The association between gait analysis measures associated with crouch gait,functional health status and daily activity levels in cerebral palsy

1. Introduction

Cerebral palsy (CP) is the most common cause of motor impairment in children [1]. The associated muscle tightness, weakness and impaired motor control often lead to difficulties during gait. There are numerous gait presentations associated with CP; however, crouch gait, defined as excessive knee flexion in stance phase, is one of the most common pathological patterns associated with bilateral CP [2]. This flexed knee posture during gait has been shown to increase the forces and demands on the knee joint [3, 4] and can potentially lead to joint degeneration and pain [5]. For these reasons, the correction of excessive knee flexion during gait is a significant focus of surgical intervention in those with CP. A number of surgical techniques have been shown to improve knee kinematics in crouch gait [6, 7, 8, 9]. It is recognised that the assessment of surgical outcomes should include outcome measures based on the International Classification of Functioning, Disability and Health (ICF) [10]. The impact of knee crouch, and its treatment on these domains is less clear. A recent long-term follow-up study by Boyer et al. [9] found that while distal femoral extension osteotomy (DFEO) and patellar tendon advancement (PTA) led to superior outcomes in terms of improving knee extension compared to alternative surgical intervention. This did not translate into better outcomes in terms of non-gait function, activity, participation, quality of life or knee pain.

The Pediatric Outcomes Data Collection Instrument (PODCI) is a measure of functional health status which has been found to be valid and reliable in children and adolescents with CP [11, 12]. The PODCI questionnaire yields scores in six domains: upper extremity function, transfers and basic mobility, sports and physical function, pain/comfort, happiness with physical condition and a global function score. A number of studies have reported improvement in the functional PODCI domains (sports and physical function, transfers and basic mobility and global function) following orthopaedic surgery in children with CP but no changes in the psycho-social domains (pain/comfort and happiness with physical condition) [13, 14, 15]. As none of these studies [13, 14, 15] reported on gait kinematics, it is not possible to ascertain if the reported post-operative improvements in PODCI domains were associated with improvement in specific gait kinematic variables post-operatively. In contrast, Stout 2008 [16] compared short-term outcomes in groups who had DFEO-only, DFEO+PTA and PTA-only and found that while stance-phase knee flexion was restored to normal range in the DFEO+PTA and PTA-only groups, none of the three groups demonstrated significant improvement in the transfers and basic mobility, sports and physical function or pain/comfort domains of the PODCI assessment. Thus, is appears that while orthopaedic intervention can potentially lead to improvements in the functional domains of the PODCI assessment. They do not seem to be related to changes in stance phase knee flexion.

Ambulatory children with CP have been shown to have significantly reduced levels of walking activity compared to typically developing peers [17]. On examining the relationship between gait kinematics and daily activity, Wilson et al. [18] found a moderate correlation ( r = 0.58) between daily step counts and an overall gait kinematic summary score (Gait Deviation Index (GDI)). Nicholson et al. [19] confirmed a moderate correlation between GDI and stride counts ( r = 0.49) in a cohort of participants with CP pre-operatively, although this correlation did not persist at surgical follow-up. However, neither study specifically examined the influence of knee kinematics. Thus, the potential relationship between crouch gait, as measured in the gait laboratory and activity levels is unclear.

The aim of this study was to examine the relationship between gait parameters related to knee crouch gait, functional health status and daily activity levels in children and adolescents ( 18 years old) with bilateral CP. Our hypothesis was that individuals with more severe crouch gait, as measured by knee flexion angle at mid-stance, would have lower measures of functional health status and reduced daily activity levels.

2. Method

3. Results

A total of 35 participants met the inclusion criteria for this study. All had knee flexion greater or equal to 19 ∘ at mid-stance in one or both limbs. Participant demographic data and gait, QoL and activity parameters are summarised in Table 1.

