Cerebellar ataxia

ABSTRACT

Background and Purpose: Weighting the axial skeleton has been recommended to improve gait for patients with ataxia. In this study the effect of axial loading on gait of individuals with cerebellar ataxia was examined. Subjects and Methods: Five subjects with ataxic gait ambulated 9.76 meters for 5 trials in each of 4 conditions: an initial unweighted phase, with weight at the shoulders, with weight at the waist, and a final unweighted phase in a repeated ABCA single-subject design. Footfall characteristics were recorded for 3.66 meters of each trial using a GAITRite" mat. The two standard deviation band method was used to analyze 6 gait parameters: velocity, cadence, standard deviation of step time, standard deviation of step length, width of the base of support, and double stance time. Results: Gait characteristics changed unpredictably in the 4 conditions, worsening more often than improving. Discussion: Effects of axial loading on gait were inconsistent. Conclusion: Using axial weights to improve gait for patients with ataxia was not supported.

Key words: ataxia, axial loading, cerebellar dysfunction, gait disturbances, weights

INTRODUCTION

Ataxia is the loss of ability to control and coordinate voluntary movements following damage to the sensory or cerebellar system.1 Cerebellar ataxia may be due to tumors, disease, cerebral vascular accident, or a number of other pathologies.2 The cerebellum is thought to play a major role in the preparation of planned voluntary activity and the monitoring of ongoing movement.3 Disorders of the cerebellum and its efferent or afferent pathways results in deficits in the rate, range, and force of voluntary movement such as dysmetria, dyssynergia, or dysdiadochokinesia. In addition, the initiation and accuracy of movement patterns is impaired, resulting in uncoordinated movements and movement decomposition. Persons who have cerebellar lesions often experience delay in the initiation of movement, as well as slowed and irregular movement and high amplitude tremor.3 A wide base, unsteadiness, and irregularity of step length and width, absent rhythm and feet lifted too high with no reciprocal arm swing characterize a cerebellar ataxic gait.3 Gait disturbances may result from disruption of the rhythmic discharges of the rubrospinal, reticulospinal, and vestibulospinal tracts or may be produced by errors in the programming of rate and force of muscle contraction. Loss of the ability to modulate the long-loop postural reflexes effectively also contributes to difficulties in gait.

There is no specific medical treatment that is effective for ataxia, so management focuses on symptomatic relief and the promotion of function through the use of adaptive devices and therapeutic exercise.3-5,7 For many years, physical therapists have used weights, either on the limbs or on the axial skeleton as an intervention to assist patients to control extraneous movements.5,6,8,9 The approach probably originated with the work of Margaret Rood. Using a neurodevelopmental framework emphasizing the effect of sensory input on movement, Rood theorized that afferent impulses from deep pressure receptors from joint compression exceeding that normally superimposed by body weight would be stimulated by loading, reflexively facilitating cocontraction of the stabilizing muscles around the compressed joints.8,9 Weighted vests or berets were recommended to facilitate cocontraction around the joint and thereby increase the patient's stability.8,9 Since a primary problem in many neuromuscular conditions including ataxia was thought to be a lack of sufficient cocontraction and stability, weights were recommended to augment the patient's ability to stabilize posture. Using similar reasoning, Morgan theorized that weights might increase the patient's sense of position by providing augmented feedback.5

As Rood's neurodevelopmental approach fell out of favor, motor control explanations were advanced to support the practice of weighting. Using motor control theories, author's suggested that adding mass in the appropriate amounts and at the appropriate locations might increase the stability of the body or body part biomechanically by changing the moment of inertia, thus reducing the severity of the ataxia or tremor and slowing the movement to allow additional reaction time.5-7 Clearly, neither of these theoretical approaches is very detailed or systematic and the authors were forced to conclude that there may be no clear or compelling rationale for using weights to improve gait in ataxic patients.

