Gait and Functional Mobility Deficits in Fragile X-Associated Tremor/Ataxia Syndrome

Abstract

Fragile X-associated tremor/ataxia syndrome (FXTAS) results from a “premutation” (PM) size CGG repeat expansion in the fragile X mental retardation 1 (FMR1) gene. Cerebellar gait ataxia is the primary feature in some FXTAS patients causing progressive disability. However, no studies have quantitatively characterized gait and mobility deficits in FXTAS. We performed quantitative gait and mobility analysis in seven FMR1 PM carriers with FXTAS and ataxia, six PM carriers without FXTAS, and 18 age-matched controls. We studied four independent gait domains, trunk range of motion (ROM), and movement transitions using an instrumented Timed Up and Go (i-TUG). We correlated these outcome measures with FMR1 molecular variables and clinical severity scales. PM carriers with FXTAS were globally impaired in every gait performance domain except trunk ROM compared to controls. These included total i-TUG duration, stride velocity, gait cycle time, cadence, double-limb support and swing phase times, turn duration, step time before turn, and turn-tosit duration, and increased gait variability on several measures. Carriers without FXTAS did not differ from controls on any parameters, but double-limb support time was close to significance. Balance and disability scales correlated with multiple gait and movement transition parameters, while the FXTAS Rating Scale did not. This is the first study to quantitatively examine gait and movement transitions in FXTAS patients. Gait characteristics were consistent with those from previous cohorts with cerebellar ataxia. Sensitive measures like the i-TUG may help determine efficacy of interventions, characterize disease progression, and provide early markers of disease in FXTAS.

Introduction

Fragile X-associated tremor/ataxia syndrome (FXTAS) is a progressive neurological disorder resulting from a “premutation” size 55–200 CGG repeat expansion in the 5′ untranslated region of the fragile X mental retardation 1 (FMR1) gene located on the X chromosome [1, 2]. The characteristic symptoms of FXTAS include cerebellar gait ataxia, kinetic tremor, peripheral neuropathy, and cognitive decline and/or dementia [1–4]. There is high phenotypic variability, with gait deficits presenting as a predominant feature in a subset of FXTAS patients resulting in progressive mobility disability and increased risk of falls [5]. Thus far, no studies have quantitatively characterized gait and movement transition deficits in patients with FXTAS. Understanding these deficits is crucial for early disease detection, therapeutic management, and more fully characterizing the natural history of FXTAS.

The traditional stopwatch-timed Timed Up and Go test (TUG) is a clinically important functional mobility task used to assess fall risk [6], but it does not capture specific gait or mobility deficits during task performance. The newer instrumented Timed Up and Go (i-TUG) utilizes body-worn inertial sensors which allow determination of the spatiotemporal aspects of gait and quantitative measures of functional movement transitions including sit to stand, turning, and turn to sit [7]. Such methods are increasingly being utilized in movement disorders such as multiple sclerosis (MS) [8] and Parkinson’s disease (PD) [9, 10], in which multiple i-TUG variables correlated with disease severity and were identified as possible markers of disease progression. While traditional marker-based approaches are the gold standard in gait analysis, high correlations between spatiotemporal aspects of gait measured with these and the inertial sensor approach have been reported [11, 12]. In addition, traditional systems require precise placement of the feet on a force plate, a difficult task for persons with ataxia, and they typically do not record enough steps for accurate measurement of gait variability [13], a key feature of ataxia [14]. The inertial sensors allow for gait analysis that is less costly, time-consuming, and invasive to the patient, and they are able to capture additional measures which are important for the assessment of gait stability and known to be affected in ataxic patients, including trunk acceleration and displacement [15].

Presently, there are no optimized treatments for patients with FXTAS. It is critical that the specific gait and mobility impairments in FXTAS be quantitatively characterized, as this will allow for design of tailored rehabilitation and new, targeted treatment interventions, and aid in the determination of disease severity and progression. Therefore, we performed this study to characterize gait patterns and mobility deficits in FMR1 premutation (PM) carriers with and without FXTAS compared to healthy age-matched controls. PM carriers without FXTAS were included to investigate whether inertial sensor gait analysis could detect any early signs of disease in those PM carriers who are clinically normal on a neurologic assessment.

