Gait and dynamic balance in adults with spina bifida

Background: Spina bifida (SB) is a complex congenital malformation, often causing impaired gait performance depending on the level and extent of malformation. Research regarding gait and balance performance in adults with SB, has not been sufficiently described yet. Research question: What are the characteristics of spatiotemporal gait parameters and balance performance in adults with SB? Further, do persons with muscle function (MF) level 3 differ regarding gait and balance performance from those with MF level 1 – 2? Methods: Cross-sectional observational study at an outpatient clinic. 41 adults with

Background: Spina bifida (SB) is a complex congenital malformation, often causing impaired gait performance depending on the level and extent of malformation.Research regarding gait and balance performance in adults with SB, has not been sufficiently described yet.Research question: What are the characteristics of spatiotemporal gait parameters and balance performance in adults with SB? Further, do persons with muscle function (MF) level 3 differ regarding gait and balance performance from those with MF level 1-2?Methods: Cross-sectional observational study at an outpatient clinic.41 adults with SB (18-65 years), who walked regularly.Spatiotemporal parameters of gait was assessed with the APDM system and balance performance with the Mini Balance Evaluation Systems Test (Mini-BESTest).Muscle strength in the legs was assessed with 0-5 manual muscle test, and participants were classified according to level of MF into groups MF1, MF2, and MF3.Two-sided t-test was used for parametric independent variables, and Cohen's d was used for effect sizes.The Mann-Whitney U test was used for non-parametric independent data and effect size was calculated by the z value (r = z/√n).Results: Mean gait speed was 0.96 (SD 0.20) m/s and mean stride length 1.08 m (SD 0.17), individuals with MF3 showed significantly slower gaitspeed and shorter stride length (p < 0.05).Lumbar rotation was 21 • (SD 11), and thoracic lateral sway 15 • (IQR 15) with significantley difference (p < 0.001 and p < 0.05) for individuals in MF3.Mini-BESTest showed a mean score of 11.3 (SD 6.9), and individuals with MF3 showed significantly lower scores (p ≤ 0.001).Significance: Gait and balance performance was reduced compared to normative data in almost all parameters, especially in persons with less muscle function.Increased knowledge from advanced gait analysis may help healthcare professionals to design rehabilitation programmes, in order to achieve and maintain a sustainable gait and balance performance.

Background
Spina bifida (SB) is a complex congenital malformation often involving vertebrae, the spinal cord, and the brain.Contractures, hip dislocation, scoliosis, and kyphosis are frequent as well as hydrocephalus and bladder and bowel dysfunction [1].The level and extent of spinal involvement define the muscular and sensory function affecting both ambulatory function and balance performance [2,3].Gait characteristics have been described for children and adolescents with SB [2,[4][5][6][7] but there are few studies characterising gait in adults [8].
Likewise, balance performance, a prerequisite for independent walking, has not been sufficiently characterised in adults with SB.
Complex gait abnormalities with compensatory mechanisms are common, involving multiple joints and anatomic structures [6], and leading to a characteristic movement pattern [5].Typically, children with SB walk with increased rotation and thoracic lateral sway, increased pelvic movements and flexed knees in the stance phase reflecting an absence of plantar flexion [5,7,9].Hip abductor weakness is the factor most contributing to abnormal gait characteristics in children with SB [4].Gait analysis has mainly been performed in laboratory settings with three-dimensional motion analysis systems with force plates, rendering a large amount of data for detailed analysis but with limited ecological validity [10][11][12].Today, wearable systems with sensors offer new opportunities to assess gait and balance performance in daily life settings [12][13][14][15].
Gait and balance performance are closely linked since independent gait requires a dynamic balance and might be even more challenging for adults with SB [16][17][18][19].Despite similar muscular prerequisites, adults with level 3 of muscle function (MF) (i.e.MF3, defined as reduced knee flexion, and only traces of hip extension, hip abduction, and below-knee muscles [3]) have been associated with an unstable gait and a large variation of mode of mobility [19,20].
The aim of this study was to describe gait and balance performance in adults with SB, with focus on spatiotemporal gait parameters.A second aim was to characterize and compare gait parameters and balance performance for participants in different levels of muscular function [3].

