Houghton, Kristin Michelle MD, MSc, FRCPC, Dip Sports Med; Guzman, Jaime MD, MSc, FRCPC
INTRODUCTION AND PURPOSE
Body balance involves numerous neuromuscular processes controlled by sensory input, central processing, and neuromuscular responses.1 Many factors may influence balance, including leg dominance, fatigue, age, gender, weight, height, foot size, habitual physical activity, and previous lower extremity injury.2–7
Proprioception is defined as the sense of posture, movement, and changes in equilibrium and the knowledge of position of limbs and body in space.8 Proprioceptive mechanoreceptors present in synovial joints, capsule, muscles, ligaments, menisci, and skin provide afferent input at the spinal level for arthrokinetic and muscular reflexes and at the supraspinal level for motor planning and motor learning.9 Smooth coordinated joint movements play a key role in dynamic joint stability.
Children with juvenile idiopathic arthritis (JIA) may have both proprioceptive deficits and proprioceptive impairments in their motor responses, making them susceptible to balance perturbations. Arthritis of lower extremity joints may be associated with loss of muscle strength and range of motion; instability of weight-bearing joints; and impairment of ambulation, mobility, and exercise tolerance.10 Delayed motor development has been reported, with deficits more evident with increasing age and increased demands for complex gross motor skills.11,12
Disturbance of the proprioceptive network by inflammatory arthritis has been shown in adults with increased postural sway and decreased balance reported in rheumatoid arthritis (RA).13,14 Poor balance has also been recognized in children and young adults with arthropathy associated with hemophilia.15–17
To date, few studies examined balance in children with JIA.18,19 It is our clinical observation that some children with JIA have poor balance, demonstrated by difficulty in single-leg standing and single-leg hopping. The purpose of the present study was to evaluate and describe dynamic postural balance by using the Biodex Balance System (BBS; Biodex, Shirley, New York) in children with JIA. Our hypothesis was that children with JIA have poor dynamic balance compared with children who are healthy. We also explored the relationship between balance and muscle strength, disease activity and functional impairment.
Subjects aged 8 to 18 years, with a definite diagnosis of JIA and swelling of lower extremity joints within the year before testing, who were followed at the Pediatric Rheumatology Program at a single tertiary care hospital were invited to participate.20 The control group was composed of children aged 8 to 18 years who were healthy. Subjects and controls were excluded from participation if they had auditory or visual impairment (reduced visual acuity allowable if corrected with lens/glasses) or orthopedic injury involving the lower extremities. Study assessments occurred between July and December 2008 for subjects with JIA and between July 2008 and August 2011 for the control group. All subjects and caregivers provided assent and consent before participation. Ethical approval was obtained from the University Clinical Research Ethics Board and the hospital's Research Review Committee.
The subjects with JIA underwent a clinical assessment by a rheumatologist (single examiner) the same day as the balance testing. The following data were collected: active joint count, age, sex, leg dominance, history of previous lower extremity orthopedic injury, JIA subtype, and current medications. Active joint count was determined as the number of joints with either swelling or painful, limited range of motion. Articular severity index was calculated as the sum of scores for joint swelling, pain on motion, tenderness, and limitation of motion.21 Subjects or their proxy (parent or guardian) completed the Childhood Health Assessment Questionnaire (CHAQ), a valid and reliable measure of function in children with JIA, with potential score of 0 to 3, where higher scores indicate greater functional impairment.22 Pain was measured on a 100-mm visual analog scale (VAS), where higher scores indicate more pain.22
Height (Harpenden Stadiometer; London, England) and body mass (Seca Electronics Model 767, Hamburg, Germany) were measured to the nearest 0.1 cm and 0.1 kg, respectively. Body mass index (BMI) was calculated (kg/m2). Obesity was defined as BMI greater than the 95th percentile for age according to reference data.23 Isometric strength of both legs was tested (hip abductors, flexors, extensors; knee extensor, flexors; ankle dorsiflexors and plantar flexors) by manual muscle testing (MMT), and the average of 3 trials was recorded.24 All muscle testing was performed by a single examiner.
