The Association Between Physical Activity/Heart Rate Variability Data Obtained Using a Wearable Device and Timed Motor Functional Tests in Patients with Duchenne Muscular Dystrophy: A Pilot Study

INTRODUCTION

Duchenne muscular dystrophy (DMD) is an X-linked muscle disease characterized by delayed motor development in early childhood that results in difficulty standing and walking due to weakness of proximal muscles. Symptoms are progressive, with wheelchair use by the age of 12 years [1]. Cardiomyopathy occurs in every patient with DMD > 18 years old. Eventually, these patients die of heart and respiratory failure. In recent years, heart failure and arrhythmias due to myocardial damage have become prognostic determinants due to the widespread use and advances in respiratory management [2]. For this reason, periodic echocardiography is an essential part of the standard of care for DMD [3]. In addition to myocardial damage, sinus tachycardia is often observed in DMD, and has been reported to be due to decreased cardiac parasympathetic activity and/or increased sympathetic activity by heart rate variability (HRV) analysis [4–6]. Associations between physical exercise and HRV have been reported in healthy children [7] and patients with pathogenic conditions [8, 9], but not in patients with DMD.

Currently, various timed function tests including time standing from supine position, 10-meter walk test (10MWT), North Star Assessment of Ambulation (NSAA), and 6-minute walk test (6MWT) [10–15] are used in the clinical evaluation for DMD. However, the collection of accurate data often fails not only because of the physical burden on the patient, but also because of confounding factors such as unfamiliar environment, mood, and physical condition that prevents accurate performance. Recent developments in wearable devices has made it possible to collect a variety of biometric information, including physical activity, for clinical practice in patients with DMD [16–20]. Wearable accelerometers have been shown to be useful for quantitative and multifaceted measurement of movement. In particular, by combining wearable sensor technology with machine learning methods, a behavioral biomarker has been constructed that can distinguish DMD from controls and predict disease course both longitudinally and cross-sectionally [21]. This biomarker, derived from a digital representation of daily activities, can predict disease progression and track response to treatment in DMD patients.

A multifunctional wireless Holter recorder (FLRA Holter ECG^®; MED LINK Co., Ltd., Toyoda City, Aichi, Japan) has been developed to reduce the physical and mental burden of the assessment procedure on the wearer and has a built-in 3-axis accelerometer that can measure heart rate, body temperature, step count, physical activity, posture, exercise intensity, and autonomic nervous system information over 24 hours [7]. The purpose of this pilot study is to investigate differences in physical activity and HRV between patients with DMD and healthy subjects using this wearable device, and to explore whether there are physical activity/heart rate variability parameters that correlate with timed function tests performed on patients with DMD.

METHODS

RESULTS

DISCUSSION

Significant differences were found in several parameters of physical activity between the patients with DMD and controls (Table 3), which is expected since DMD is a motor disorder. On the other hand, the patients with DMD consistently leaned backward, while controls leaned forward in their walking posture. Regarding age, height, weight, and BMI, the basal metabolic rate of both groups showed a strong positive correlation with body weight and BMI (Table 4). Basal metabolic rate is influenced by body surface area, age, gender, body size, and body temperature. Physical activity and fat burning have been reported to be significantly higher with increasing age [7]. However, no correlation was found for heart rate with any of the morphometric measurements because of the small age and body size variation in both groups.

When HRV was compared between the two groups, there was no significant difference in mean heart rate, and maximal heart rate was lower in the DMD group (Table 5). While most previous reports have included patients of a wide age range or investigated patients older than 10 years [6], it is likely that the healthy boys included in this study were younger and had higher heart rates, which would explain why there was no difference between the two groups. It has been reported that heart rate does not decrease with age in patients with DMD, unlike normal subjects [22], and the difference between the two groups may become clearer when older age groups are included in the analysis.