Subjects and Methods

Subjects

Thirteen PM carriers were recruited from the Fragile X-Associated Disorders Program at Rush University Medical Center (RUMC) and were part of a larger ongoing neurological phenotype study in PM carriers. Most were identified through family members of known children with fragile X syndrome. A detailed medical history, neurological examination, and MRI, when available, were used to diagnose FXTAS according to clinical and radiological criteria [2]. Inclusion criteria were (1) FMR1 CGG repeat size of 55–200 (for PM carrier subjects) and (2) ability to ambulate independently in the community with or without an assistive device. For the FXTAS group, seven PM carriers presented to the movement disorders clinic with signs of cerebellar ataxia on neurological exam and were included in the study. This was to characterize the salient features of ataxic gait in FXTAS. Age-matched control participants were recruited through friends and family members in the above database and through university employee networks. Exclusion criteria were (1) lower extremity orthopedic surgery within the past year, (2) any additional neurological or musculoskeletal disorder that could potentially cause gait and balance problems, or (3) inability to follow directions for the testing protocol. All participants signed an informed consent, approved by the Institutional Review Board at RUMC. Subject demographics are listed in Table 1.

Table 1.

Demographic, clinical, and molecular characteristics of study participants

Variable	Healthy controls (n=18)	PM carriers without FXTAS (n=6)	PM carriers with FXTAS (n=7)
Age	68.78±5.06 (62–80)	65.33±7.45 (60–79)	69.57±4.89 (63–76) ns
Sex	(6 men, 12 women)	(1 man, 5 women)	(4 men, 3 women)
Height (cm)	167.72±8.07	166.5±9.38	170.0±7.90 ns
Weight (kg)	68.38±12.13	65.4±18.3	81.73±23.5 ns
Body mass index	24.65±4.73	23.15±4.25	28.09±7.59 ns
FMR1 CGG repeats		77.5±12.49 (66–100)	88±5.52 (80–94) ns
FXTAS-RS		6.8±4.21 (4–14)	31±14.75 (12–52)**b
			Men (41±10.98)
			Women (21±6.48)c
FXTAS diagnosis	N/A	N/A	4 Definite (3 men, 1 woman), 3 probable (1 man, 2 women)
FXTAS stage of disease	N/A	0	Men (2.75±0.83), Women (3±0.82)c
Activation ratios (AR) of normal allele in women	N/A	0.37±0.15 (0.2–0.5)	0.48±0.60 (0.05–0.9) ns
i-TUG total duration (s)	16.29±2.02	18.21±3.96	23.69±4.93****a, *b
BBS (out of 56)	55.89±0.32 (55–56)	56	52.83±2.64 (49–56)***a, **b
FIM mobility subscale scores (out of possible 35)	35	34.83±0.41 (34–35)	33.43±1.27 (32–35)***a, *b
Neuropathy score		6±5.66 (2–10)	7.6±5.73 (0–14) ns

Age, height, weight, body mass index, CGG repeats, FXTAS-RS scores, FXTAS diagnosis, FXTAS stage of disease, activation ratio (AR), i-TUG total duration, Berg Balance Scale (BBS), Functional Independence Measure (FIM), and neuropathy scores were compared between healthy controls, premutation (PM) carriers without FXTAS, and PM carriers with FXTAS groups. All data (except sex) reported as mean±SD with range in brackets

ns not significantly different between groups

^*p<0.05;

^**p<0.01;

^***p<0.001;

^****p<0.0001

^aSignificantly different from controls

^bSignificantly different from premutation (PM) carriers without FXTAS

^cNot significantly different between men and women with FXTAS

Molecular Analyses

Blood samples were sent to the Rush University Molecular Diagnostic Laboratory (Dr. Berry-Kravis lab) for molecular testing. DNA was isolated from peripheral blood leukocytes of all PM carriers. FMR1 PCR with quantification of allele-specific CGG repeat length and activation ratio (AR) (the percent of cells with the normal FMR1 allele on the active X chromosome) in women PM carriers was performed as previously described [16, 17].

