The Gait Deviations of Ankylosing Spondylitis with Hip Involvement

Clinical Rheumatology (2019) 38:1163–1175 https://doi.org/10.1007/s10067-018-4401-y

ORIGINAL ARTICLE

The gait deviations of ankylosing spondylitis with hip involvement

Guoning Zhang1 & Jia Li2 & Zhengliang Xia3 & Weidong Xu2

Received: 19 June 2018 /Revised: 25 November 2018 /Accepted: 10 December 2018 /Published online: 4 January 2019 # International League of Associations for Rheumatology (ILAR) 2019

Abstract Objective The aim of the study was to investigate the gait deviations of ankylosing spondylitis (AS) patients with hip involvement. Methods Thirty-six subjects, including 18 AS patients with hip involvement (AS group) and 18 healthy people (control subjects, CS group), were enrolled in the study. Three-dimensional gait analysis of the AS group and CS group was performed. Kinematic parameters, kinetic parameters and surface electromyography (sEMG) during the gait cycle were measured. Results The AS patients with hip involvement had a lower gait velocity, shorter step length and shorter stride length. In the hip angles, there was significantly decreased flexion, excessive abduction and excessive external rotation; there was excessive flexion in the knee and reduction in plantar flexion of the ankle. AS patients had increased forward trunk flexion, excessive obliquity and restricted rotation of the trunk during the gait cycle. The hip moments of the AS group showed a significant reduction in flexion, abduction and external rotation during the gait cycle. The root mean square amplitude of the sEMG for the rectus femoris in the AS group was higher than that in the CS group. Conclusion The gait deviations in AS patients with hip involvement were described in this study. The gait analysis results demonstrated statistically significant alterations regarding the kinematic and kinetic gait parameters for the patients included in the sample. Coordination and balance were impaired by the disease. An efficient physical exercise plan can be formulated according to the results of gait analysis.

Keywords Ankylosing spondylitis . Gait analysis . Kinematics . Kinetics

Introduction

Ankylosing spondylitis (AS) is a chronic, painful, degenera- Guoning Zhang, Jia Li and Zhengliang Xia contributed equally to this tive and inflammatory form of arthritis that affects the spine work. and sacroiliac joints and may cause fusion of the spine, resulting in complete trunk rigidity [1]. The rigid spine, from * Weidong Xu occiput to sacrum, causes patients to present with a stooped [email protected] position. Guoning Zhang More than 20% of patients with AS have hip involve- [email protected] ment and 47–90% of those patients experience hip in- Jia Li volvement bilaterally [2–4]. Hip pain was the first com- [email protected] plaint (opposed to spinal discomfort) in some patients Zhengliang Xia with AS, particularly in juvenile-onset patients. In these [email protected] patients, hip aggravation is rapid, and many patients re- quire total hip arthroplasty. Hip involvement with AS is 1 Department of Orthopedics, Tongren Hospital, Shanghai Jiao Tong often associated with severe functional impairment, work University School of Medicine, Shanghai, China disability and a compromised quality of life. The mecha- 2 Department of Orthopedics, Changhai Hospital Affiliated to the nism leading to hip involvement remains unknown. Second Military Medical University, No. 168 Changhai Road, Shanghai, China Gait analysis can provide information regarding kinematics, kinetics parameters and muscle activity to further delin- 3 School of Kinesiology, Shanghai University of Sport, Shanghai, China eate the relationship among joint disease, joint impairments 1164 Clin Rheumatol (2019) 38:1163–1175 and the compensatory gait strategies adopted to overcome criteria included the following: (1) unilateral involvement of painful and disabling deformities [5, 6]. the hip, (2) patients with infection, trauma, or other recent In previous studies, a trend toward gait velocity and stride surgery, (3) patients with concomitant rheumatoid immune length reduction was found in AS patients with spinal kypho- diseases, (4) age more than 70 years, (5) patients with neuro- sis [7]. Most patients demonstrated hyperextension of the hip logical or psychiatric diseases, (6) patients with severe visual joint and hyperflexion of the knee and ankle joints during or auditory impairments and (7) patients with a Cobb angle ambulation to maintain balance and avoid falls [8, 9]. These greater than 50° (to eliminate the disturbance of severe spinal tendencies indicate that patients with AS may employ a more kyphosis on gait). The control group consisted of normal sub- cautious gait pattern [10]; moreover, reduced gait velocity, jects enrolled from the population of hospital personnel. together with shorter stride length, could contribute to an increase in patient fatigue during ambulation [11, 12]. Clinical data collection The most prevalent characteristic of abnormal gait in patients with AS is spine ankylosis. Postural and gait The basic clinical data for all the patients were collected, changes can also be observed in peripheral joints, par- including gender, age, body mass index (BMI), age of ticularlyinthehip[13]. Although it is suggested that onset and disease duration. Related scores were tabulated, the lower extremity joints may contribute to the preser- including the Harris hip score [21], Bath Ankylosing vation of compensatory balance in the presence of Spondylitis Disease Activity Index (BASDAI) [22], Bath changes in the spinal column [14], it should be noted Ankylosing Spondylitis Functional Index (BASFI) [23], that arthritis and enthesis may occur in the lower ex- Bath Ankylosing Spondylitis Metrology Index (BASMI) tremity joints, and this may affect adaptation. Gait char- [24], Visual Analog Scale for Pain (VAS pain) [25]and acteristics can indicate the motor function of patients. the Berg balance scale [26]. A full-length spinal X-ray, Biomechanical characteristics, in terms of joint moment, which included the cervical vertebrae, thoracic vertebrae, muscle activity and range of lower limb motion, can be lumbar vertebrae and pelvic and bilateral hip joints, was investigated with gait analysis [15–17]. However, there taken for each patient. BASRI for the sacroiliac joint are few studies relating to gait characteristics in patients (BASRI-SIJ) [20] was graded by the X-ray. The basic with ankylosing spondylitis involving hip joint. The clinical data for the control group, including gender, age mechanism of AS with hip involvement is uncertain, and BMI, were collected. and how to prevent the exacerbation of hip dysfunction in AS is still unclear. Gait analysis Thus, the aim of the study was to perform analysis of gait deviations of AS patients with hip involvement. The gait test was performed across a distance of 10 m, as Biomechanical characteristics were investigated with three- determined by the length of the laboratory. Participants dimensional gait analysis. were instructed to ambulate at a comfortable pace, and gait cycles for each limb were recorded across three trials at our laboratory. Patients were allowed to rest; however, Materials and methods they were not given any physical assistance during the test. Kinematic data were acquired utilising a 10-camera, Subjects three-dimensional motion-capture system (Vicon, Oxford, UK, 100 Hz). This approach was combined with data Thirty-six subjects, including 18 AS patients with hip involve- derived from a ground reaction force dynamometer ment (AS group) and 18 healthy people (control subjects, CS (Kistler, Switzerland, 1000 Hz) to derive three- group), were enrolled in this study. The minimum sample dimensional kinetics variables. number per group was 17, as determined by GPower3.1 sam- A total of 51 reflective markers were placed on the sub- ple size calculation software [18]. Patients with AS were re- jects at the following anatomical land marks: head, trunk, cruited from the orthopaedic and rheumatism departments in arms, pelvis, thighs, shanks and feet (Fig. 1). Raw kinematic the Changhai Hospital between November 2017 and and kinetic data (C3D file format) were imported to Visual March 2018. Patients were required to meet the New York 3D software (C-Motion, Inc., USA) to filter and compute the criteria [19]. All the patients with AS received a full-length gait parameters. Marker trajectories were filtered using a spinal X-ray, which included the hip joint. The Bath 4th-order zero-lag low-pass Butterworth filter, with a cut- Ankylosing Spondylitis Radiology Hip Index (BASRI-h) off frequency of 10 Hz. The kinematic parameters, including [20] was used to evaluate the hip joint; the bilateral hips spatiotemporal gait parameters (gait velocity, cadence, step should be graded on a BASRI-h scale of 1–4 to aid in estab- length, stride length, stride width) and joint angles, were lishing the AS diagnosis of the involved hip. The exclusion determined from markers placed on the ankles, knees, hips Clin Rheumatol (2019) 38:1163–1175 1165

