Alterations of Autonomic Nervous System Activity in Children with Spina Bida: A Case-Control Study

We compared heart rate variability parameters of patients with spina bida with those of a control group during urodynamic study to evaluate the autonomic nervous system dysfunction of spina bida. Continuous heart rate variability parameters were recorded during 3 successive periods (P0: 2 minutes before the start of lling; P1: start of lling to the rst desire to void; P2: P1 to the end of lling or the start of voiding). Children with vesicoureteral reux who underwent video-urodynamic study were established as a control group. We included 11 patients with spina bida and 9 controls. At baseline, patients with spina bida had lower values of RMSSD, pNN50, and HF, while LF/HF ratio was increased (5.04 ± 4.75 vs 0.67 ± 0.42, p = 0.014). During bladder lling, LF/HF values increased in the control group (P0 0.67 ± 0.42 vs P1 0.89 ± 0.34 vs P2 1.21 ± 0.64, p = 0.018) while it was decreased in spina bida patients (P0 5.04 ± 4.75 vs P1 3.96 ± 4.35 vs P2 3.26 ± 4.03, p < 0.001). HF were signicantly increased in spina bida children during bladder lling (p = 0.002). In time domain, SDNN was increased only in control group during bladder lling. Parasympathetic activity domains were decreased in children with spina bida at baseline. During the bladder lling phase, parasympathetic activity increased with xed sympathetic activity in spina bida group while the control group demonstrated a shifted balance toward sympathetic preponderance at the end of bladder lling. These ndings may be related to the pathophysiology of neurogenic bladder in spina bida.


Introduction
Spina bi da lies within the spectrum of neural tube defects caused by a failure of the caudal neural tube to fuse normally in early development, affecting the central nervous system and resulting in permanent disability. It is the most common birth defect with an average incidence of 1 to 10 in 1000 live births. 1 Magnetic resonance imaging can provide an early diagnosis and untethering may reduce potential lifelong disabilities. Spina bi da can result in severe disability of multiple organ systems, including neurological de cits, sphincter dysfunction, and limb deformities. 2 Variable impact on the somatic, parasympathetic, and sympathetic innervation of the bladder affects the patient's ability to store and empty urine and can ultimately cause chronic kidney disease due to poor bladder dynamics. As the lower urinary tract (LUT) is closely coupled to the autonomic nervous system (ANS), 3 an indicator of ANS activity may offer objective information about bladder sensations and further understanding about underlying pathophysiology for speci c diseases. Baselines and ANS activity changes during bladder lling has been studied in healthy controls or patients with lower urinary tract symptoms (LUTs) without neurological de cit. 4,5 However, there are no studies regarding how the ANS affects bladder pathophysiology in patients with spina bi da.
One of the most practical, reproducible, and non-invasive ways to monitor the ANS is heart rate variability (HRV). HRV measures the spontaneous change of the R-R interval in continuous electrocardiogram (ECG) as a response to physiologic demand which re ects the continuous interplay between sympathetic and parasympathetic in uences. 6 Although HRV was initially used to evaluate autonomic function in cardiovascular diseases, it has also been used to assess underlying dysfunction of ANS related to LUT disease in several studies. 5,7,8 Therefore, our objectives were (1) to compare ANS activity using HRV between children with spina bi da and a control group at baseline and (2) to investigate and compare responses to bladder lling during urodynamic study in both groups.

Results
There were no signi cant differences between the 2 groups with respect to baseline characteristics ( Table 1). The control group consisted of 7 patients previously diagnosed with VUR and 2 patients who underwent examination to evaluate if VUR was present. Seven patients (63.6%) were in use of clean intermittent catheterization in the study group. Low bladder compliance less than 20 ml/cmHO was observed in 7 patients (63.6%) of the study group while none were present in the control group.