Methods
A cross-sectional design was used, data collection commenced in February 2020 and concluded in December 2020.

Participants
A near-total regional cohort of adults with SB enrolled at a specialised outpatient clinic for individuals with spinal cord disorders in Stockholm, Sweden, included in a previous study by Bendt et al. [17], were screened for inclusion.Also, participants enrolled at the outpatient clinic after the previous study were screened for inclusion at the time of their regular follow-up.Inclusion criteria were adults with SB (age 18-65 years), community or household ambulators as defined by Hoffer et al. [21] and a minimum level of MF corresponding to MF3 [3] and a minimum of 30 m of walk distance.Exclusion criteria were other diagnoses substantially affecting gait or balance performance and participants without muscular impairments in lower limbs.
Seventy-five individuals met the inclusion criteria, 13 did not reply, 19 declined and thereby 43 agreed to participate.Two individuals were excluded due to too weak hip flexion and knee extension.In total 41 individuals were included.

Procedure
Data were collected through instrumented gait analysis, clinical assessment of balance, motor and sensory function, contractures and measurement of height and weight, together with a semi-structured interview.As this is a small and rare patient group, there are few validated instruments available.The Hoffer classification [21] and the levels of MF [3] have been developed for persons with SB.Whilesome instruments have been validated for persons with spinal cord injury [22][23][24], a diagnosis with similarneurologic prerequites for walking or persons with other neurological diagnoses or for older individuals [14,25,26].
Pain was registered according to the International Spinal Cord Society (ISCOS) pain questionnaire [23].Ambulatory function was registered according to the criteria by Hoffer, a scale designed for persons with SB [21,27] and self-reported regular walking distance was categorised as > 1000 m, > 100 m or ≤ 100 m.Fear of falling was investigated with a single question "In general, are you afraid of falling?" answer options yes or no [28].
Falls during the last year were registered and concerns about falling were assessed with the Falls Efficacy Scale International (FES-I) [25,29] previously used for persons with spinal cord injury [30], ranging from 16 indicating "no concerns about falling", to 64 indicating "severe concerns about falling".Physical activity in daily life was reported in accordance with Frändin & Grimby [26], where level 1 indicates"Hardly any physical activity", level 2 "Mostly sitting, sometimes a walk, light gardening, or similar tasks, sometimes light household activities such as heating up food, dusting, or clearing away"", level 3  Level of neurological impairment was assessed according to the International Standards for Neurological Classification of Spinal Cord Injury [22].Muscle strength in the lower extremities was assessed using a 0-5 graded manual muscle test [31], and participants were classified according to level of muscle function (MF) into groups MF1, MF2, and MF3 [32].MF1 with weakness in foot intrinsic muscles and plantar flexors grade 4-5, MF2 with foot plantar flexion grade 3, knee flexion grade 3, hip extension and/or hip abduction grade 2-3 and MF3 with hip flexion and knee extension grade 4-5, knee flexion grade 3, and only traces of hip extension, hip abduction, and below-knee muscles.In case of asymmetry the most severe side was used for classification [3].Contractures were measured with a goniometer and registered if > 20 degrees in the hip, knee, or ankle joints were registered together with those with an arthrodesis in ankles [33,34].
Temporal and spatial parameters of gait were assessed with the Mobility Lab™ using six Opal™ wearable sensors (APDM wearable technologies Inc. [15,35]), and steady-state gait was analyzed with APDM's Gait Track algorithm [15].The Mobility Lab is a portable gait and balance laboratory providing a valid system for gait data collection [14], which has previously been used for children with SB and cerebral palsy [36], adults with parkinsons disease and multiple sclerosis [37,38] and older people [13].The sensors were placed at: sternum, waist (at the level of the fifth lumbar vertebra), bilaterally on dorsal wrists and bilaterally on dorsal metatarsus via elastic straps.Participants were instructed to "Stand as still as possible.When you hear a tone, start walking at a natural and comfortable pace".
Participants walked along a straight corridor for 12 m back and forth for 30 s and used their regular walking shoes (not high heels or sandals), and/or orthoses/walking aids as used in their daily life.
This was repeated three times and rest was offered between the walking trials.The mean number of analysed gait cycles was 44 (SD 11) with a range from 11 to 65.For one person Mobility Lab was not able to calculate gait characteristics.Variables of left and right side were averaged to a single value for each trial.In case of muscular asymmetry, the most affected side was used (n = 9).An average of all three trials was used in the statistical analysis.For two individuals one trial deviated > 0.15 m/s from the others, these two trials were excluded from the analysis.Defintions of spatial and temporal gait characteristics by Mobility lab are presented in supplementary Table 1.The normative data are based of the Mobility Lab on an average of a pool of 70 subjects, age 20 -90 years, with 5 men and 5 women per decade, see Table 2.The normative range is reported as the 5th-95th percentile to provide a representation of the actual range [39].
The Mini Balance Evaluation Systems Test (Mini-BESTest) was used to assess dynamic balance [40].It is composed of 14 items (max score of 28) divided into four subsections: Anticipatory postural adjustments, Postural responses, Sensory orientation, and Dynamic gait.The instrument has not been tested for validity and reliability in adults with SB but for persons with spinal cord injury (SCI) who have similar neurological prerequisites [24] and for people with multiple sclerosis [41] with similar neurologic dysfunction and balance related problems.The Timed Up and Go (TUG) test was used to assess functional mobility [42].