Balance was assessed with the BBS, a commercially available balance device. The BBS is a multiaxial tilting platform that allows objective measurement of the ability of a subject to maintain dynamic single- and double-limb postural stance on an unstable platform. The movable balance platform provides up to 20° of surface tilt within a 360° range of motion. This stabilometric technique allows assessment of overall (OA) single- and double-limb postural stability in addition to anterior/posterior (AP) and medial/lateral (ML) stability. The OA stability index represents the variance of foot platform displacement in degrees, from level, in all motions during a test. For static measures, the stability index is measured as the angular excursion of a patient's center of gravity. The variance of foot platform displacement in degrees, from level, for motion in the sagittal plane is represented by the AP stability score, and the variance of foot platform displacement in degrees, from level, for motion in the frontal plane is represented by the ML stability score. A high score is indicative of a lot of movement and indicates poor balance. The OA stability score is believed to be the best indicator of the OA ability of the patient to balance the platform.25 The stability platform allows for varying levels of difficulty of stability testing from level 12 (static) to level 1 (least stable). We assessed average OA, AP, and ML stability scores under the following 5 conditions: right-leg static balance, left-leg static balance, bilateral static balance, bilateral dynamic balance at BBS level 2 (very unstable), and bilateral dynamic balance at BBS level 7 (moderately unstable). Subjects were trained for 3 minutes on the BBS before testing. Each condition was then tested in random sequence as a set of 3 trials of 20 seconds each, separated by at least 10 seconds between trials and a rest period of 2 minutes between conditions. If a subject was unable to balance for the full 20 seconds of a trial, trial times were recorded to the nearest 1/100 of a second. Loss of balance included touching the floor with the non–weight-bearing foot during single-leg balance, touching the handles, or shifting the weight-bearing foot/feet. All tests were performed barefoot. During bilateral stance, subjects were instructed to step on the BBS platform and assume a position of comfort with slight knee flexion and their arms placed across their chest or on their hips. During single-leg stance, subjects were asked to hold their non–weight-bearing leg in hip and knee flexion (thigh parallel to floor).
Sufficient data were not available for BBS balance outcomes in pediatric patients on which to base an accurate sample size calculation for this preliminary study. Since at least 10 subjects per group are needed for balance measures when using the BBS, we decided to enroll 25 subjects with JIA because we anticipated this group to have greater variance in balance measures than children who were healthy. Controls were not individually matched to each subject with JIA because of difficulties with enrollment, but they were within the same age range as a group. Descriptive statistics were calculated for subject demographics and disease characteristics. Mean and standard deviation are presented for normally distributed data. Median values and ranges are presented for nonnormally distributed data. Difference in balance stability scores between subjects with JIA and controls was tested using an independent sample t test. Pearson correlation coefficients were calculated to evaluate the JIA group for possible relationships between balance and strength. Correlation coefficients of 0.3 to 0.5 were considered low, 0.5 to 0.7 moderate, and 0.7 to 1.0 as high. Post hoc t tests were done to compare dynamic balance measures for those subjects with JIA with and without pain, functional impairments, and active disease. Significance level for all tests was set at α < .05. Statistics were performed using STATA 10 (StataCorp LP, College Station, Texas).
Twenty-five subjects with JIA (14 boys), aged 8.7 to 18.2 years (median, 13.8 years), and 36 controls (22 boys), aged 8.2 to 18.0 years (median, 9.5 years), participated. Controls were younger than subjects with JIA. Demographic, clinical, and disease factors are shown for subjects with JIA in Table 1. One subject with JIA and no subjects in the control group were obese. Lower extremity joints involved in subjects with JIA within the year before testing included the knee (20), ankle (15), foot (11), and hip (7). All but 1 subject had multiple joints involved. Median joint count on the day of testing was 0 (0–2), and only 5 subjects had active arthritis of the lower limb at the time of testing. Median articular severity index score was 2 (0–26). Median CHAQ score was 0 (0–1.5) and median VAS pain score was 5 (0–73). All but 1 subject were on medication for arthritis. Thirteen (52%) were on nonsteroidal anti-inflammatory drugs, 17 (68%) were on disease-modifying antirheumatic therapy, 4 (16%) were on prednisone, and 7 (28%) were on biologic therapy.