Fig. 1 The placement of reflective markers

and other segments, and joint-centre displacement data were Results analysed. Kinetic parameters, including ground reaction force (GRF) and joint moments, were filtered using a 4th- Clinical characteristics order, zero-lag, low-pass Butterworth filter with a cut-off frequency of 50 Hz. The clinical characteristics of the subjects are displayed in Active surface electrodes (Trigno Wireless EMG System, Table 1. All the subjects were male. No significant differences Delsys, USA, sample rate 2000 Hz) were used to record sur- were found in relation to age or BMI. The average Harris Hip face electromyography (sEMG) signals. The sEMG for the Score was 68.94 ± 8.48, which indicated that the hip function bilateral gluteus maximus, gluteus medius, rectus femoris, bi- of the AS group was poor and impaired. The BASDAI, ceps femoris, tibialis anterior and gastrocnemius of each pa- BASFI, BASMI and VAS were not considered poor. The av- tient was measured. The electrodes were placed on the muscle erage score of the Berg balance scale was decreased in the AS bellies of each muscle. The placement of the respective sur- group; however, the score did remain above 40, indicating no face muscle electrode was according to the predecessor’sref- risk of falls during ambulation [26]. The full-length radio- erence [27]. Integrated electromyography and root mean graph of the spine indicated that all the sacroiliac joints of square amplitude were used to evaluate muscle function while the patients with AS and hip involvement were classified the patient was walking in circles. within the three to four grades of the BASRI-SIJ.

Statistical analysis Kinematic parameters

SPSS19.0 was used to analyse the data. The mean ± standard In the spatiotemporal gait parameters, the AS group demon- deviation for each gait parameter was calculated for both strated a lower gait velocity (p < 0.05), shorter step length groups across three trials; values were compared among the (p < 0.01) and shorter stride length (p < 0.01) contrast to the groups. The data were linearly interpolated to 100 different CS group. The average step length difference between the points to compare joint angles and moments at different ve- lower extremities in the AS group was larger than the differ- locities across the whole gait cycle. Independent samples t ence in the CS group (p < 0.05) (Table 1). tests were used to compare the gait parameters between the Both the initial contact and the toe-off flexion angle of the AS group and control group if the data represented normal hip joint in the AS group were less than those in the CS group distributions. Non-parametric tests, including the Mann– (p < 0.05). The abduction angle of the hip in the AS group was Whitney U test and the Wilcoxon rank sum test, were used larger than that in the CS group at the time of toe-off if the data demonstrated non-normal distributions. The corre- (p < 0.05). The maximal flexion, adduction and internal rota- lation between gait parameters and clinical characteristics of tion angles in the AS group were smaller than those in the CS AS patients with hip involvement was assessed by calculating group (p < 0.05). The maximal abduction angle of the hip in the Spearman correlation coefficient. All p values < 0.05 (two- the AS group was larger than that in the CS group (p <0.05). sided) were considered statistically significant. During one gait cycle, there was significantly decreased 1166 Clin Rheumatol (2019) 38:1163–1175

Table 1 The clinical characteristics and spatiotemporal gait parameters of the AS group and CS group

AS group (n = 18) (mean ± standard deviation) Control group (n = 18) (mean ± standard deviation) p value

Gender 18 Males 18 Males Age (years) 40.17 ± 6.37 40.67 ± 5.30 0.800 BMI (kg/m2) 25.36 ± 3.07 23.93 ± 0.97 0.073 Onset age (years) 26.00 ± 5.94 Disease duration (years) 14.06 ± 5.27 Harris hip score 68.94 ± 8.48 BASDAI 4.52 ± 1.15 BASFI 4.31 ± 1.97 BASMI 7.00 ± 3.36 VAS pain score 3.89 ± 1.64 Berg balance scale 43.67 ± 9.11 BASRI-h grade (n/%) BASRI-h grade 1 2(11.11) BASRI-h grade 2 6(33.33) BASRI-h grade 3 6(33.33) BASRI-h grade 4 4(22.23) BASRI-SIJ (n/%) BASRI-SIJ grade 3 10 (55.56) BASRI-SIJ grade 4 8 (44.44) Stride period (s) 1.06 ± 0.11 1.06 ± 0.05 0.904 Left single stance period (%) 38.29 ± 2.62 38.61 ± 1.55 0.569 Right single stance period (%) 38.12 ± 3.95 38.49 ± 1.66 0.642 Gait cadence (/s) 0.95 ± 0.09 0.94 ± 0.04 0.601 Gait velocity (m/s) 1.15 ± 0.21 1.25 ± 0.09 0.009 Left step length/height (m/m) 0.34 ± 0.06 0.38 ± 0.19 0.009 Right step length/height (m/m) 0.35 ± 0.39 0.38 ± 0.26 0.002 Step length difference (m) 0.04 ± 0.05 0.02 ± 0.02 0.024 Stride length/height (m/m) 0.70 ± 0.97 0.76 ± 0.42 0.002 Stride width/height (m/m) 0.08 ± 0.06 0.04 ± 0.01 <0.001