Although the approach is an old one, one of the current major texts used in many physical therapy programs recommends weights as one technique to be used in the management of patients with ataxia, suggesting that the practice is still in common use.3 At the 1999 Annual Conference & Exposition of the American Physical Therapy Association, Dr. Patricia Montgomery called for further research into the use of weights as a treatment alternative for children with ataxia.10

Results of the 3 studies in which the use of weights to decrease ataxia were examined arc variable.5,6,11 Findings by Morgan' were generally positive. He studied the effect of weighting the upper extremity on reaching movements in 58 patients with incoordinated arm movements from a variety of causes.The patients included at least 24 with cerebellar pathologies and 16 with pathologies that may or may not have affected the cerebellum. He reported that there was a specific weight that had the greatest effect on the ataxia for each individual. He also described the effects of weights on hips, thighs, and ankles to reduce ataxic gait, reporting his subjective judgement that I out of 14 patients improved and the more ataxic patients required more weight.

In contrast findings by 2 other investigators were not in favor of using weighting to decrease ataxia. Lucy and Hayes,6 who investigated the clinical effectiveness of using a 1.36 Kg weight across the shoulder to reduce oscillations in the center of pressure of subjects with ataxia and healthy controls, reported highly variable results. Some patients were benefited but others became more destabilized. Manto, Godaux, and Jacquy11 investigated the effects of placing weight of 200 and 500 gm on the MCP joints of patients with cerebellar dysfunction for coordination of agonist and antagonist muscle activity for wrist flexion and extension. The investigators found that in patients with cerebellar dysfunction, the timing of agonist/antagonist activity was disrupted with the addition of weights to distal joints, and therefore, accuracy of movement was compromised. They concluded as well, that use of distal loading may be more detrimental to stability than axial loading. Based on these findings, Bastian, has cautioned clinicians not to weight the limbs of patients with ataxia.

Guidance from published research regarding the clinical use of weights as an intervention technique for patients with ataxia is fragmentary and, for the most part, very old. One common suggestion, however, is that axial weighting may be superior to distal weighting in providing stability. In addition, none of the studies directly used objective measures of the effect of weight to effect improvement of ataxic gait. Therefore the purpose of this study was to determine whether the addition of 10% of body weight to the axial skeleton at the waist or at the shoulders would improve of gait parameters in subjects exhibiting ataxic gait patterns. We anticipated that an improved gait would be indicated by increased velocity and cadence, and decreased width of the BOS, double stance time, variability of step length, and variability of step time.

METHODS

Subjects

Five ambulatory individuals with ataxic gait patterns participated in 5 repeated single-subject design investigations. Three subjects were recruited by contacting area physical therapists to suggest patients who have ataxia from their caseload who met the criteria of being ambulatory and able to cooperate with the research protocol. Three subjects were recruited through area therapists who referred patients that met the criteria of being ambulatory and able to cooperate with the research protocol. Patient characteristics are reported in Table 1. Subjects with ataxia were designated Subect 1, 2, 3, 4, and 5 consecutively by age with Subject I being the youngest and Subject 5 the oldest. The patient history, family report, and information supplied by referring professionals all supported a diagnosis of cerebellar ataxia, but physician records were not available. The diagnosis of cerebellar injury was reported to have been confirmed by MRI in only 3 of the 5 cases. None had an acute lesion and all were community ambulators. None wore an orthoses or used a gait device. Presence of an ataxic gait pattern was confirmed by a physical therapist experienced in treating patients with neurological deficits, using the criteria of wide base of support, unsteadiness, and irregularity of steps and foot placement.

For a pretest, 5 subjects with no known gait deviations were recruited in a purposive sample from acquaintances and family of the investigators. The nondisabled subjects were paired to resemble the experimental subjects in age, weight, and gender. Nondisabled subjects were tested to validate the criteria used to measure improvement or reduction in the ataxic gait pattern and to investigate the effect of weighting on gait for persons with no gait deviation. Each subject or guardian signed an informed consent document approved by the TTUHSC Institutional Review Board.