Single-leg balance was impaired in many subjects with JIA. Ten subjects (40%) were unable to complete any of the 3 trials. One subject could not complete the test on the right leg, 6 on the left leg, and 3 were unable to complete the test on either leg. Ninety-two percent were right-side dominant; a left-side dominant subject accounted for the incomplete test on the right leg. All of the controls were able to successfully complete the single-leg balance test on at least 1 trial. Comparison of each of the 3 trials (completed vs not completed) between the 2 groups is shown in Table 2.
Postural stability results are shown in Table 3. One child, an 11.9-year-old girl with polyarticular rheumatoid factor negative disease, was unable to complete BBS level 2 (very unstable). She had disease duration of 14 months and a history of arthritis affecting both knees, ankles, subtalar joints, and some metatarsophalangeal joints. She did not have active arthritis on the day of testing. The AP, ML, and OA stability indices in the JIA group were significantly higher than controls at level 2 (very unstable). There was no significant difference between the JIA group and control group for bilateral static balance or dynamic balance at level 7 (moderately unstable).
Correlations between strength and balance are shown in Table 4. Isometric hip abduction and hip flexion strength had a low correlation with single-leg balance. Hip abduction, hip flexion, knee flexion, and ankle plantar flexion strength had a low correlation with bilateral balance.
Further analysis of subjects with JIA revealed no difference in balance at BBS level 2 for those with or without pain (VAS > 0, VAS = 0, respectively), with or without functional impairments (CHAQ >0, CHAQ = 0, respectively), and with or without active joints on the day of testing. Correlation analyses for BMI were not done as only one subject with JIA had a BMI outside of the healthy weight range.
To our knowledge, our study is the first to report on a comprehensive assessment of balance in children with JIA and healthy controls. We found that compared with healthy controls, children and adolescents with JIA and history of lower extremity arthritis in the last year had gross impairments in single-leg balance and mild impairments in bilateral dynamic balance.
Single-leg balance was significantly impaired, with 40% of the children with JIA unable to stand on one leg on a stable surface for 20 seconds with eyes open. Errors or loss of balance were common in both groups and diminished over the trials in the control but not in the JIA group, suggesting a training or learning effect in the controls. Few studies describe single-leg balance in active adolescent populations. Normative data suggest that adolescents can stand on a single leg on a static surface with eyes closed for 25 seconds.26 Most balance tests include eyes-open dynamic and eyes-closed static and dynamic conditions; no reference data are available for single-leg standing with eyes open. In a case report on balance in juvenile arthritis, a 10-year-old girl successfully completed single-leg balance tests on BBS level 4 (moderately unstable).27 Two previous reports of single-leg balance measures in children with juvenile arthritis have been published.18,19 Neither study had a control group nor was balance the primary outcome. Andre19 reported improvement in single-leg balance in 48 children with foot arthritis after wearing orthotics, and Bacon and colleagues18 reported no improvement in single-leg balance after a 6-week aquatic exercise program in 11 children with juvenile RA.
We found mild impairments in bilateral dynamic balance tested with BBS at level 2 (very unstable) and no impairments in bilateral static balance (level 12) or dynamic balance at level 7 (moderately unstable).