(p value <0.05) were statistically significant flexion and excessive abduction as well as external rotation in the shoulder–pelvic flexion angle at initial contact and toe-off, the AS group compared with those in the CS group was larger in the AS group than that in the CS group (p <0.05). (p <0.001)(Table2)(Fig.2). The average shoulder–pelvic angle of right obliquity throughout The knee flexion angle during the gait cycle in the AS group the entire gait cycle, as well as the shoulder–pelvic angle of right was significantly larger than that in the CS group (p <0.001). obliquity during initial contact, was larger in the AS group than There was a larger degree of knee flexion during the stance stage that in the CS group (p < 0.05). The average shoulder–pelvic of gait and a smaller degree of knee flexion during the swing stage angle of rotation throughout the entire gait cycle, as well as the in the AS group (p < 0.001). Maximal ankle plantar flexion in the shoulder–pelvic angle of rotation at initial contact, was smaller in AS group was larger than that in the CS group (p < 0.05). Ankle the AS group than that in CS group (p < 0.05). Over all, the AS plantar flexion in the AS group was reduced during one gait cycle group demonstrated more excessive right obliquity and restricted (p < 0.01) (Table 2)(Fig.2). trunk rotation during the gait cycle compared with the CS group The shoulder–pelvic angle is defined as the angle between a (Table 3)(Fig.3). line drawn between the two acromial processes and a line drawn between the two anterior superior iliac spines. This angle, in the Kinetic parameters sagittal, coronal and horizontal planes, represents trunk flexion– extension, obliquity and rotation, respectively, during the gait The joint moments throughout the gait cycle are reported in cycle [28]. In this study, we found that the average shoulder– Table 4. The hip moments for patients in the AS group dem- pelvic flexion angle throughout the entire gait cycle, as well as onstrated a significant reduction in flexion, abduction and Clin Rheumatol (2019) 38:1163–1175 1167

Table 2 Hip, knee (flexion/extension) and ankle (plantar/dorsiflexion) angles (°) of the AS group and CS group

AS group (mean ± standard CS group (mean ± standard p value deviation) deviation)

Initial contact hip flexion/extension angle (flexion is positive) 18.74 ± 10.35 28.82 ± 5.41 <0.001 Initial contact hip abduction/adduction angle (abduction is positive) 4.28 ± 5.60 3.76 ± 3.73 0.532 Initial contact hip external/internal rotation angle 7.27 ± 9.38 4.93 ± 6.13 0.092 (external rotation is positive) Toe-off hip flexion/extension angle (flexion is positive) 1.90 ± 11.26 5.34 ± 6.25 0.030 Toe-off hip abduction/adduction angle (abduction is positive) 9.38 ± 5.71 6.60 ± 4.14 0.002 Toe-off hip rotation angle (external rotation is positive) − 1.48 ± 6.80 − 1.59 ± 6.69 0.927 Maximal hip flexion angle during gait cycle 21.82 ± 11.23 34.54 ± 7.20 <0.001 Maximal hip extension angle during gait cycle 6.87 ± 10.07 4.58 ± 5.45 0.103 Maximal hip abduction angle during gait cycle 10.87 ± 5.69 8.59 ± 3.73 0.007 Maximal hip adduction angle during gait cycle 0.02 ± 5.81 4.61 ± 3.59 <0.001 Maximal hip external rotation angle during gait cycle 10.24 ± 8.15 8.32 ± 4.37 0.090 Maximal hip internal rotation angle during gait cycle 5.20 ± 5.13 7.25 ± 5.09 0.026 Hip flexion/extension angle during gait cycle (flexion is positive) 7.76 ± 8.81 15.51 ± 12.98 <0.001 Hip abduction/adduction angle during gait cycle 4.65 ± 2.66 1.18 ± 3.78 <0.001 (abduction is positive) Hip external/internal rotation angle during gait cycle 0.78 ± 2.62 − 1.26 ± 2.76 <0.001 (external rotation is positive) Initial contact knee flexion/extension angle (flexion is positive) 5.70 ± 5.54 5.51 ± 4.22 0.827 Toe-off knee flexion/extension angle (flexion is positive) 53.31 ± 10.64 46.16 ± 15.89 0.003 Maximal knee flexion angle during gait cycle 65.86 ± 9.71 66.44 ± 17.35 0.821 Minimal knee flexion angle during gait cycle 1.81 ± 4.90 0.80 ± 17.35 0.821 Knee flexion/extension angle during gait cycle (flexion is positive) 28.73 ± 15.56 26.35 ± 20.41 0.001 Knee flexion/extension angle during stance stage 21.34 ± 9.99 14.52 ± 9.26 <0.001 (flexion is positive) Knee flexion/extension angle during swing stage 42.66 ± 14.69 48.65 ± 16.66 <0.001 (flexion is positive) Initial contact ankle plantar/dorsiflexion angle − 5.06 ± 29.89 − 2.09 ± 5.02 0.464 (dorsiflexion is positive) Toe-off ankle plantar/dorsiflexion angle (dorsiflexion flexion is − 13.45 ± 33.18 − 11.22 ± 9.53 0.626 positive) Maximal ankle dorsiflexion angle during gait cycle 17.58 ± 33.59 10.48 ± 3.29 0.075 Maximal ankle plantar flexion angle during gait cycle 27.86 ± 27.96 19.56 ± 7.69 0.018 Ankle plantar/dorsiflexion angle during gait cycle − 0.65 ± 5.33 − 1.88 ± 5.09 0.003 (dorsiflexion is positive)

(p value <0.05) were statistically significant external rotation during the gait cycle (p < 0.001). The knee reported in Table 5. No significant differences were found in flexion moment for patients in the AS group was larger than the PvGRF between the two groups. that in the CS group (p < 0.001). The ankle plantar flexion Surface electromyography (sEMG) is a common tool moment for patients in the AS group was larger than that in utilised to obtain the measurement the electrical signals that the CS group (p <0.001)(Fig.4). accompany muscle contractions. sEMG is an important way Ground reaction force (GRF) refers to the force of the body to detect muscle activity on the surface of the skin. Integrated contacting the ground when the body is standing, walking or electromyography (IEMG) and root mean square amplitude running. The sole of the foot is reacting with the ground, and (RMS)wereusedtoevaluatethetime–domain criteria of ground forces are acting on the foot. GRF can also be de- sEMG. The IEMG and RMS of the two groups are reported scribed as contact force of the foot. GRF is measured utilising in Table 6. We found that most comparisons between the two force platforms, as they capture the resultant mechanical in- groups demonstrated no significant differences. The RMS for teraction between the foot and the ground. The GRF is then the rectus femoris, during both the single-leg-stance phase and represented by the platform coordinate system [29]. The peak the swing phase of gait, was higher in the AS group than that vertical ground reaction force (PvGRF) during the gait cycle is in the CS group (p < 0.05). The RMS for the gluteus maximus 1168 Clin Rheumatol (2019) 38:1163–1175

Fig. 2 The joint angles (degrees) during the gait cycles (stance phase 0–70%, swing phase 70–100%) of the AS group and CS group and the IEMG for the gluteus medius during the swing phase Correlation research of gait were higher in the AS group than those in the CS group (p <0.05). The correlations between the gait parameters and clinical characteristics of AS patients with hip involvement are

