to the younger adults (100.79 ± 7.99). Additionally, they pointed out that low level of functional mobility was correlated with lower value of GVI [24]. GVI score was also validated for patients with mild to moderate Parkinson's disease. The results showed that a mean, overall value of GVI was 97.5 ± 11.7 and mean GVI value for the more a ffected side was 94.5 ± 10.6 [8]. In our study GVI value while walking barefoot was 74.2 ± 9.48 for a ffected leg and 78.6 ± 7.67 for non-a ffected leg and increased significantly when the subjects were wearing AFO (83.1 ± 8.74 for affected leg, 86.5 ± 8.32 for non-a ffected leg).