Quantitative Gait Analysis

The i-TUG was performed using the commercially available APDM Mobility Lab™ six inertial sensor system (APDM™; Oregon) [7]. The sensors were attached 4 cm above each malleolus, at the dorsum of the wrists, on the lumbar trunk at the level of L5, and on the upper trunk 2 cm below the sternal notch. The distance of the traditional TUG was increased from 3 to 7 m to allow sufficient gait cycles for analysis. Subjects stood up from a chair, walked 7 m at their self-selected speed, turned around, walked back, and sat down. Subjects performed the i-TUG six times without an assistive device, and the mean value for each gait parameter was calculated. Intraindividual gait variability was determined by the coefficient of variation (standard deviation/mean×100) for each gait and mobility parameter. Based on models of key gait domains in older adults proposed to reflect independent features of the neural control of locomotion [18, 19], we a priori selected variables from four of the five domains of gait. The domains were (1) gait speed (stride velocity and gait cycle time), (2) rhythm (cadence), (3) gait variability (stride length, stride velocity, and cadence variability), and (4) gait cycle phase (percentage (%) of gait cycle spent in swing and double-limb support phases). We also assessed trunk range of motion (ROM) in all three planes (horizontal, sagittal, and frontal), given that trunk ROM was previously found to be increased in patients with cerebellar ataxia [15]. We created a movement transition domain (turn duration, step time before turn, number of steps to turn, and turn-to-sit duration) to assess the impact of FXTAS on these transitions.

Neurological Rating Scales

All PM carriers were scored on the FXTAS Rating Scale (FXTAS-RS) by a movement disorder neurologist, which rates tremor, postural sway, gait, parkinsonism, coordination, dystonia, speech, and oculomotor deficits [20], in order to assess the presence and severity of FXTAS signs. Subjects were also administered the total neuropathy score (TNS), modified to exclude nerve conduction velocity testing [21]. Testing for neuropathy is important given that it is found in a significant number of PM carriers with and without FXTAS [2, 22] and affects performance on spatiotemporal measures of gait [23].

Balance and Functional Disability Scales and Measures

All subjects were administered the following measures: (1) Berg Balance Scale (BBS), a test that measures functional balance activities [24], (2) the mobility subscale of the Functional Independence Measure (FIM), an activities of daily living (ADL) and disability rating scale [25], and (3) the Activities-specific Balance Confidence scale (ABC), a self-reported scale of confidence when performing functional walking tasks [26]. These tests were done in order to determine whether self-report and performance-based balance and disability measures correlated with gait and movement transition measures.

Statistical Analysis

Statistical analysis was performed using GraphPad Prism 6. Normal distribution of i-TUG parameters was analyzed using the Shapiro-Wilk test. Group differences were analyzed using one-way ANOVA with Tukey post hoc comparisons. All mean values of i-TUG variables were normally distributed, except for stride length, stride velocity, and cadence variability for which the Kruskal-Wallis test was applied. To accommodate for multiple comparisons, Bonferroni corrections were applied for the six gait domains and the significance level was thus reduced to p≤0.008. Spearman’s rank correlation coefficient (rho) was used to assess the relationship between the FXTAS-RS, BBS, FIM, and ABC scores and i-TUG parameters (statistical significance was set at p<0.05).

Results

Subject Characteristics

Our study included seven PM carriers with FXTAS and signs of cerebellar gait ataxia on clinical exam, six PM carriers without FXTAS, and 18 age-matched controls. Demographic, clinical, and molecular data are presented in Table 1. There were no significant age differences between groups or differences in CGG repeat size or AR (in women) between PM carriers with and without FXTAS. As expected, PM carriers with FXTAS had worse scores on the FXTAS-RS than those without FXTAS (p=0.0087). However, the TNS did not differ between the two groups. PM carriers with FXTAS scored lower on the BBS than controls and PM carriers without FXTAS (p=0.0004 and 0.0014). All controls and all but one PM carrier without FXTAS had perfect FIM mobility scores. PM carriers with FXTAS had significantly reduced FIM mobility subscale scores compared to controls (p=0.0001) and PM carriers without FXTAS (p=0.027) due to requiring an assistive device when walking in the community, handrail support for stair climbing, and/or assistance with bathtub transfers.