Table 3 Shoulder–pelvic angles (°) of the AS and CS groups

AS group (mean ± standard deviation) CS group (mean ± standard deviation) p value

Initial contact flexion/extension angle (flexion is positive) 5.08 ± 14.28 − 9.20 ± 4.22 <0.001 Initial contact obliquity angle (right is positive) 2.98 ± 2.98 1.67 ± 1.60 0.028 Initial contact rotation angle 3.08 ± 2.17 4.50 ± 2.22 0.012 Toe-off flexion/extension angle (flexion is positive) 4.98 ± 14.42 − 10.47 ± 5.06 <0.001 Toe-off oblique angle 3.59 ± 3.08 3.35 ± 1.90 0.702 Toe-off rotation angle 2.91 ± 1.84 3.27 ± 2.02 0.446 Maximal flexion angle during gait cycle 7.48 ± 13.69 − 7.47 ± 4.57 0.033 Maximal extension angle during gait cycle − 3.17 ± 14.45 12.37 ± 4.57 <0.001 Maximal oblique angle during gait cycle 4.58 ± 2.85 5.54 ± 1.54 0.091 Minimal oblique angle during gait cycle 1.87 ± 3.18 0.24 ± 0.65 0.005 Maximal rotation angle during gait cycle 4.49 ± 2.08 6.13 ± 1.25 <0.001 Minimal rotation angle during gait cycle 1.00 ± 1.49 0.03 ± 0.02 <0.001 Flexion/extension angle during gait cycle (flexion is positive) 5.53 ± 0.33 − 9.53 ± 0.66 <0.001 Oblique angle during gait cycle (right is positive) 3.09 ± 0.20 2.51 ± 0.73 <0.001 Rotation angle during gait cycle 2.75 ± 0.22 3.00 ± 0.97 <0.001

(p value <0.05) were statistically significant Clin Rheumatol (2019) 38:1163–1175 1169

Fig. 3 The shoulder–pelvic angles (degrees)during the gait cycles (stance phase 0–70%, swing phase 70–100%) of the AS group and CS group displayed in Table 7. The majority of the spatiotemporal This study was the first to examine the gait cycle in a popula- gait parameters, hip angles and flexion moments were tion of AS patients with hip involvement. positively correlated with the Harris hip scores and Berg The average age of onset of AS for patients in this study Balance Scale scores (all p values< 0.01) but negatively was 26.00 ± 5.94 years old, and the average disease duration correlated with the BASFI and BASRI-h grades (all p- was14.06±5.27years.HipinvolvementinASpatientsinthis values< 0.01). The shoulder–pelvic flexion angles and age demographic has a major impact on work and quality of oblique angles were positively correlated with disease du- life [32]. The average BMI of patients in the AS group was ration, BASFI, BASMI and BASRI-h (all p values < 0.05) 25.36 ± 3.07, indicating an overweight classification in the but negatively correlated with the Harris hip scores and Chinese population. Hyemin Jeong et al. [33]reportedthat Berg Balance Scale scores (all p values < 0.01). Most high BMI was one of the factors related to advanced baseline BASDAI and VAS pain scores demonstrated no correla- presence of hip arthritis in patients with AS. The BASRI tion with the gait parameters of the AS patients (p >0.05). scores for patients related to the sacroiliac joint were all grades 3–4, indicating that all the AS patients with hip involvement were in the middle and late stages of sacroiliac joint. Therefore, no meaningful relationship was observed between Discussion the sacroiliac joint and gait deviation in this study. In regard to the spatiotemporal gait parameters, the AS There have been three previous studies [8, 9, 30] evaluating group demonstrated decreased gait velocity, shorter step the gait parameters in patients with AS [31]. However, none of length and shorter stride length compared with the CS group. these studies focused on AS patients with hip involvement. Both Del Din S [8] and Zebouni L [9]reportedBatrend 1170 Clin Rheumatol (2019) 38:1163–1175

Table 4 Hip, knee (flexion/extension) and ankle (plantar/dorsiflexion) moments (N m/kg) of the AS group and CS group

AS group (mean ± standard CS group (mean ± standard p value deviation) deviation)

Initial hip contact flexion/extension moment (flexion is positive) 0.05 ± 0.22 0.09 ± 0.21 0.304 Initial hip contact abduction/adduction moment (abduction is positive) 0.05 ± 0.14 0.03 ± 0.14 0.506 Initial hip contact rotation moment (external rotation is positive) 0.01 ± 0.06 − 0.01 ± 0.06 0.231 Toe-off hip flexion/extension moment (flexion is positive) 0.19 ± 0.19 0.29 ± 0.12 0.002 Toe-off hip abduction/adduction moment (abduction is positive) − 0.04 ± 0.08 − 0.03 ± 0.06 0.558 Toe-off hip rotation moment (external rotation is positive) − 0.01 ± 0.04 − 0.01 ± 0.02 0.214 Maximal hip flexion moment during gait cycle 0.74 ± 0.22 1.03 ± 0.20 <0.001 Maximal hip extension moment during gait cycle 0.83 ± 0.32 0.88 ± 0.22 0.340 Maximal hip abduction moment during gait cycle 0.42 ± 0.22 0.44 ± 0.28 0.728 Maximal hip adduction moment during gait cycle 0.38 ± 0.31 0.47 ± 0.34 0.127 Maximal hip external rotation moment during gait cycle 0.15 ± 0.08 0.21 ± 0.05 <0.001 Maximal hip internal rotation moment during gait cycle 0.11 ± 0.05 0.10 ± 0.05 0.165 Hip flexion/extension moment during gait cycle (flexion is positive) 0.06 ± 0.29 0.13 ± 0.37 <0.001 Hip abduction/adduction moment during gait cycle 0.19 ± 0.17 0.26 ± 0.25 <0.001 (abduction is positive) Hip external/internal rotation moment during gait cycle (external rotation 0.01 ± 0.03 0.04 ± 0.05 <0.001 is positive) Initial contact knee flexion/extension moment (flexion is positive) − 0.01 ± 0.10 0.03 ± 0.11 0.084 Toe-off knee flexion/extension moment (flexion is positive) 0.07 ± 0.11 0.08 ± 0.09 0.769 Maximal knee flexion moment during gait cycle 0.73 ± 0.27 0.63 ± 0.17 0.010 Maximal knee extension moment during gait cycle 0.27 ± 0.10 0.35 ± 0.08 <0.001 Knee flexion/extension moment during gait cycle (flexion is positive) 0.16 ± 0.20 0.08 ± 0.16 <0.001 Initial contact ankle plantar/dorsiflexion moment 0.01 ± 0.01 0.01 ± 0.01 0.406 (dorsiflexion is positive) Toe-off ankle plantar/dorsiflexion moment (dorsiflexion is positive) − 0.01 ± 0.07 − 0.01 ± 0.06 0.621 Maximal ankle dorsiflexion moment during gait cycle 0.11 ± 0.05 0.28 ± 0.39 0.002 Maximal ankle plantar flexion moment during gait cycle 1.09 ± 0.25 1.03 ± 0.39 0.296 Ankle plantar/dorsiflexion moment during gait cycle − 0.30 ± 0.32 − 0.22 ± 0.27 <0.001 (dorsiflexion is positive)