Statistical analyses
Analyses were performed using SPSS version 28 (IBM Corp., Armonk, NY, USA).Descriptive data were presented as numbers and proportions.To analyse normal distribution kurtosis and skewness values, Shapiro-Wilks values, visual inspection of QQ-plots and histograms were used.Mean and standard deviation (SD) was used for normally distributed variables, and median with interquartile range (IQR) was used for nonnormally distributed variables.The one-way anova was used to analyse differences in age and BMI between the MF groups.To examine associations between gait speed and age, dynamic balance and physical activity Spearman rank correlation analysis was used and classified according to Dancey and Reidy [43].Normally distributed independent variables were analysed with two-sided t-tests, and Cohen's d was used for effect sizes and presented with point estimate.The Mann-Whitney U test was used for non-normally distributedindependent data, and the effect size was calculated by using the z value (r = z/√n).Statistical significance was determined to p ≤ 0.05.

Results
Forty-one individuals were included, mean age 37 years (range 19-57 years), 20 (49 %) women.The participants had a level of MF from MF1 to MF3, with the MF3 group being the largest with 26 (64 %) individuals, there were no difference in age between the three levels of MF groups (p = 0.5877).In total 36 (88 %) participants had myelomeningocele (3/3 in MF1, 9/12 in MF2 and 24/26 in MF3) and 5 (12 %) had lipomeningocele (3/9 in MF2 and 2/24 in MF3).According to the International Standards for Neurological Classification of Spinal Cord Injury [22], level of neurological impairment was L3 for 32 participants (79 %).There were 27 (66 %) community ambulators and 14 (34 %) household ambulators.Their median level of physical activity according to Frändin and Grimby [26] was 2, corresponding to: "Light physical activity such as easy household tasks: heating up food, dusting, straightening up, sometimes a walk, or easy gardening".Orthoses were used by 19 participants (46 %), of whom 13 persons used ankle-foot orthoses, and three used walking aids (canes or crutches).Seventeen participants (41 %) had perceived pain during the last week, in most cases located in the lower extremities and load related.Table 1 presents demographic data, neurologic level of spinal malformation, level of MF, walk distance, physical activity, and pain for the total cohort.Twenty-six (63 %) participants had fallen during the last 12 months and 18 (44 %) were afraid of falling with no significant differences between individuals in the three MF levels (p 0.85252).Concerns about falling were assessed with FES-I, mean for the cohort was 27.0 (SD 9.3) with no significant difference between MF levels (p[TS8201 0.577577).