Among the HRV parameters, LF norm was significantly lower in patients with DMD. LF is considered to be an index influenced by both sympathetic and parasympathetic nerve activity. On the other hand, the HF component, a measure of parasympathetic activity, did not differ between the two groups, suggesting that patients with DMD may have abnormal sympathetic activity. Loss of vagal tone and increased sympathetic tone [23], as well as abnormal autonomic function as measured by HRV have been reported in patients with DMD [23–26]. On the other hand, all frequency domain components of HRV in ambulatory DMD have been reported to be within the normal range with no difference compared to age-matched healthy controls [27]. These HRV results are controversial, do not examine all parameters, and may be influenced by factors such as the age, duration of illness, and degree of physical condition. We noted a trend toward lower VLF during arousal in patients with DMD (Table 5). Only one report has shown a significant decrease in VLF components in DMD [28]. VLF decreases with sympathetic activity and increases with parasympathetic activity [29, 30]. Recently, the clinical significance of VLF has attracted much attention because of a report that it is an independent risk factor for chronic heart failure [31]. The significance of reduced LF and VLF in patients with DMD needs further investigation, including reproducibility.

A relationship between daily physical activity and basic motor performance has been reported [32]. Therefore, we investigated the relationship between physical activity parameters and various motor functions in DMD-affected children (Supplementary Table 1). Contrary to expectations, our results showed that none of the physical activity parameters showed significant correlations with motor functions. However, an inverse correlation (r = –0.729) has been observed between the time from supine to standing and the degree of forward leaning during walking, although the correlation was just insignificant (p = 0.06). Since a forward shift of the body’s center of gravity is important for standing up, posterior pelvic tilt has been reported to cause difficulty in standing stand up [33, 34]. Thus, the backward tilt in patients with DMD may be related to the time taken to stand up from a supine position. 10MWT, NSAA, and 6MWT showed no significant correlation with any of the physical activity parameters. The reason for this is unknown, but we believe that it cannot be concluded without further collection of data over time and more subjects. In dystrophic dogs, acceleration and angular velocity have been reported to correlate with spontaneous locomotion [35, 36], and we believe that it may be necessary to consider body acceleration during movement.

Finally, we examined the relationship between HRV and various timed motor functional tests (Table 6) because an association between HRV and exercise capacity has been noted in healthy subjects [7], people with obesity [8], and heart failure patients [9]. Time from supine to standing did not correlate with any of the HRV parameters. This test examines momentary motor function and may not be affected by heart rate or the autonomic nervous system. The 10MWT, on the other hand, correlated inversely with mean heart rate in all conditions. The NSAA values were inversely correlated with the mean heart rate during arousal, as well as the 10MWT, and were also positively correlated with TF and VLF. TF indicates total autonomic activity, while VLF indicates both sympathetic and parasympathetic activity, as mentioned earlier, and reduced activity of these autonomic nerves may affect these motor function tests. In fact, a positive correlation between VLF and physical activity has been reported in stroke patients [37]. Thus, the VLF component of HRV may be associated with physical activity in certain pathologies. The 6MWT was not associated with heart rate, but was inversely correlated with LF/HF ratio in all conditions. TF, VLF, LF norm, and HF norm also showed high, albeit slightly non-significant, correlation coefficients (Table 6). The 6MWT was inversely correlated with the LF/HF ratio (the higher the sympathetic activity, the shorter the 6MWT). It has been reported that aerobic capacity and LH/HF are inversely correlated [38], and autonomic balance (LH/HF) in DMD may be related to exercise tolerance. Although the small number of cases in this study makes it impossible to draw definitive conclusions, autonomic abnormalities in DMD may play a role in motor function and exercise tolerance. We plan to clarify this in more detail in our next study.

The weaknesses and limitations of this study are as a pilot study of a small number of patients and normal controls and only generated hypotheses that should be confirmed in larger studies. An additional weakness of the study was the failure to check for reproducibility, perform multivariate analysis, and monitor the natural history in each patient.

CONCLUSION

In this pilot study, the wearable device demonstrated differences in various physical activity/HRV parameters in daily life of patients with DMD and control subjects and the possibility that HRV parameters are related to timed motor function tests. However, these results are hypothetical and longitudinal studies in large cohorts are needed to draw conclusions.