Gait Analysis

A computerized GAITRite" system from CIR Systems (CIR Systems-GAITrite, Clifton, NJ) was used to analyze the gait patterns of the subjects. The GAITRite system provided an automated method to measure the spatial and temporal parameters of gait.12 The active area of the GAITRite walkway is 61 cm (24 inches) X 366 cm (144 inches) with a grid of 48 X 288 sensors on 1.27 cm (0.5 inch) centers. As the subject ambulated along the walkway, the sensors detected and recorded each footstep. The mat recorded the specific activated sensors, the distance between the activated sensors, and the time of activation/deactivation. Sensitivity of the sensors could be adjusted for the subject size so that even the footfalls of young or light subjects could be detected. The electronic walkway transferred this information to a nearby personal computer via the interface cable. Application software processed the raw walkway data into footfall patterns that were printed out for visual inspection, and also computed temporal and spatial parameters including velocity, cadence, step time, step length, width of the base of support, and double and single stance time, along with additional variables not analyzed. Because the subjects with ataxic gait sometimes staggered or stepped off the mat, the GAITrite sometimes identified the right foot as the left or vice versa. These errors were corrected by hand editing of the data. Selby-Silverstein and Besser13 compared the GAITRite system of analysis to paper measures of footfall imprints and concluded that the GAITRite's reliability for spatial parameters such as step time and step length is acceptable. Unpublished preliminary data comparing the GAITRite Gold to a gait laboratory analysis suggests agreement in the range of approximately r = .93, suggesting a coefficient of variation of approximately .80 for spatial parameters.14 (Since the sampling rate is higher in the GAITRite Gold, the data would not be representative for temporal parameters.) Reliability of the GAIT Rite for individuals with ataxia has not been specifically tested. Weighted Vests

Velvasoft(TM) vests used to secure the weights on smaller patients were provided by M.W Sales and Service Inc. (M.W Sales and Service Inc., San Antonio, Tex) The 10% of body weight amount used in this study was based on recommendations that come with Velvasoft vests,15 because therapists may use such recommendations to choose the amount of weight to apply when working with a patient. Weight was determined by asking the patient or family. The vests were made of velvet-textured fabric to which variable amounts of weight (increments of 1/8 lb, 1/4 lb, and 1/2 lb) were attached by Velcro(TM) at the desired location on the shoulders or at the waist. If the weight of the subject was over 63 kg, the Velvasoft(TM) vest would not hold 10% of body weight. For these subjects, the weights were attached to the shoulders by encasing a weight that was 5% of the subject's body weight in each of 2 pieces of Stockinette(TM) (for the total of 10% needed). These weights were then placed on the subject's shoulders and the Stockinette was tied under the opposite arm to secure the weight for the phase with weight above the waist. In the condition with weight at the waist, the weights in the cloth tubing were balanced in front and back and tied around the subject at the waist.

Procedure Each subject walked a distance of 3.05 m (10 ft) before reaching the GAITRite mat, the length of the GAITRite mat 3.66 m (12 ft), and another 3.05 m (10 ft) for a total of 9.76 m (32 ft) a minimum of 20 times. The subject walked this distance 5 times with no weight as a baseline, 5 times with the weights on the shoulders, 5 times with the weights at the waist, and, finally, 5 times without weight to assess carryover effects, in an ABCA single subject design. Three subjects tugged at the vest while walking, possibly indicating a negative reaction to the weight and showed signs of fatigue or mild distress. To minimize fatigue, subjects had a 1 to 2-minute rest between trials. Each subject completed all trials in one session of testing, which lasted 1 to 2 hours. It was necessary to repeat trials for many subjects due to equipment error or noncompliance by the subject. Data from those trials were discarded. Rest breaks were allowed between trials when the subject appeared to be fatigued. Young subjects required coaching and demonstration. Small incentives such as snacks, play, or inexpensive toys were provided to children and Subject 4 based on the advice of the guardian.

Subject 4 had difficulty staying on the mat long enough for it to measure the parameters of gait.This was due to the restricted width of the mat and his severe staggering. As a result of this difficulty, it was necessary to have the subject repeat many trials and to have another person walking beside him to assist him to stay on the mat by limiting his deviation from the mat using gentle nudges at the waist or shoulders. This intervention may have affected his performance, but without this assistance, no data could have gathered. Approximately equal assistance was required for each of the 4 phases of testing, so that the results should allow a reasonable comparison of the 3 conditions.