Our findings are consistent with studies of adults with RA that report increased postural sway and decreased balance.14,28 The BBS has been used in 2 studies to evaluate postural control in adults with arthritis: one study in ankylosing spondylitis and one in RA.14,29 In a study of 70 patients with ankylosing spondylitis, no effect on postural balance was found,29 whereas a study of 74 patients with RA found significant deficits in postural stability compared with healthy controls.14 In the study of patients with RA, balance was tested at level 8 (most stable; static level on BBS with levels 1-8) and level 2 (more unstable). Both the OA and AP indices were higher for the RA group for static balance. The static stability OA scores in our cohort of children with JIA (0.612 ± 0.440) are better than those of adults with RA (2.7 ± 0.9) and adults with ankylosing spondylitis (3.29 ± 1.54). Static stability OA scores for adults who are healthy vary in the literature (2.2 ± 0.7; 3.08 ± 1.18). No reference data for static stability scores in children and adolescents are available for comparison, and in this study, we found no difference on static bilateral stability testing between our subjects with JIA and a small control group. Dynamic balance (level 2) was significantly impaired in the RA group, with 15.9% of patients compared with 7.1% of controls unable to complete the test.14 In our study, only 1 child, an 11.9-year-old girl with polyarticular rheumatoid factor negative disease, was unable to complete BBS testing at level 2.
We found correlations between muscle weakness as measured by MMT and poor balance in children with JIA. Studies in adults have shown that as lower extremity strength decreases, the ability to balance decreases as well.30,31 Furthermore, poor proprioception increases the functional consequences of weakness.31,32
In children and young adults with hemophilia, greater muscle strength around affected joints may help protect joints from hemarthrosis, increase joint stability, and reduce injury risk.15,17 Indeed, prophylactic physical therapy focused on improving periarticular muscle strength was demonstrated to reduce the frequency of hemorrhage.16
We found that isometric hip abduction and hip flexion strength had a low correlation with single-leg balance; and hip abduction, hip flexion, knee flexion, and ankle plantar flexion strength had a low correlation with bilateral static and dynamic balance. Significant muscle weakness has been reported in children with JIA, with weakness most pronounced in the muscles surrounding the active joints.33 Knee extension has been most commonly studied, and all studies have reported decreased isometric knee extension strength and torque.33–35 Lindehammar and Backman33 found reduced isometric muscle strength in knee extensors in 20 children with juvenile chronic arthritis. Strength was found to be 45% to 65% of the expected value in muscles near an inflamed joint and only slightly decreased (80%-90% of the expected value) in muscles not adjacent to active arthritis. Isometric and isokinetic ankle dorsiflexor strength were reduced only with ankle arthritis. Brostrom and colleagues36 reported on ankle strength in 10 girls with polyarticular JIA and 10 controls. The girls with JIA had lower ankle plantar flexor and dorsiflexor isometric strength, lower concentric plantar flexion torque, and lower dorsiflexion eccentric and concentric torque than children in the control group. All children were stronger in plantar flexion than dorsiflexion. Hedengren and colleagues34 found lower isometric knee extension, ankle plantar flexion, and dorsiflexion torque in 11 children with juvenile chronic arthritis. The fact that we did not find knee extension deficits to be correlated with decreased balance may be because of our use of manual isometric muscle testing, whereas others used measures of isometric torque or isokinetic strength.33–36
We found no differences in balance scores between children with and without pain, functional impairment, or active disease. In contrast, a positive correlation between balance, measured with the BBS, and function as measured by the Health Assessment Questionnaire is described in adults with RA.14 The CHAQ has a recognized ceiling effect, and better discriminatory measures of function are likely required to identify a significant correlation between function and the relatively mild impairments in dynamic balance shown in this study. Sixteen of our patients (64%) had a CHAQ score of 0, but none of our patients reported optimal sport and activity participation on direct questioning. Similar to our results, postural balance was not affected by disease activity in patients with RA.14
Increased rates of injury, including ankle sprains and anterior cruciate ligament knee injuries, are reported in both high school and collegiate athletes with poor single-leg balance.37–39 Balance training is a key component of sports injury rehabilitation, and recent studies support proprioception training as a strategy to reduce sports-related injuries in children and adolescents.40–43 Proprioceptive training may also decrease joint damage and improve athletic performance in children and young adults with hemophilia.15,16 It follows that proprioceptive exercises may be of benefit to children with JIA, particularly those with poor single-leg balance. Myer and colleagues27 described successful specialized neuromuscular training to improve biomechanics and enhance safe sport participation in a 10-year-old girl with quiescent juvenile RA. Her single-leg balance stability index score improved (right, 56%; left, 46%) after a twice weekly 5-week neuromuscular training program.27 The authors suggest that normalization of biomechanics and teaching correct sport-specific techniques may help children with JIA participate in sport with decreased injury risk. This approach expands upon current recommendations promoting moderate fitness and strengthening exercises for children with arthritis.44 Balance and activity-specific tasks (eg, soccer-specific movements for a child playing soccer) are not routinely assessed, and it is the authors' opinion that such an assessment and subsequent directed rehabilitation may better prepare children with JIA for safe sport participation.