(p value <0.05) were statistically significant towards reduction^ (no significant difference) in gait velocity muscle weakness. Excessive abduction of the hip is an and stride length. These results indicated that hip involvement indicator of abductor contracture, lower limb length asym- exerted influence on the spatiotemporal gait parameters for metry or scoliosis with pelvic inclination [34]. Excessive patients with AS. The step length difference of the AS group external rotation of the hip is indicative of a contracture of was larger than that of the CS group, which suggested that the the external rotation muscles and purposeful positioning, disease impaired coordination of the gait cycle. which avoids taxing inflamed joints or the over use of the The present study demonstrated decreased flexion and gluteus maximus [34]. In this study, the average knee excessive abduction and external rotation of the hip an- flexion angle during the gait cycle in the AS group was gles during the gait cycle in patients with AS. The reduc- significantly larger than that in the CS group. There was a tion of the hip flexion angle is an expression of several larger degree of knee flexion during the stance stage and a factors, including hip flexor muscle weakness, hip pain, smaller degree of flexion during the swing stage. This hamstring muscle spasm, reduction of hip extensor de- result was different from those of previous studies [8, mand and foot dragging [34]. Hip flexor muscle weakness 9]. Compensation for the reduction of hip flexion and can also contribute to low gait velocity and shorter step hamstring muscle contracture could be an explanation length [35]; the results of our study related to spatiotem- for the excessive flexion of the knee during the stance poral gait parameters were consistent with this phenome- stage. During the swing stage, weakness of the quadriceps non. These findings suggest that hip involvement in pa- and hip flexors could contribute to the reduction in knee tients with AS might contribute to hip pain and hip flexor joint flexion [34]. The ankle angles during the gait cycle Clin Rheumatol (2019) 38:1163–1175 1171

Fig. 4 The joint moments (N m/kg) during the gait cycle (stance phase 0–70%, swing phase 70–100%) of the AS group and CS group of patients in the AS group were largely variable. between the pelvis and trunk and poor coordination of Decreased ankle plantar flexion was present throughout the upper limb swing during ambulation could contribute most of the gait cycle, which was similarly observed in to the reduction of trunk rotation in AS patients with hip a previous study [8]. The ankle dysfunction may be com- involvement [30]. pensation for the reduction of hip flexion. The hip movement of the AS group demonstrated a signif- In our study, the AS group demonstrated increased for- icant reduction in flexion, abduction and external rotation dur- ward trunk flexion, excessive obliquity and restricted ro- ing the gait cycle. The flexion movement of the hip is depen- tation during the gait cycle, which was in contrast to the dent upon two factors: flexor muscle force and the flexor lever CS group. These results were divergent from the findings arm. This result indicated that the strength of the hip flexor of a previous study [8]. Additional factors beyond anterior muscle group for patients in the AS group is lower than that in spinal flexion, including muscle weakness of the quadri- the control group. The diminished hip flexor strength could be ceps femoris and flexion contracture of the hip joint, due to the restricted hip flexion present in patients with AS could contribute to forward trunk flexion during the gait and hip involvement. Over time, as patients become less able cycle. Abduction myasthenia of the hip and scoliosis of to actively engage the hip flexors, or less willing to engage the the spine could contribute to the excessive trunk obliquity hip flexors due to pain, myasthenia and disuse atrophy be- in AS patients with hip involvement. Poor coordination come complications [36]. The abduction and external rotation

Table 5 The peak vertical ground reaction force (PvGRF) (N/kg) AS group (mean ± standard CS group (mean ± standard p during the gait cycle of the AS deviation) deviation) value group and CS group Left PvGRF (N/kg) 10.84 ± 1.19 11.16 ± 0.53 0.155 Right PvGRF (N/kg) 11.06 ± 1.33 11.26 ± 0.73 0.446 Average PvGRF (N/kg) 10.95 ± 1.19 11.21 ± 0.56 0.255 1172 Clin Rheumatol (2019) 38:1163–1175

Table 6 sEMG of the AS group and CS group

AS group (mean ± standard deviation) CS group (mean ± standard deviation) p value

Standardised IMEG for single-leg-stance phase Gluteus maximus 0.23 ± 0.14 0.19 ± 0.06 0.191 Gluteus medius 0.25 ± 0.16 0.25 ± 0.94 0.911 Rectus femoris 0.28 ± 0.16 0.23 ± 0.08 0.251 Biceps femoris 0.19 ± 0.08 0.14 ± 0.06 0.061 Tibialis anterior 0.19 ± 0.10 0.11 ± 0.06 0.013 Gastrocnemius 0.32 ± 0.18 0.35 ± 0.18 0.657 Standardised IMEG for swing phase Gluteus maximus 0.23 ± 0.13 0.20 ± 0.06 0.296 Gluteus medius 0.23 ± 0.15 0.14 ± 0.06 0.019 Rectus femoris 0.22 ± 0.16 0.19 ± 0.06 0.476 Biceps femoris 0.30 ± 0.14 0.31 ± 0.14 0.837 Tibialis anterior 0.29 ± 0.09 0.27 ± 0.10 0.632 Gastrocnemius 0.16 ± 0.13 0.10 ± 0.05 0.057 Standardised RMS for single-leg-stance phase Gluteus maximus 0.38 ± 0.08 0.35 ± 0.08 0.374 Gluteus medius 0.35 ± 0.08 0.37 ± 0.08 0.401 Rectus femoris 0.34 ± 0.08 0.29 ± 0.04 0.019 Biceps femoris 0.37 ± 0.09 0.34 ± 0.07 0.379 Tibialis anterior 0.35 ± 0.07 0.37 ± 0.06 0.449 Gastrocnemius 0.40 ± 0.07 0.35 ± 8.27 0.087 Standardised RMS for swing phase Gluteus maximus 0.39 ± 0.07 0.33 ± 0.06 0.025 Gluteus medius 0.36 ± 0.09 0.36 ± 0.09 0.905 Rectus femoris 0.38 ± 0.07 0.29 ± 0.06 <0.001 Biceps femoris 0.35 ± 0.06 0.33 ± 0.05 0.218 Tibialis anterior 0.40 ± 0.79 0.45 ± 0.07 0.083 Gastrocnemius 0.38 ± 0.10 0.37 ± 0.08 0.784