Data Analysis

The variables chosen for analysis were velocity, cadence, standard deviation (SD) of step time and of step length, width of the base of support (BOSS, and double stance time (% of gait cycle). Standard deviation was used to indicate the degree of variability of step time and step length. Improvement was indicated by an increase in cadence or velocity, or a decrease in the other 4 measures.The two-standard-deviation band analysis for single-subject designs was used to determine whether there were differences in the subjects' gait characteristics between the initial unweighted condition and the weighted and carryover conditions.6 The mean and standard deviation of data points in the baseline phase are calculated. Lines are then drawn two standard deviations above and below the mean and extended into the intervention and carryover phases. If 2 or more successive data points in the intervention or carryover phase fall outside the two standard deviation band, change from baseline to intervention or carryover is considered significant.

In the pretest, data were analyzed for each nondisabled subject in the same manner as was done for the subjects who had an ataxic gait. Paired t-tests were used to compare the 2 groups (subjects with ataxia and those with no gait deviation) on each gait characteristic for the initial unweighted phase to assess the appropriateness of the chosen gait characteristics to distinguish ataxic from typical gait. Bonferoni correction was applied to establish an alpha level of .008 for the t-tests. To investigate inter-trial variability, intra-class correlation coefficients were computed using model 3 as described by Portney and Watkins" for each variable for the first 5 trials for the nondisabled subjects.

RESULTS

Inter-trial Variability

Intra-class correlation coefficients for the various gait parameters investigated were respectively: velocity ICC (3,1) = 0.80, cadence ICC (3,1) = 0.78, step time ICC (3,1) = 0.92, step length ICC (3,1) = 0.98, width of the base of support ICC (3, 1) = 0.65, percent of gait cycle in double support ICC (3, 1 ) = 0.41.

Healthy Individuals

The single-subject analysis for nondisabled subjects showed inconsistent changes between weighted and unweighted trials, with more changes in the direction of deterioration than of improvement. Means for each variable chosen differed between the subjects with ataxia and the nondisabled subjects in the expected direction for all gait characteristics measured in the baseline condition (See Table 2). Mean velocity and cadence were lower for the group with ataxia than for the group with no gait deviation. The SD of step time and step length, width of the BOS, and double stance percent of the gait cycle all were greater for the group with ataxia than for the group with no gait deviation. However, of the 6 gait characteristics, the difference was significant only for SD of step length (t = 10.59, p

Patients with ataxia

Results of the single-subject studies for subjects with ataxia are illustrated in Figures 1-5. Visual inspection of the graphs does not reveal any obvious consistent trends in the data. Statistically significant changes from baseline are reported below.

Subject 1 decreased SD of step length with weight on the shoulders (improvement) and increased double stance time with weight at the waist (deterioration).

Subject 2 decreased velocity in both weighted conditions and carryover, increased SD of step time with weight at the waist, increased SD of step length with weight on the shoulders, and increased double stance time with weight at the waist (all deterioration).

Subject 3 increased velocity with weight on the shoulders and increased cadence with weight at the waist (improvement).

Subject 4 increased double stance time in both weighted conditions (deterioration).

Subject 5 had increased velocity with weight at the waist and in carryover (improvement). DISCUSSION The use of weights in these 5 patients with cerebellar ataxia produced variable responses in the parameters of gait that were measured. Although the response of subjects with ataxia to axial weighting was quite variable, deterioration of gait characteristics appears to occur more often than improvement. In addition, in the few cases where improvement was seen, the improvement was inconsistent. No consistent differences were seen related to weight placement on the shoulders or at the waist.

The gait parameters chosen were not significantly dif ferent for the pretest subjects and those with ataxia in this small sample, with the exception of SD of step length, but the fact that all 6 parameters varied in the direction expected does not contradict the assumption that the variables are appropriate to indicate whether the gait pattern is showing increased or decreased ataxia.