Discussion
To our knowledge, this is the rst study to objectively compare the baseline pro le of ANS activity and responses to bladder lling in children with or without spina bi da using spectral analysis of cardiovascular signal variability methodology. Several ndings in our study are noteworthy. First, the baseline pro le of HRV was altered in spina bi da patients in terms of decreased values related with the peripheral nervous system (RMSSD, pNN50, HF) and increased LF/HF ratio representing sympathetic dominance. The response to bladder lling also signi cantly differed between the two groups, most pronounced in the LF/HF ratio (increasing value in the control group vs decreasing value in the spina bi da group with bladder lling). These results provide an explanation of the underlying neuropathology in spina bi da.
There are a variety of urological consequences depending on the severity of the fusion abnormality and location of the lesion. An upper motor neuron lesion with detrusor overactivity and detrusor sphincter dyssynergy are most likely to develop, but an acontractile detrusor and sphincter denervation are also seen as a result of spinal cord tethering. 9,10 The mainstay in treatment of spina bi da is still combination of pharmacological agents and clean intermittent catheterization, which has had limited success. 11 Current treatment strategies have focused on preventing the consequences of neurogenic bladder rather than trying to understand the disease, therefore the understanding of the disease has not yet improved.
Since the rst report of changes in HRV by Hon and Lee, 12 HRV has been used extensively as a quantitative marker of ANS activity. Previous studies demonstrated that ANS dysfunction proven by HRV re ects the underlying pathology of LUTs. 5,8 The most frequently studied LUTs associated with HRV is overactive bladder (OAB), which is controversial as the results, as well as study settings, of how ANS differs from healthy controls were heterogenous. Hubeaux et al. reported a predominance of parasympathetic activity with the bladder emptied and a preponderance of sympathetic activity at the end of bladder lling in women with OAB syndrome. 13 Choi and Kim observed a decrease in HRV indexes, including HF, SDNN, RMSSD, in women presenting with OAB compared to healthy women. 14 In children with bladder bowel dysfunction, HF was signi cantly lower than healthy controls at baseline. 15 In healthy subjects, LF/HF ratio showed a stable sympathovagal balance until rst desire to void, while the balance demonstrated a shift towards sympathetic activation before strong desire to void. 4 This response is thought to be caused by a similar mechanism to the vesicovascular response that is mediated by sympathetic nerves, such as the hypogastric nerve. 16 Unlike previous studies, LF/HF ratio increased until rst desire to void and continued to rise until the end of lling in our study. This result may be due to the younger age group in our study not being capable of expressing exact bladder sensation and also the wide variation of HRV value. During bladder lling, contrary to the control group, only HF increased without increasing LF in the spina bi da group, resulting in a decrease in LF/HF ratio. These results suggest ANS dysfunction, predominantly a sympathetic ANS dysfunction, in spina bi da. Given that the thoracolumbar sympathetic efferent pathways in the hypogastric and pelvic nerves induce an inhibition of detrusor muscle and an excitation of the bladder base and urethra, the alterations observed in sympathetic activity during bladder lling may lead to a sensation of urgency and a problem in urine storage. In our study, neurogenic detrusor overactivity (NDO) was seen in only 4 patients (36.3%); however, on past urodynamic studies, NDO was observed in all except 1 patient. These urodynamic results may be associated with the decreased LF/HF ratio during bladder lling.
In addition, SDNN in time domain was signi cantly increased with bladder lling in the control group, whereas in remained unchanged in spina bi da patients. Given that SDNN represents overall HRV, 17 the unchanged value of SDNN in spina bi da may be due to decreased neurotransmission of ANS during bladder lling. With extrapolation from the enteric nervous system, which has similar nerve distribution as the bladder, neural loss and decreased nerve ber density in the myenteric plexus was seen in spina bi da patients, which correlated with severity of bowel dysfunction. 18 These alterations may be caused by reaction to disrupted extrinsic innervation, resulting in trans-neuronal degeneration. Therefore, we infer that these evidences may be related to an overall decreased HRV value during bladder lling in spina bi da, but they cannot explain the sympathetic predominance at baseline.
Recently, autonomic cardiovascular function was evaluated in wheelchair-using children with MMC, and RMSSD at rest were reduced compared to the control group, which agrees with our results of reduced RMSSD, pNN50, and HF in the spinal bi da group at baseline. 19 The pathology underlying reduced vagal tone and increased LF/HF ratio in spina bi da patients is uncertain. In patients with thoracic spinal cord injuries, higher heart rates and reduced vagal tone has been documented. 20 This may re ect compensation for decreased stroke volume and compensatory reductions in vagal tone to maintain autonomic balance. 21,22 Deteriorating vascular properties (small diameter, low ow, and high shear stress) were present in patients with spina bi da 23 , and these results suggest that decreased vagal tone and sympathetic predominance in spina bi da may occur via similar mechanism demonstrated in spinal cord injury.
The main limitation of this study was the small number of subjects, therefore, HRV analysis was not performed for each subtype according to urodynamic study. Considering the large deviation of baseline HRV, a larger study is required to con rm our results. The second limitation is HRV measurements in our study were performed with a non-medical Bluetooth device, not with conventional ECG. However, recent studies have demonstrated that Bluetooth devices have provided an acceptable agreement for the measurement of HRV when compared with ECG. Therefore, the measurement method in this study has validity. 24,25 Although other criteria could be considered to see the change in HRV in terms of time, it is di cult to divide periods based on a speci c time because each patient has different lling time inherent to their bladder capacity. Patients with spina bi da often lose their sense of bladder and demonstrate underactivity of bladder, which leads confusion in deciding proper point of each period. Despite these limitations, the criteria we applied to divide each period could be suboptimal method to observe the trends of HRV changes on bladder lling in both of control and spina bi da. Another concern is that HRV may be affected by emotional stress caused by the arti cial setting of the study, however, the same setting was applied in both groups and the examination was conducted with su cient time to adapt.
Another limitation may be that the control group consisted of children with VUR who may have abnormal bladder dysfunction that may be related to HRV, although this has not been studied.
At baseline, HRV parameters representing PNS activity are decreased in children with spina bi da compared to control group. During bladder lling, parasympathetic activity was relatively increased with a xed sympathetic activity in spina bi da while control group demonstrated a shift in sympathetic/parasympathetic balance towards the sympathetic component. These results may be related to underlying neuropathology caused by spina bi da, which could be used to better understand and manage spina bi da.