(p value <0.05) were statistically significant

of the hip joint during the gait cycle may indicate weakness of two groups in the IEMG measurements. These results suggest the muscles of abduction and external rotation. This scenario that, in patients with AS and hip involvement, the behaviour of increased hip abduction and external rotation could be dis- of the neuromuscular system is not influenced. The RMS for ease-induced, related to limited adduction and externally ro- rectus femoris for patients in the AS group, in both the single- tated protective positioning of the hip joint [37]. leg-stance phase and swing phase of gait, was higher than that In our study, no difference was found in the PvGRF be- of the CS group, indicating that rectus femoris needs to recruit tween the two groups. The PvGRF reflects the supportive more muscle fibres during the gait cycle in the presence of AS capability of the lower limbs during ambulation. The results and hip involvement. This result could also account for the indicated that the working load of the lower limbs during reported increase in subject fatigue during ambulation. ambulation was not influenced by the disease. Most gait parameters for the AS patients were correlated sEMG is utilised in research, across various disciplines, to with the Harris hip, BASFI, BASFI-h and Berg Balance Scale investigate a wide range of research questions. sEMG is per- scores. These results indicated that the severity of hip joint haps most useful for providing insight into the behaviour of involvement (both joint dysfunction and radiographic display the neuromuscular system [38]. sEMG is a measurement of of damage) and the overall functional status of patients with muscle excitation, as it measures changes in the polarity of the AS had great influence on the gait deviations of AS patients muscle fibre membrane resulting from neural excitation. In with hip involvement. The BASDAI and VASpain score were our study, we found no significant differences between the not correlated with the gait parameters, suggesting that the lnRemtl(09 38:1163 (2019) Rheumatol Clin

Table 7 Correlation between gait parameters and clinical characteristics of AS patients with hip involvement

Disease duration Harris hip score BASDAI BASFI BASMI VAS pain score Berg Balance Scale BASRI-h Grade

rprprprprprpr p r p

− 0.002 < 0.001 − − < 0.001 − − < 0.001 < 0.001 − < 0.001 – Gait velocity 0.351 0.468 0.148 0.215 0.513 0.222 0.060 0.521 0.566 0.701 1175 Left step length/height − 0.054 0.651 0.427 < 0.001 − 0.016 0.892 − 0.574 < 0.001 − 0.257 0.029 − 0.207 0.081 0.584 < 0.001 − 0.688 < 0.001 Right step length/height 0.099 0.410 0.310 0.008 0.012 0.920 − 0.407 < 0.001 − 0.195 0.100 − 0.134 0.260 0.501 < 0.001 − 0.605 < 0.001 Stride length/height − 0.034 0.776 0.406 < 0.001 0.006 0.958 − 0.562 < 0.001 − 0.252 0.033 − 0.170 0.154 0.567 < 0.001 − 0.670 < 0.001 Gait cadence − 0.494 0.002 0.852 < 0.001 − 0.072 0.674 − 0.783 < 0.001 − 0.317 0.060 − 0.318 0.059 0.877 < 0.001 − 0.836 < 0.001 Initial contact hip flexion angle − 0.618 < 0.001 0.231 0.051 0.075 0.531 − 0.387 0.001 − 0.390 0.001 − 0.133 0.266 0.337 0.004 − 0.347 0.003 Maximal hip flexion angle − 0.557 < 0.001 0.292 0.013 0.071 0.552 − 0.381 0.001 − 0.309 0.008 − 0.167 0.160 0.376 0.001 − 0.402 < 0.001 Maximal hip extension angle 0.517 < 0.001 0.073 0.543 − 0.062 0.605 − 0.013 0.915 0.148 0.214 0.029 0.811 − 0.031 0.794 0.107 0.370 Maximal hip abduction angle 0.008 0.950 0.045 0.709 0.021 0.860 0.017 0.888 0.228 0.054 0.009 0.941 0.015 0.901 − 0.149 0.210 Maximal hip adduction angle − 0.198 0.096 − 0.411 < 0.001 − 0.159 0.183 − 0.330 0.005 − 0.277 0.019 − 0.397 0.001 − 0.378 0.001 − 0.243 0.040 Maximal hip external rotation angle − 0.087 0.467 0.282 0.017 0.018 0.879 − 0.133 0.264 − 0.061 0.611 − 0.078 0.517 0.152 0.202 − 0.033 0.782 Maximal hip internal rotation angle 0.029 0.809 0.045 0.708 0.032 0.793 − 0.057 0.635 0.045 0.708 0.176 0.139 − 0.095 0.428 0.215 0.070 Initial contact shoulder–pelvic flexion angle 0.767 < 0.001 − 0.324 0.006 0.033 0.783 0.429 < 0.001 0.351 0.003 0.135 0.260 − 0.338 0.004 0.300 0.010 Maximal shoulder–pelvic flexion angle 0.768 < 0.001 − 0.365 0.002 0.015 0.901 0.465 < 0.001 0.366 0.002 0.165 0.166 − 0.376 0.001 0.346 0.003 Minimal shoulder–pelvic flexion angle 0.759 < 0.001 − 0.341 0.003 0.004 0.976 0.452 < 0.001 0.349 0.003 0.127 0.286 − 0.355 0.002 0.325 0.005 Initial contact shoulder–pelvic obliquity angle 0.240 0.043 − 0.581 < 0.001 − 0.158 0.186 0.753 < 0.001 0.689 < 0.001 0.060 0.618 − 0.556 < 0.001 0.329 0.005 Maximal shoulder–pelvic oblique angle 0.310 0.008 − 0.143 0.230 − 0.172 0.148 0.104 0.382 − 0.125 0.294 − 0.071 0.556 − 0.004 0.976 0.125 0.296 Minimal shoulder–pelvic oblique angle 0.576 < 0.001 − 0.508 < 0.001 − 0.057 0.632 0.767 < 0.001 0.512 < 0.001 0.092 0.443 − 0.519 < 0.001 0.462 < 0.001 Initial contact shoulder–pelvic rotation angle − 0.119 0.317 0.088 0.463 0.079 0.510 − 0.138 0.249 − 0.032 0.792 0.037 0.761 0.080 0.503 − 0.097 0.418 Maximal shoulder–pelvic rotation angle 0.168 0.159 0.138 0.247 0.359 0.002 − 0.179 0.133 − 0.178 0.133 0.376 0.001 0.121 0.312 − 0.024 0.836 Minimal shoulder–pelvic rotation angle 0.169 0.156 − 0.143 0.229 0.454 < 0.001 0.351 0.003 0.171 0.151 0.087 0.469 − 0.245 0.038 0.147 0.218 Maximal hip flexion moment 0.304 0.010 0.378 0.001 0.012 0.920 − 0.286 < 0.015 0.184 0.122 0.162 0.173 0.410 < 0.001 − 0.444 < 0.001 Maximal hip extension moment 0.114 0.342 − 0.208 0.080 0.084 0.484 0.046 0.700 − 0.041 0.735 0.385 0.001 − 0.339 0.004 0.516 < 0.001 Maximal hip abduction moment 0.176 0.140 0.170 0.153 − 0.090 0.453 − 0.013 0.916 0.195 0.100 − 0.095 0.425 0.180 0.130 − 0.272 0.021 Maximal hip adduction moment − 0.244 0.039 0.037 0.757 − 0.172 0.148 − 0.063 0.597 − 0.148 0.216 − 0.214 0.071 0.104 0.385 − 0.102 0.392 Maximal hip external rotation moment − 0.386 0.001 0.104 0.386 − 0.373 0.001 − 0.127 0.286 − 0.018 0.879 − 0.211 0.075 0.206 0.083 − 0254 0.031 Maximal hip internal rotation moment − 0.114 0.341 − 0.061 0.610 − 0.321 0.006 0.010 0.932 0.040 0.738 − 0.207 0.082 − 0.049 0.680 0.024 0.839