Gait Performance

Subjects took an average of 208 steps over the six i-TUG trials, meeting the recommendation of greater than 30 steps to accurately assess gait variability indices [13].

Between-Group Comparisons

Significant differences were found between controls and PM carriers with FXTAS on all gait and movement transition domains, except for trunk ROM variables and number of steps to turn (Table 2). PM carriers with FXTAS had significantly impaired performance in the domains of gait speed, rhythm, cycle phase, variability, and movement transition (p=0.008 to <0.0001). Similarly, PM carriers with FXTAS performed significantly worse than those without FXTAS on all of these variables except stride velocity and swing and double-limb support time (before Bonferroni corrections). No significant differences between controls and PM carriers without FXTAS were found on any i-TUG variable, but increased double support time observed in asymptomatic carriers was close to reaching significance (p=0.06) (Fig. 1).

Table 2.

Descriptive data and statistics for between-group comparisons for all gait and movement transition domains and their specific variables

i-TUG domain parameters	Control (n=18) Mean	SD	No FXTAS (n=6) Mean	SD	FXTAS (n=7) Mean	SD	Control vs. No FXTAS p value	Control vs. FXTAS p value	No FXTAS vs. FXTAS p value
Gait speed
Stride velocity (m/s)	1.36	0.16	1.31	0.20	1.09	0.18	0.834	0.008	0.090
Gait cycle time (s)	1.04	0.08	1.06	0.09	1.23	0.17	0.929	0.001	0.021*
Gait rhythm
Cadence (steps/min)	116.4	8.70	114.3	9.19	99.27	11.99	0.895	0.001	0.023*
Gait cycle phase
Double-limb support (%)	18.52	3.60	22.76	3.76	24.94	4.40	0.063	0.002	0.569
Swing (%)	40.74	1.81	38.62	1.88	37.54	2.20	0.637	0.002	0.576
Gait trunk ROM (degrees)
Horizontal	4.37	1.59	3.86	1.23	5.42	2.42	0.808	0.387	0.263
Sagittal	4.37	0.82	4.78	1.49	5.02	0.97	0.663	0.325	0.903
Frontal	8.18	2.50	8.48	2.29	9.07	1.20	0.956	0.645	0.882
Gait variability
Stride length (m) CoV	2.25		2.07		3.99		0.932	0.002	0.007
Stride velocity (m/s) CoV	3.08		3.23		6.71		0.982	<0.001	0.003
Cadence (steps/min) CoV	2.36		2.22		4.48		0.969	0.001	0.005
Movement transition
Turn duration (s)	2.03	0.33	2.08	0.52	2.83	0.75	0.977	0.003	0.026*
Step time before turn (s)	0.53	0.04	0.53	0.06	0.61	0.08	0.987	0.003	0.027*
# Steps to turn	4.58	0.59	4.41	0.64	4.76	0.81	0.833	0.814	0.595
Turn-to-sit duration (s)	3.53	0.51	3.98	0.89	5.34	1.02	0.400	<0.0001	0.006

Gait and movement transition domain variables using the instrumented Timed Up and Go (i-TUG) test were compared between age-matched controls, premutation (PM) carriers without FXTAS (No FXTAS), and PM carriers with FXTAS (FXTAS) via one-way ANOVA. CoV coefficient of variation which was determined by the formula CoV=SD/mean×100. Significant p values (p<0.008 following Bonferroni corrections) are shown in bold. Values that did not significantly differ between PM carriers with and without FXTAS (except trunk ROM) are shown in italics

^*Values that were significant before Bonferroni corrections

Fig. 1.

Mean differences in double-limb support time (a) and swing time (b) between controls, premutation carriers without FXTAS (No FXTAS), and premutation carriers with FXTAS (FXTAS). Data reported as mean+SEM. **p<0.01

Correlations

Correlations are shown in Table 3. The ABC and the FIM correlated with multiple gait and movement transition parameters (p=0.044 to<0.0001), while the FXTAS-RS only correlated with stride length and velocity variability (p<0.03). The BBS significantly correlated with swing and double-limb support time (p=0.007), turn-to-sit duration (p=0.012), and variability of stride length and velocity (p<0.02). CGG repeat size did not correlate significantly with any i-TUG parameters but did with both FXTAS-RS and BBS scores (p=0.013 and 0.048). There were no correlations between AR in women and i-TUG parameters. The ABC correlated very highly with almost all i-TUG domains, except for gait variability (Fig. 2).