Reliability of the selected gait variables has not been reported for individuals with ataxia. Reliability was good for step length, step time, velocity, and cadence and poor reliability for the percent of the gait cycle in double support and the width of the base of support." Reliability for step length and step time for this study exceed that recommended for most clinical instruments." It would be incorrect to assume that the ICC reflects only measurement error, however, since it is likely that there was actual variation in gait parameters from one trial to another.

The decision to use 10% of body weight was arbitrary, based upon recommendations of the manufacturer of the vests used in the study. It is likely that the amount of weight used in the current study exceeded the weight used in the Morgan' study of gait and definitely exceeded the weight used in the Lucy and Hayes(' study of static standing posture. Although Morgan does not report age or weights, the diagnoses reported suggest adults. Morgan used weights of 2 to 3.5 kg. The average weight of patients in the Lucy and Hayes study was 70.1 kg and weights of 1.36 kg (approximately 2% of body weight) were used. It is possible that less weight may have had a more beneficial effect for gait as suggested in the Morgan study and that the heavier weights used in the current study may have had a more variable or negative impact on gait characteristics. In addition, improvement in static balance may not translate into improvement in dynamic gait.'- Finally, negative reaction of some patients suggests that the weight chosen may have been too heavy. Nevertheless, there does not appear to be any support from this study for the beneficial effects of axial weighting for the gait of patients with ataxia.

An incidental finding suggesting that the SD of step length may be an effective measure to differentiate ataxic from typical gait may, however, be useful clinically. The findings from this small sample suggest that the SD of step length for patients with ataxia may be almost twice that of persons with no gait deviation. If true, a simple and inexpensive ink-footprint pedograph18 could be used to monitor severity or improvement in gait for patients who have ataxia by measuring the step length and computing and recording the SD. This observation should, however, be replicated and validated on a larger sample.

Some of the limitations of the study were overcome by design, but others were not. Recruiting a large number of ambulatory patients with cerebellar ataxia who could cooperate with the research protocol was difficult. The singlesubject design compensated for this difficulty by using each subject as her or his own control. Subject's fatigue was a limitation. If the study is repeated, it might be beneficial to spread the trials out over a longer time period. Rest periods were provided for the subjects, but may not have been sufficient for full recovery. No consistent deterioration of performance in later trials was observed, however, suggesting that fatigue did not prevent gathering meaningful data.

An arbitrary baseline of 5 trials rather than continuing until there was a stable baseline in the initial condition may have led to misinterpretation of the data. Limiting the baseline to 5 trials meant that for some variables for some subjects, there was not a stable baseline. Subjects who required repeated trials may have encountered a practice effect, although there is no trend in the data to suggest such an effect. Tactile guidance may have changed the gait of the subject who required assistance to stay on the mat. Finally, the width of the GAITRite mat of 61 cm also posed a problem in that some patients with ataxia had difficulty staying on the mat. It is, however, difficult to think of a way to collect gait parameters that does not involve walking in a restricted path.

The results of this investigation found inconsistent support for the practice of using axial weighting for patients with ataxia. A thorough review of the literature attempting to find a detailed rationale for the use of weights to improve function for patients with ataxia revealed only a very limited theoretical basis for the intervention, as discussed previously. Articles presenting research evidence supporting the practice are also surprisingly weak given the persistence of the recommendations suggesting that the approach is useful.

CONCLUSION

No consistent effects of axial loading were found for the gait of persons with ataxia, although deterioration of gait characteristics appeared to be slightly more prevalent than improvement. The use of axial loading as an intervention technique to improve gait characteristics for patients with ataxia is not supported by these data. The theory that weights may improve movement coordination by increasing proprioceptive input or by slowing ataxic movement was not supported.

For these 5 subjects with ataxia, SD of step length differentiated their gait from the gait of subjects with no gait deviation, suggesting that SD of step length may be clinically useful data for assessing ataxic gait. Further investigation on a larger sample would be needed to confirm this observation.