(p value <0.05) were statistically significant 1173 1174 Clin Rheumatol (2019) 38:1163–1175 activity of the disease and the associated pain were not the Bath Ankylosing Spondylitis Functional Index; BASMI, Bath Ankylosing VAS pain main factors affecting the gait deviations in AS patients with Spondylitis Metrology Index; , Visual Analog Scale for Pain; BASRI-h, Bath Ankylosing Spondylitis Hip Radiology Index; BASRI-SIJ, hip involvement. Bath Ankylosing Spondylitis Radiology Index on Sacroiliac Joint; GRF, Some limitations in our study should be noted. First, the ground reaction force; PvGRF, peak vertical ground reaction force; sample size was limited. Limited sample size may affect the sEMG, surface electromyography; IEMG, integrated electromyography; RMS reliability of sampling inference. Second, this study was a , root mean square amplitude Publisher’sNoteSpringer Nature remains neutral with regard to jurisdic- cross-sectional study, which did not take into account the tional claims in published maps and institutional affiliations. exact causality of gait abnormalities among the sample population. Third, this study speculated that gait deviations could be due to hip flexor muscle weakness and abductor References muscle contracture; however, gait analysis could not direct- ly measure muscle strength. In future studies, a dynamic 1. Jiménez-Balderas FJ, Mintz G (1993) Ankylosing spondylitis: clin- simulation model based on gait analysis results should be ical course in women and men. J Rheumatol 20(12):2069–2072 built to investigate the effects of muscle strength on gait 2. Yacoub YI, Laatiris A, Hajjaj-Hassouni N (2012) Gender and disease features in Moroccan patients with ankylosing spondylitis. parameters. Clin Rheumatol 31(2):293–297. https://doi.org/10.1007/s10067- In our study, the gait deviations of AS patients with hip 011-1819-x involvement were described. The gait analysis results demon- 3. Cruyssen BV, Muñozgomariz E, Font P, Mulero J, Vlam KD, strated statistically significant alterations regarding the kine- Boonen A, Vazquezmellado J, Flores D, Vastesaeger N, Collantes E (2010) Hip involvement in ankylosing spondylitis: epidemiology matic and kinetic gait parameters for the patients included in and risk factors associated with hip replacement surgery. the sample. Coordination and balance were impaired by the Rheumatology (Oxford) 49(1):73–81. https://doi.org/10.1093/ disease. Long-term gait deviations may further aggravate the rheumatology/kep174 degree of hip joint in patients with AS, so timely treatment and 4. Kim TJ, Kim TH (2010) Clinical spectrum of ankylosing spondylitis in Korea. Joint Bone Spine 77(3):235–240. https://doi.org/10. targeted rehabilitation training should be carried out for these 1016/j.jbspin.2009.11.015 AS patients. These results can provide data support for reha- 5. Turner DE, Helliwell PS, Siegel KL, Woodburn J (2008) bilitation training and guide doctors or therapists to carry out Biomechanics of the foot in rheumatoid arthritis: identifying abnor- relevant treatment. For AS patients with hip involvement, the mal function and the factors associated with localised disease 'impact. Clin Biomech 23(1):93–100. https://doi.org/10.1016/j. results of gait analysis could be utilised to formulate efficient clinbiomech.2007.08.009 physical exercise or rehabilitation plans to increase the range 6. Broström EW, Esbjörnsson AC, Von HJ, Iversen MD (2012) Gait of motion of affected joints, stretch the flexor and abductor deviations in individuals with inflammatory joint diseases and os- muscles of the hip and improve balance during ambulation. teoarthritis and the usage of three-dimensional gait analysis. Best Pract Res Clin Rheumatol 26(3):409–422. https://doi.org/10.1016/ j.berh.2012.05.007 Acknowledgements The authors thank all the patients and volunteers 7. Barkham N, Coates LC, Keen H, Hensor E, Fraser A, Redmond A, that participated in the study. Cawkwell L, Emery P (2010) Double-blind placebo-controlled trial of etanercept in the prevention of work disability in ankylosing Authors’ contributions Guoning Zhang, Jia Li and Zhengliang Xia con- spondylitis. Ann Rheum Dis 69(11):1926–1928. https://doi.org/ tributed equally to the manuscript; they conceived and designed the study. 10.1136/ard.2009.121327 Guoning Zhang and Jia Li collected the clinical data. Guoning Zhang and 8. Din SD, Carraro E, Sawacha Z, Guiotto A, Bonaldo L, Masiero S, Zhengliang Xia conducted the gait analysis and performed the statistical Cobelli C (2011) Impaired gait in ankylosing spondylitis. Med Biol analysis. Weidong Xu approved the final version. Eng Comput 49(7):801–809. https://doi.org/10.1007/s11517-010- 0731-x Funding sources This research was supported by funding from the 9. Zebouni L, Helliwell PS, Howe A, Wright V (1992) Gait analysis in BEstablishment of parachute training computer simulation system for ankylosing spondylitis. Ann Rheum Dis 51(7):898–899 parachute training^ (Grant No. BKJ14J004). 10. Helliwell PS, Smeathers JE, Wright V (1989) Shock absorption by the spinal column in normals and in ankylosing spondylitis. Proc – Compliance with ethical standards Inst Mech Eng H 203(4):187 190. https://doi.org/10.1243/PIME_ PROC_1989_203_037_01 11. Sahin N, Ozcan E, Baskent A, Karan A, Kasikcioglu E (2011) Muscular kinetics and fatigue evaluation of knee using by isokinetic All patients provided written informed consent. The study was conducted dynamometer in patients with ankylosing spondylitis. Acta by the orthopaedic department of our hospital and was approved by the Reumatol Port 36(3):252 clinical research ethics committee of Changhai Hospital (CHEC2013- 12. Turan Y, Duruöz MT, Bal S, Guvenc A, Cerrahoglu L, Gurgan A 194). (2007) Assessment of fatigue in patients with ankylosing spondylitis. Rheumatol Int 27(9):847–852. https://doi.org/10.1007/ Disclosures None. s00296-007-0313-x 13. Aydog E, Depedibi R, Bal A, Eksioglu E, Unlu E, Cakci A (2005) Abbreviations AS, ankylosing spondylitis; BMI, body mass index; Dynamic postural balance in ankylosing spondylitis patients. BASDAI, Bath Ankylosing Spondylitis Disease Activity Index; BASFI, Rheumatology 45(4):445–448. https://doi.org/10.1093/ rheumatology/kei192 Clin Rheumatol (2019) 38:1163–1175 1175