Table 3.

Spearman’s correlation coefficients (rho) between i-TUG domain parameters, ABC, BBS, FXTAS-RS, and FIM scores

i-TUG domain parameters	ABC	BBS	FXTAS-RS	FIM
Gait speed
Stride velocity (m/s)	0.712 ***	0.218	−0.257	0.316
Gait cycle time (s)	−0.533**	−0.312	0.369	−0.449*
Gait rhythm
Cadence (steps/min)	0.547 **	0.312	−0.369	0.443 *
Gait cycle phase
Double-limb support (%)	−0.580**	−0.480**	0.270	−0.444*
Swing (%)	0.580 **	0.476 **	−0.271	0.441 *
Gait trunk ROM (degrees)
Horizontal	−0.205	−0.092	0.271	−0.193
Sagittal	−0.213	−0.134	−0.299	−0.240
Frontal	−0.280	−0.293	−0.236	−0.161
Gait variability
Stride length (m) CoV	−0.187	−0.429*	0.633 *	−0.308
Stride velocity (m/s) CoV	−0.260	−0.423*	0.703 *	−0.458**
Cadence (steps/min) CoV	−0.453	−0.336	0.531	−0.365*
Movement transition
Turn duration (s)	−0.414*	−0.297	0.429	−0.409*
Step time before turn (s)	−0.510**	−0.307	0.334	−0.393*
# Steps to turn	−0.208	−0.298	0.568	−0.191
Turn-to-sit duration (s)	−0.695***	−0.452*	0.559	−0.595***

Significant correlations following Spearman’s rank correlation analyses are shown in bold

i-TUG instrumented Timed Up and Go, ABC Activities-specific Balance Confidence scale, BBS Berg Balance Scale, FXTAS-RS FXTAS Rating Scale, FIM Functional Independence Mobility scale, CoV coefficient of variation

^*p≤0.05;

^**p<0.01;

^***p<0.001

Fig. 2.

Spearman’s correlations (rho) between double-limb support time (a), stride velocity (b), turn-to-sit duration (c), cadence (d), and Activity-specific Balance Confidence (ABC) scale scores

Discussion

This is the first study to quantitatively examine gait and functional movement transitions in PM carriers with FXTAS who have cerebellar gait ataxia. PM carriers with FXTAS were globally impaired in every gait domain compared to controls. They were significantly slower on multiple measures of gait and turning speed (stride velocity, gait cycle time, cadence, turn duration, step time before turn, and turn-to-sit duration) compared to controls. These findings suggest that PM carriers with FXTAS needed to move at a slower velocity to compensate for lack of stability and avoid falling. Additionally, they spent more time than controls in the double-limb support phase of gait and less time in swing phase, which is consistent with previous studies of patients with cerebellar gait ataxia [14, 27, 28]. This is thought to be a compensatory mechanism for decreased balance stability [29]. Our finding of significant correlations between gait cycle phase variables and the BBS supports this hypothesis.

Our finding that PM carriers with FXTAS demonstrated significantly greater gait variability than both controls and PM carriers without FXTAS is also consistent with prior studies in cerebellar ataxia [14, 27, 30]. Increased gait variability with cerebellar pathology is thought to be due to difficulty in coordinating limb movements requiring step-by-step adjustments of the trunk and limbs during gait and movement transitions [14, 30]. The cerebellar Purkinje cells are important in pacemaking [31], and their dysfunction or loss in FXTAS is likely to cause problems with the timing and coordination of locomotion. Increased gait variability is also associated with increased fall risk [32].