ACKNOWLEDGEMENTS

This study has been previously presented as a poster at the Texas Physical Therapy Association Annual Conference in Austin, TX, October 6-7, 2000, and the American Physical Therapy Association Combined Sections Meeting in San Antonio,TX February 17, 2001.

This study was approved by the Institutional Review Board for the Protection of Human Subjects of Texas Tech University Health Sciences Center and University Medical Center.

We would like to thank the following people and organizations who helped make this research possible: Institutional support from the Physical Therapy Program at Texas

Tech University Health Sciences Center; Steve Sawyer, FIT, PhD; CIR Systems (GAITRitel system); M.W Sales and Services Inc. (Velvasoft(TM) weighted vests); Georgia Blessey, PT, PCS; Kathryn Goodwyn, PT; Tammy Quisenberry, PT; the subjects, parents, and families who participated in the study.

REFERENCES

Bastian AJ. Ataxia:"If Only I Felt Steadier." PT Magazine. September, 1998:62-68.

Adams RD, Victor M, Ropper AH. Principles of Neurology. 6th ed. Companion Handbook. New York, NY: McGraw-Hill Co Inc; 1998:43,56-57.

Melnick ME, Oremland B. Movement dysfunction associated with cerebellar problems. In: Umphred DA, ed. Neurological Rehabilitation. 4th ed. St. Louis, Mo: Mosby-Year Book Inc; 2001:717.

Bastian AJ. Mechanisms of ataxia. Phys Ther. 1997;77: x,72-X75

Morgan MH. Ataxia and weights. Physiotherapy. 1975; 61:332-334.

Lucy SD, Hayes KC. Postural sway profiles: Normal subjects and subjects with cerebellar ataxia. Physiotherapy Canada. 1985;37:140-148.

Montgomery PC. Achievement of gross motor skills in two children with cerebellar hypoplasia: Longitudinal case reports. Pediat Phys Ther 2000; 12:68-76.

Goff B. The application of recent advances in neurophysiology to Miss M. Rood's concept of neuromuscular facilitation. Physiotherapy. 1972;58:409-415.

Stockmeyer S. An interpretation of the approach of Rood's treatment of neuromuscular dysfunction. Am J Phys Neurol. 1967;62:937.

10 Montgomery PC. Clinical considerations in childhood ataxia. presented at:Annual Conference & Exposition of the American Physical Therapy Association; June 6, 1999; Washington, DC.

Manto M, Godaux E, Jacquy J. Cerebellar hypermetria is larger when the inertial load is artificially increased. Ann Neurol. 1994;35:45-52.

12 GAITRitel system [user's guide]. Clifton, NJ: CIR Systems; 1999.

13 Selby-Silverstein L, Besser M. Accuracy of the Gaitrite system for measuring temporal-spatial parameters. Phys Then. 1999;79:S59.

14 Karakostas T. Personal Communication, Sept. 10, 2002. 19 Velvasoft(TM) [brochure insert]. San Antonio, TX: M.W Sales and Services Inc.; 1999.

16 Portney LG, Watkins MP Foundations of Clinical Research:Application to Practice. Norwalk, Conn:Appleton & Lange; 1993:222-223.

Winstein CJ, Gardner ER, McNeal DR, et al. Standing balance training: Effect on balance and locomotion in hemiparetic adults. Arch Phys Med Rehabil. 1989;70: 755-762.

18 Ogg HL. Measuring and evaluating the gait patterns of children. Phys Ther 1963;43:717-720.

Nancy Clopton, PT, PhD1 Dana Schultz, MPT2 Catherine Boren, MPT2 Jennifer Porter, MPT2 Tandra Brillhart, MPT2

1Associate Professor, Physical Therapy Program, Department of Rehabilitation Sciences. Texas Tech University Health Sciences Center, Lubbock, TX

2Recent graduates of the Physical Therapy Program, Texas Tech University Health Sciences Center, Lubbock, TX

Copyright Neurology Report Mar 2003
Provided by ProQuest Information and Learning Company. All rights Reserved