14. Bot SD, Caspers M, Royen BJV, Toussaint HM, Kingma I (1999) 26. Berg K, Wood-Dauphinee S, Williams JI (1995) The Balance Scale: Biomechanical analysis of posture in patients with spinal kyphosis reliability assessment with elderly residents and patients with an due to ankylosing spondylitis: a pilot study. Rheumatology 38(5): acute stroke. Scand J Rehabil Med 27(1):27 441–443 27. Heinrichs K (1998) Introduction to surface electromyography. 15. Cappozzo A (1984) Gait analysis methodology. Hum Mov Sci 3(1– Pearson Education/Prentice Hall, pp 527–531 2):27–50. https://doi.org/10.1016/0167-9457(84)90004-6 28. Park WH, Kim YH, Lee TR, Sung PS (2012) Factors affecting 16. Simon SR (2004) Quantification of human motion: gait analysis— shoulder–pelvic integration during axial trunk rotation in subjects benefits and limitations to its application to clinical problems. J with recurrent low back pain. Eur Spine J 21(7):1316–1323. https:// Biomech 37(12):1869–1880. https://doi.org/10.1016/j.jbiomech. doi.org/10.1007/s00586-012-2280-5 2004.02.047 29. Camomilla V, Cereatti A, Cutti AG, Fantozzi S, Stagni R, Vannozzi 17. Chiu M-C, Wang M-J (2007) The effect of gait speed and gender on G (2017) Methodological factors affecting joint moments estima- perceived exertion, muscle activity, joint motion of lower extremity, tion in clinical gait analysis: a systematic review. Biomed Eng ground reaction force and heart rate during normal walking. Gait Online 16(1):106. https://doi.org/10.1186/s12938-017-0396-x. Posture 25(3):385–392. https://doi.org/10.1016/j.gaitpost.2006.05. 30. Mangone M, Scettri P, Paoloni M, Procaccianti R, Spadaro A, 008 Santilli V (2011) Pelvis-shoulder coordination during level walking – 18. Cohen J (1988) Statistical power analysis for the behavioral sci- in patients with ankylosing spondylitis. Gait Posture. 34(1):1 5. ences. Technometrics 31(4):499–500. https://doi.org/10.1080/ https://doi.org/10.1016/j.gaitpost.2011.02.002 00401706.1989.104886 31. Carroll M, Parmar P, Dalbeth N, Boocock M, Rome K (2015) Gait characteristics associated with the foot and ankle in inflammatory 19. Linden SVD, Valkenburg HA, Cats A (1984) Evaluation of diag- arthritis: a systematic review and meta-analysis. BMC nostic criteria for ankylosing spondylitis. A proposal for modifica- Musculoskelet Disord 16:134. https://doi.org/10.1186/s12891- tion of the New York criteria. Arthritis Rheum 27(4):361–368 015-0596-0 20. Mackay K, Mack C, Brophy S, Calin A (2010) The bath ankylosing 32. He C, He X, Tong W, Zheng W, Zhang T, Zhao J, Xu W (2016) The spondylitis radiology index (BASRI): a new, validated approach to effect of total hip replacement on employment in patients with disease assessment. Arthritis Rheum 41(12):2263–2270. https:// ankylosing spondylitis. Clin Rheumatol 35(12):2975–2981. doi.org/10.1002/1529-0131(199812)41:12<2263::AID-ART23>3. https://doi.org/10.1007/s10067-016-3431-6 0.CO;2-I 33. Jeong H, Eun YH, Kim IY, Kim H, Lee J, Koh EM, Cha HS (2017) 21. Harris WH (1969) Traumatic arthritis of the hip after dislocation Characteristics of hip involvement in patients with ankylosing . and acetabular fractures: treatment by mold arthroplasty An end- spondylitis in Korea. Korean J Intern Med 32(1):158–164. https:// result study using a new method of result evaluation. J Bone Joint doi.org/10.3904/kjim.2015.229 – Surg Am 51(4):737 755 34. Perry J, Burnfield J M, Cabico L M(2010) Gait analysis: normal 22. Garrett S, Jenkinson T, Kennedy LG, Whitelock H, Gaisford P, and pathological function. Slack Calin A (1994) A new approach to defining disease status in anky- 35. Burnfield JM, Josephson KR, Powers CM, Rubenstein LZ (2000) losing spondylitis: the Bath Ankylosing Spondylitis Disease The influence of lower extremity joint torque on gait characteristics – Activity Index. J Rheumatol 21(12):2286 2291 in elderly men. Arch Phys Med Rehabil 81(9):1153–1157. https:// 23. Calin A, Garrett S, Whitelock H, Kennedy LG, O'Hea J, Mallorie P, doi.org/10.1053/apmr.2000.7174 Jenkinson T (1994) A new approach to defining functional ability in 36. Rasch A, Byström AH, Dalén N, Martinezcarranza N, Berg HE ankylosing spondylitis: the development of the Bath Ankylosing (2009) Persisting muscle atrophy two years after replacement of Spondylitis Functional Index. J Rheumatol 21(12):2281 the hip. J Bone Joint Surg Br Vol 91(5):583–588 24. Van der Heijde D, Landewé R, Feldtkeller E (2008) Proposal of a 37. Beaulieu ML, Lamontagne MBeaule PE (2010) Lower limb bio- linear definition of the Bath Ankylosing Spondylitis Metrology mechanics during gait do not return to normal following total hip Index (BASMI) and comparison with the 2-step and 10-step defi- arthroplasty. Gait Posture 32(2):269–273. https://doi.org/10.1016/j. nitions. Ann Rheum Dis 67(4):489–493. https://doi.org/10.1136/ gaitpost.2010.05.007 ard.2007.074724 38. Vigotsky AD, Halperin I, Lehman GJ, Trajano GS, Vieira TM 25. Mccormack HM, Horne DJ, Sheather S (1988) Clinical applications (2017) Interpreting signal amplitudes in surface electromyography of visual analogue scales: a critical review. Psychol Med 18(4): studies in sport and rehabilitation sciences. Front Physiol 8:985. 1007–1019 https://doi.org/10.3389/fphys.2017.00985