Patients with cerebellar ataxia may show an increase in head and trunk ROM [15]. However, we did not find significant differences between PM carriers with FXTAS and controls in any trunk ROM variable. The i-TUG algorithm for measuring trunk ROM is determined by the pelvic and sternal sensors. However, the excessive trunk motion seen in ataxic patients during gait probably represents the entire body center of mass moving excessively [14] and in an irregular pattern and not the movement of the upper trunk relative to the pelvis. The apparent discrepancy with prior findings will need to be explored in future studies.

PM carriers with FXTAS were significantly impaired in movement transitions including turn duration, step time before turn, and turn-to-sit duration. They did not require a greater number of steps to perform turns, suggesting that they took longer to perform each step. These findings are consistent with a previous study where patients with cerebellar ataxia needed more time to complete 180-degree turns [33]. The same study showed that patients had slower deceleration into a turn, and our increased step time before turn is consistent with these findings. Slower turning is postulated to be due to lack of anticipatory postural control [34] and control over one’s center of mass [33].

PM carriers without FXTAS performed similarly to controls on all gait parameters, indicating that our small cohort had normal gait and movement transitions during the i-TUG. However, there was a trend toward increased time spent in the double support phase of the gait cycle in asymptomatic carriers which did not differ from PM carriers with FXTAS. This finding is typical in individuals with cerebellar ataxia and gait instability [14, 27, 30] and may be predictive for falls in the elderly [35]. PM carriers without FXTAS were not significantly different from those with FXTAS on stride velocity and double support and swing phase times. This finding might suggest that some PM carriers are developing abnormal gait patterns associated with the ataxic FXTAS phenotype. Our recent report of abnormal balance control in PM carriers without FXTAS using computerized posturography [17] and a study reporting increased postural sway (defined as ataxia) in 30 % of asymptomatic PM carriers [36] also support the feasibility of early detection of FXTAS signs with quantitative measurement devices.

The BBS, ABC, and FIM scales correlated with many gait and movement transition parameters, including double-limb support and turn-to-sit duration. ABC scores were highly correlated with all i-TUG domains, except for gait variability and trunk ROM, suggesting that subjects were aware of their balance deficits or lack thereof. Moreover, the ABC has been shown to predict falls in patients with PD and MS [37, 38] and similar future studies in FXTAS would be informative. The fact that the variables captured by the inertial sensors were highly correlated with these widely used functional measures of balance, balance confidence, and disability demonstrates their validity and clinical relevance as potential outcome measures in future studies.

Although the FXTAS-RS correlated with the BBS and FIM, it correlated with very few gait parameters. This is interesting given that it is currently the main clinical assessment tool for disease severity and progression in FXTAS, and the gait disorder in FXTAS is one of the most debilitating aspects of the disease. This may be due to the fact that only 4 of the 44 items on the FXTAS-RS pertain to gait and walking abilities. Our study suggests the need to revise the FXTAS-RS so that gait is weighted more heavily.

It is important to acknowledge that the findings of our study may have been limited by low sample size. However, our gait performance results in PM carriers with FXTAS are consistent with previous studies in patients with other conditions causing cerebellar ataxia. The lack of correlations between FMR1 molecular variables and gait parameters is likely due to small sample size. We recently reported that increased CGG repeat size interacted with increasing age to predict balance deficits in 44 PM carriers with and without FXTAS and that AR was associated with delayed response latencies to balance perturbations [17]. Previous studies have shown that increased CGG repeat size is associated with earlier onset and more severe motor dysfunction in PM carrier men [39] and women (only when AR was incorporated in the analyses) [20], but these also included higher subject numbers. Another limitation of our study was the inability of the inertial sensor system to measure step width, which is increased in some patients with cerebellar ataxia, and may be a compensatory mechanism for lack of postural stability during locomotion [14, 30].

In conclusion, we detected abnormalities in many gait domains and functional movement transitions in PM carriers with FXTAS that distinguish them from both asymptomatic PM carriers and controls. Additionally, we identified a trend toward the abnormal FXTAS gait phenotype in one gait parameter that might prove useful as a surrogate marker of disease if confirmed in future prospective longitudinal studies. Larger studies utilizing this type of gait and motion analysis might also be useful to distinguish an ataxia- versus tremor-dominant form of the disease [40] and provide a sensitive method to monitor disease severity and the natural history of FXTAS.