Mean gait speed was 0.96 (SD 0.20) m/s and mean stride length was 1.08 (SD 0.17) m.Cadence was 107 steps per minute (SD 11) and during the gait cycle participants spent a mean of 37 % (SD 2.9) in the swing phase and 63 % (SD 2.9) in stance phase.Mean lateral step variability was 4.8 (SD 1.6) cm.The lumbar rotation was 21 degrees (SD 11) and the thoracic lateral sway in the frontal plane was 15 degrees (IQR 15).In Table 2 spatial and temporal gait characteristics are presented with mean and SD for the total cohort and for participants divided into levels of MF1 to 3. In line with this Fig. 1 illustrates gait characteristics (mean and SD) for gait speed (m/s), double support phase in percentage of gait cycle time (GCT %), lumbar rotation (degrees), and thoracic lateral sway (degrees) for participants divided into MF 1-3.In supplementary Table 3b, spatial and temporal gait characteristics (presented with mean and SD) are presented for those with a muscular asymmetry in the legs (n = 9).
Mini-BESTest showed a mean score of 11.3 (SD 6.9) out of a maximum of 28.The participants had difficulties with all subsections (Anticipatory postural adjustments mean 2.1 of max 6, Postural responses mean 2.7 of max 6, Sensory orientation mean 3.1 of max 6, and Dynamic gait mean 4.4 of max 10), see Fig. 2. Individuals in the MF3 group showed significantly lower scores on the Mini-BESTest mean 8.9 (SD 5.9) compared to others (MF3 vs. MF1 merged with MF2 p = 0.001, MF3 vs. MF2 p = 0.004).In the subsections Anticipatory postural adjustments, Postural responses, and Sensory orientation the individuals with MF3 performed lowerthan the those with MF1 and MF2 (p < 001, A weak correlation [43] was seen between gait speed and age (r s − 0.33, p = 0.039) while there was a moderate correlation between gait speed and dynamic balance (Mini-BESTest) (r s =.59, p = <.001)Further, there was no correlation between gait speed and level of physical activity (r s =.26, p = 0.106) [43].

Discussion
This is the first study characterising spatial and temporal gait parameters as well as dynamic balance in adults with SB.Our main finding is that gait and dynamic balance are highly affected, illustrating a gait pattern outside the normative range in more than half of the analysed gait parameters.The gait was characterised by slower speed, shorter stride length, longer stance phase and a larger lateral step variability.However, most distinctive was the large thoracic lateral sway as well as a large lumbar rotation.We also demonstrated that dynamic balance in adults with SB was considerably affected and extensive, affecting all subsections, and even more pronounced in those with a lower level of muscle function (i.e MF3).
Previous reports regarding gait characteristics on individuals with SB describe gait in children [4][5][6][7]9] and the few studies on gait in adults have focused on ambulation in relation to neurologic impairment [19,20].Only one study of individuals with SB has investigated gait with movement analysis and the focus was adults over 50 years [8].
We found that adults with SB walked slower and with shorter strides than the normative data [11].A study of adults with SCI [44], with similar neurologic impairments, also showed a slower gait speed and