Sports and physical function, sedentary time/day, mean steps/day, knee flexion at initial contact and age were normally distributed. Transfers and basic mobility, pain/comfort, happiness with physical condition, global Function, knee flexion at mid-stance, Normalised speed and normalised step length were not normally distributed ( p ⩽ 0.01 in all cases).

The relevant correlation coefficients between gait, PODCI and daily activity variables and associated p-values are summarised in Figs 1 and 2. Analysis of variance inflation factors demonstrated no collinearity between any of the included parameters; values ranged from 1.18 to 5.82.

OPEN IN VIEWER

OPEN IN VIEWER

There were no significant correlations (Bonferroni adjusted p -value 0.008) between knee flexion at mid-stance and any of the PODCI or daily activity measures. In contrast, normalised step length significantly correlated with three PODCI domains:ports and physical function (Spearman ρ = 0.524; p = 0.001), transfers and basic mobility (Spearman ρ = 0.565; 0.001) and global function (Spearman ρ = 0.486; p = 0.003). Knee flexion at initial contact demonstrated a significant correlation with transfers and basic mobility (Spearman ρ = - 0.508; p = 0.002). There were no significant correlations between pain/comfort or happiness with physical condition and any of the gait variables. There was a significant correlation between mean steps/day and normalised gait speed (Spearman ρ = 0.561; p = 0.001). However, sedentary time/day was not associated with any of the gait variables. There was no significant correlation between season and either mean steps/day (Spearman ρ = 0.236; p = 0.173) or sedentary time/day (Spearman ρ = - 0.315; p = 0.065). Age was found to be associated with transfers and basic mobility (Spearman ρ = 0.545; p = 0.001) while GMFCS level was associated with sports and physical function (Spearman ρ = - 0.495; p = 0.003) and transfers and basic mobility (Spearman ρ = - 0.458; p = 0.006).

The separate multivariable linear regression analyses are summarised in Table 2. The analysis found that normalised step-length was independently associated with three PODCI domains (sports and physical function, transfers and basic mobility and global function). Normalised gait speed was significantly associated with steps/day.

OPEN IN VIEWER

Correlation coefficients between PODCI and daily activity variables and associated p -values are summarised in Fig. 3. There was a significant correlation (Bonferroni adjusted p -value 0.010) between mean steps/day and the sports and physical function domain of the PODCI assessment (Pearson r = 0.433; p = 0.009).

As crouch gait has been shown to increase the forces and demands on the knee joint [3, 4], our hypothesis was that, as a result, those with more severe crouch would have lower measures of functional health status and reduced daily activity levels. However, our analysis demonstrates that there was no correlation between knee flexion angle at mid-stance and any of the PODCI domains or daily activity measures analysed.

Instead, the current results suggest that step-length is better associated with functional health status rather than the more specific kinematic angle of knee flexion during stance. While knee flexion contributes to step-length, it has also been suggested that factors such as selective motor control also influence step-length [33]. Normalised step-length was found to be independently associated with the three functional PODCI domains (sports and physical function, transfers and basic mobility and global function). Knee flexion in stance was not related to functional health status or daily activity levels. It is consistent with the findings of Steele et al. [34] who recently reported that knee flexion in stance also demonstrated poor correlation with energy expenditure of gait. These authors further noted that oxygen consumption during gait better correlated with knee flexion angle at initial contact which the authors hypothesised may in turn lead to shorter step lengths. The fact that step-length is better associated with functional health status potentially has important implications in treatment planning in children with CP crouch gait. While gait analysis has proven its effectiveness and usefulness in surgical planning in those with CP [35, 36, 37, 38], the current study suggests that surgical interventions targeted at improving specific gait kinematics may not directly influence activity and participation. While correlation does not imply causation, our results suggest that to best improve functional health status, step-length should be considered. This possibly re-enforces the importance of appropriate orthoses where indicated as Ankle Foot Orthoses (AFOs) have been shown to improve spatiotemporal parameters such as speed and step length in addition to the beneficial effects of surgery [39, 40]. The current results suggest that further research is warranted about a potential relationship between improvements in the spatiotemporal parameters of gait with AFOs and potential changes in functional health status.