This study has several strengths and limitations. First, we used a standard objective method for measuring balance in children with JIA and compared their results with those of subjects who were healthy and of similar age and sex distribution. We believe that this is appropriate in the absence of reference values for balance measurement in children. Our control group was younger than the subjects with JIA, and as balance improves with age, the findings of balance deficits in children with JIA are even more striking.
Second, the subjects with JIA who volunteered for this study might not represent all children with JIA. For example, relative to our general JIA clinic population, our group had more boys and greater use of biologic agents and corticosteroids.
Third, the cross-sectional study design does not allow assessment of the course of the balance impairments and its temporal relations with disease activity and arthritis treatments.
Fourth, we did not formally assess habitual physical activity or sport participation and therefore cannot comment on any relationships between these and the observed balance impairments.
And finally, we used MMT to examine isometric strength. Manual muscle testing is prone to both intra and interrater variability.45 To minimize this variability, all measures were assessed by the same clinician. Manual muscle testing is a practical measure of muscle strength in the clinic setting. However, use of isokinetic strength testing protocols may allow better quantification of muscle strength and endurance and identification of specific muscles related to balance deficits.
Despite these limitations, we were able to demonstrate, for the first time, significant balance deficits in a small sample of children with JIA. The majority of children in this study had well-controlled disease, and we hypothesize that children with uncontrolled disease and active lower extremity arthritis may have greater impairments in balance. A relationship between upper extremity and/or spinal arthritis and balance dysfunction has not been described, but should be explored in future studies. Future longitudinal studies in larger samples, incorporating standard balance measurement, isokinetic muscle strength testing, and measurement of physical activity and sport participation will be better able to characterize the effect and correlates of the balance impairments shown in this study. Such studies will be most informative in designing balance rehabilitation interventions for children with JIA and evaluating their effectiveness.
The fact that single-leg static balance was grossly impaired in many of our subjects suggests that this might be a good screening test in clinical settings. Those patients who fail to balance for 20 seconds after 3 attempts could then be referred for detailed assessment. Alternatives for clinical screening is the timed eyes-closed dynamic testing (testing on a foam surface) proposed by Emery and colleagues26 or the use of the Balance Error Scoring System. The Balance Error Scoring System is used in standard assessment of concussion and includes three 20-second tests with different stance conditions (double leg, single leg, tandem; all conditions with eyes closed).46,47 The use of Balance Error Scoring System has also been proposed for screening collegiate athletes for postural deficits following lower extremity injury.48
To our knowledge, our study is the first to report a detailed assessment of balance in a group of children with JIA. We found that compared with healthy controls, children and adolescents with JIA and lower extremity involvement in the previous year had grossly impaired single-leg balance and mildly impaired bilateral dynamic balance. Quantifiable impaired balance may serve as a new measure of disability in children with arthritis and provide reference for the design of proprioceptive exercises as an integral part of therapy in the treatment of children with JIA. Improved proprioception in children with arthritis may play a role in improving balance and strength, thereby decreasing risk of injury, improving quality of life, and facilitating full activity and sport participation.
A significant proportion of children with leg arthritis have impaired balance. Proprioceptive exercises may emerge as an important therapy in the treatment of children with JIA and lower extremity arthritis.