Table 2
Spatial and temporal gait characteristics (presented with mean and SD) for the total cohort, for participants divided into levels of muscular function (MF) and with MF1 and 2 together.Levels of significance and effect size values are shown for participants in MF3 versus participants in MF1 and 2 merged as there was only three participants in MF1 and for participants in MF3 versus MF2.Normative data by Mobility Lab is also presented.Parameters in bold are those deviating from the 5th-95th percentile by the normative data of Mobility Lab and p values in Italian shows significant differences.A The normative data of the Mobility Lab system is based on an average of a pool of 70 subjects, age 20 -90 years, with 5 men and 5 women per decade.
shorter strides.We found shorter swing phase and longer stance phase, also seen as longer double support phase.A longer time spent with both feet on the ground and less swing phase has been associated with imbalance and muscle weakness in elderly adults [45].In line with findings from ambulatory persons with SCI [30], two out of three (64 %) in our study reported falls during the last year and 44 % were afraid of falling.History of falls, fear of falling and gait speed below 1.0 m/s have all been shown to be strong predictors for falls in persons with SCI and older adults [30,46].A mean gait speed of 0.98 m/s for the total cohort and 0.92 m/s for individuals in MF3 implies a higher risk of falling in persons with SB, especially in MF3.Also, a slower walking speed is suggested to result in lower gait quality [47].The group's mean total score on the FES-I was 27 out of 64, indicating a moderate to high concern about falling [25,29], with no difference between MF groups.In a previous study on adults with SB above 50 years old [8], the participants had a similar gait pattern and corresponding concerns about falling (as seen in this cohort).This can possibly indicate that this cohort (mean age 37 years) is more affected as they at a younger age have a similar gait pattern and concerns about falling.
In this study, adults with SB showed a larger thoracic movement  compared to the normative data [39] especially lateral sway but also thoracic tilt (forward/backward), in addition to a larger lumbar rotation.Similar excessive trunk movement has been described in children with SB [5,7] and has been explained by weakness of hip abductors [7].
The lateral sway of the trunk is needed to move the center of mass over the hip joint in the stance phase so that the other leg can lift in swing phase [5].As expected, participants with least MF (i.e MF3) had a more deviating gait pattern, in the lower limbs as well as in the lumbar-and thoracic movements, also seen in children with SB [5,9] they also walked significantly slower and showed a significantly larger lateral trunk sway than other participants.This cohort had three times larger lateral trunk movements as compared to young and old adults walking at the same speed [48].Increased trunk movement has also been seen in persons with stroke [49].In young healthy persons trunk sway increases with faster walking speed [48], however this pattern has not been seen in older adults with presence of gait abnormality where the relationship between trunk sway and gait speed differs [50].
The Mini-BESTest showed that dynamic balance performance was substantially affected and individuals with MF3 displayed low results in two of the four subsections (Postural responses and Sensory orientation with 46 % respectively 35 % scoring 0).A possible explanation can be the large impact of reduced or absent sensory and MF below the knees, giving little or no sensory feedback regarding the position of the feet.Absence of hip abduction-and hip extension muscles [4] also makes it demanding to change direction and hold position in both static and dynamic conditions.In the previous study including people with SCI [30] the median Mini-BESTest total score was 20 with a range of 0-28 i. e., indicating a larger proportion of individuals with less affected dynamic balance compared to the present SB cohort, also the present cohort had a considerably larger amount of individuals with AIS A-C compared to the study by Jørgensen et al. [24] (AIS A 66 % respective 7 %).Also, we found a moderate correlation between gait speed and dynamic balance, those who walked slower also had a lower dynamic balance, showing prerequisites for eachother.

Strengths and limitations
Objective measurements of gait are often expensive, time-consuming and complicated to use requiring well-equipped laboratories and skilled technicians [11].However, a sensor-based system is flexible and user-friendly in a clinical setting.For the participant, the test situation is more like a daily walk, as the sensors are small and attached outside the clothes, and allows a gait pattern with higher ecological validity.This is further emphasised as the participants used the orthoses (n = 19) and/or canes/crutches (n = 3) as normally used in daily life.The use of orthoses and/or assistive devices inevitably affect gait characteristics, although this impact is outside the scope of this study.Instead, we were interested in their walking performance in everyday life with the aids they normally use.Also, this system assesses gait over numerous gait cycles (in this study an average of 44), making it possible to analyse stride to stride variability implying a normal gait pattern.
The cohort comprises a large regional cohort of adults with SB, assumed to represent more than 90 % of adults with SB in the Greater Stockholm area [17,19] and estimated to represent around 25 % of the national cohort.Nevertheless, the participants were recruited from only one clinic, and it is a rather small number of participants, especially when divided into MF groups, which affected the power of the sub-analyses, which is a limitation of this study.To reduce this effect, we merged the MF1 and MF2 group.However, the MF distribution was comparable to the earlier studies [17,19].The normative data of the Mobility Lab [38] had a larger age span (20-90 years) than this cohort (19-59 years) and the present cohort was most likely also shorter in height (mean 162 cm), potentially affecting the gait analysis.