In terms of daily activity, we found moderate correlations between mean steps/day and both normalised gait speed and sports and physical function. It is probably not surprising that the more functional walkers in terms of speed were also the most active. There was no association between daily activity and any of the specific kinematic values, which differs somewhat from the findings of both Wilson et al. [18] and Nicholson et al. [19]. They both report a moderate correlation between step/stride count and an overall gait kinematic summary score (GDI). This may suggest that kinematic values other than those examined in this study have more of an influence on step counts and daily activity.

There are a number of limitations that should be considered when interpreting the results of this study. Our population size of 35 potentially limits the wider interpretation of the results. However, the sample size is broadly in line with, or larger than, similar studies on participants with CP [23, 41] and in all our regression analyses, the number of fitted variables is less than one-tenth of the number of observations as recommended [42]. This study specifically examined children and adolescents with ambulant bilateral CP and therefore, the generalizability of the results to older, more involved participants is somewhat limited. The majority of participants in this study were GMFCS I and II and only a limited number of GMFCS III participants were included as might be expected in a study of participants with ambulant CP. Accordingly, Fig. 1 highlights that the knee flexion at mid-stance values are somewhat skewed to the left with a smaller number of participants in more severe crouch greater than 40 0 . This is similarly the case for mean steps/day as plotted in Fig. 2 and reported in Table 1. There was a large range of recorded values with the relatively small number of participants who were GMFCS level III likely having significantly reduced step counts compared to less involved participants [29, 43]. Likewise, a ceiling effect is seen in some domains of the PODCI assessment, particularly pain/comfort, which again reflects the relatively high levels of function in this population of ambulant CP participants. The mean age of participants in this study was 9.82 years and all participants were under 18 years of age. Based on retrospective gait studies, it has long been suggested that crouch severity increases with age [2, 44]. While a recent prospective study has highlighted that this is not the case in all those with CP crouch gait [45], the findings of this study should not necessarily be extrapolated to older, more involved people with CP.

The gait kinematics analysed in this study were collected barefoot as per routine clinical analysis protocols and therefore may not fully reflect community ambulation in this population. As highlighted above, the spatiotemporal parameters of gait would be expected to improve with orthoses [39, 40]. Therefore, future work is warranted to examine step length during community-based assessment to examine their impact on daily activity and functional health status. Additionally, routine walking aids were not used during gait data collection and the five participants who routinely used a walking aid (3 posterior walker, 2 bilateral crutches) were instead assessed with hand-held assistance. Krautwurst et al. [46] have highlighted that gait data collected without a walking aid is likely to be more flexed than gait with routine walking aid. While the hand-held support provided may have mitigated this to some degree, it is still likely that gait data as assessed in the clinical gait laboratory is likely to have differed somewhat compared to routine day to day walking.

A control group, not walking in crouch, might have potentially strengthened the findings and theoretically, as functional PODCI domains and daily activity levels do not appear to be related to knee flexion, the non-crouch group would have demonstrated similar functional health status and daily activity levels. However, as crouch gait has been suggested to progress with age in children with CP [44], a non-crouch control group would potentially be younger which in turn influences both PODCI scores and daily activity levels [12, 29].

5. Conclusion

The results of this study confirm that functional gait variables are associated with the functional domains of the PODCI assessment (sports and physical function, transfers and basic mobility and global function), although not the psycho-social domains (pain/comfort and happiness with physical condition). Analysis of specific gait parameters suggest that it is step length rather than knee flexion in stance which is best associated with functional health status in children/adolescents with CP crouch gait. This highlights the importance of considering spatiotemporal measures of gait when assessing walking function in CP and examining the potential impact of footwear and orthoses on these parameters.