Clinical implications
It is important for healthcare professionals to characterise gait pattern and balance performance of adults with SB to find the most suitable advice, treatment, and assistive devices and thereby improve quality of life and reduce falls and potential injuries.
A sensor-based system is user-friendly and a clinically applicable movement analysis systems that enables analysis of gait parameters, and gives direct feedback to the professionals as well as the patients to a relatively low cost.
This study provides a deeper insight into gait characteristics, balance, and stability in adults with SB which may contribute to the design of clinical rehabilitation programs and measurements used at followups.Gait is largely affected as almost all gait parameters are affected.This is important to consider when performing two tasks at the same time for example walk while carring a cup of coffe or walk and text at the same time.A narrow sidewalk can also be challenging as the area is limited.It is also important for clinicians to take in consideration that balance control is largely affected in in this population and many subsystemes influencing balance is affected.This has most likely an impact on everyday life.
In a study including only adults with SB over 50 years [8] the participants had perceived deterioration in gait function within their fourth decade.We found a weak correlation between age and gait speed however, our study excluded indivuduals above 65 years and had therefore a limited age-perspective.In a previous study by our group on adults with SB [19] we found that individuals who had survived to an age over 46 had less complex medical conditions and better physical functions than those under 46 years.This indicates a more complex relation between age and gait speed.As well as age-related changes are seen in adults with SB, those that reach a higher age might also be the ones with milder injury from the start.The age aspect will likely play a different role in the future as the population living with SB today is getting older and surviving longer.That will also likely reguire more tailored continious interventions regarding rehabilitation, orthoses and assistive devices.
Unfortunately, the system used only analysed the forward and backward motion of the arms since these are typically involved in walking.Instead, we observed that the participants had an excessive arm movement in the frontal plane.Arm movements are important for gait and adults with SB seem to assist their balance during gait leading to a "flying" position of the arms.We have not been able to find other papers describing these distinctive arm movements although the importance of the arms for balance has been reported earlier [51].Objero et al. [52] also concluded that their data indicated arm movements are vital for the control of mediolateral sway when balance is challenged, as in persons with SB.There is a lack of knowledge regarding the interplay between upper and lower limbs, and their interaction with thoracic movement.Therefore, further research is needed to better understand the entire body movement during gait in adults with SB.Also, a next level of analysis is how gait and balance are affected in a dual task situation.

Conclusion
Gait in adults with SB was characterised as slower with shorter strides, larger step width variability and a longer double support phase than the corresponding normative data.However, most distinctive was the large lumbar rotation and the large thoracic lateral sway.Balance performance was reduced within all balance subdomains.Participants with a lower level of muscle function had larger deviations in both gait characteristics and balance performance.These results can contribute to deeper knowledge regarding gait in adults with SB and help healthcare professionals to improve rehabilitation.

Fig. 1 .
Fig. 1.Illustrates gait characteristics (mean and SD) for (A) gait speed (m/s), (B) double support in percentage of gait cycle time (GCT %), (C) lumbar rotation (degrees), and (D) thoracic lateral sway (degrees) for participants divided into muscle function (MF) levels 1-3.Grey fields illustrate the normative data, 5th-95th percentile (based on an average of a pool of 70 subjects, age 20 -90 years, with 5 men and 5 women per decade).

Fig. 2 .
Fig. 2. Illustrates the mean and standard deviation for (A) total the Mini-BESTest score and (B) the median, range and IQR (box plots) of the Mini-BESTest sub-scores divided into the muscle function (MF) levels 1-3.

Table 1
Demographic data, neurologic level of spinal malformation, level of muscular function, walk distance, physical activity, and pain for the total cohort.