Patient selection
The present study protocol was approved by the Yonsei University Health System Institutional Review Board (4-2015-0332). The study was performed in accordance with all applicable laws and regulations, good clinical practices, and the ethical principles described in the Declaration of Helsinki. A retrospective pilot study was carried out between July 2015 and March 2016 and this study con rmed the method of HRV measurement and identi ed the difference between control group and spina bi da patients. Subsequently, the prospective study was conducted from April 2016 to February 2017 to complement retrospective data after IRB approval of study protocol. Participants were informed about the purpose and procedure of the study prior to participation and that they could withdraw from the study at any time without explanation. In both of a pilot and prospective study, informed consent was obtained from parents and also children through age-appropriate agreement document in the presence of legal representatives or parents. The inclusion and exclusion criteria of prospective protocol were retrospectively applied to the pilot data. Regardless of the timing of the study, HRV was measured with a xed protocol during video urodynamic study.
All patients were evaluated via a combination of clinical history, physical examination, and questionnaire for LUTs. Inclusion criteria for the study group consisted of patients older than 4 years old who underwent detethering surgery due to spina bi da, either due to a lipomeningomyelocele (LMMC) or meningomyelocele (MMC). The control group contained patients who underwent a video urodynamic study to evaluate vesicoureteral re ux (VUR) or other anomalies in upper urinary tract. Exclusion criteria were as follows: medication of anticholinergics and a-blocker within one month; diagnosis of neurological disorders affecting the autonomic nervous system except spina bi da; diagnosis of hypertension or arrhythmia; diabetes; history of previous urological surgery including augmentation cystoplasty; anatomical abnormalities in bladder and urethra; having had coffee, tea, cigarettes and other foods which could affect autonomic nervous system before examination. In control group, patients who had LUTs such as daytime incontinence or urgency were also excluded.
In sixteen of pilot cases, four patients were retrospectively excluded following criteria due to LUTs requiring medication. Five cases with spina bi da were also excluded due to history of augmentation cystoplasty and taking anticholinergics. The remaining seven patients were recruited retrospectively. From April 2016, twenty patients were prospectively screened for the study and 13 patients were enrolled. A total of 20 children (9 in the control group and 11 in the study group) were included for analysis.

Urodynamic study and HRV measurement
Urodynamic study was performed using a recommended technique by the International Children's Continence Society. 26 We used 6-Fr double-lumen catheters for the urethra and 12-Fr uid lled balloon catheters for the rectum. A saline solution was warmed to body temperature for infusion at lling rates of 5-10% of a known or predicted capacity. The expected bladder capacity was estimated by using the following formula (in mL): [30 + (age in years×30)]. In children with spina bi da, bladder lling was terminated under the following conditions: child has a strong urge to void, continuous leakage is observed, detrusor pressure reaches greater than 40 cm H 2 O, or the patient felt a sensation of bladder fullness or abdominal discomfort. 27 The patients were in the supine position and were not under anesthesia during the examination. Surface electrodes were used and EMG patch were positioned symmetrically, perineally, left and right of the anus.
In the control group, bladder lling was continued until children felt a strong desire to void and could no longer delay micturition. Maximal cystometric capacity is the bladder volume at the end of the lling phase or when "permission to void" is given. 26 To prevent the anxiety caused by examination from affecting baseline HRV levels, patients were given su cient stability after setting up the examination. After setting and calibration of urodynamic 3 channels, continuous HRV data was obtained during three periods of the bladder lling (Fig. 2).
Period 0 (P0): two minutes before the start of bladder lling  17 Comparable to most studies, time domain measures were 1) the standard deviation of all NN intervals (SDNN), 2) the root mean squared of successive differences of successive R-R normal intervals (RMSSD), 3) the successive percentage of R-R interval differences greater than 50 ms (pNN50). RMSSD and pNN50 are correlated with high frequency (HF), which is a response to changes in parasympathetic activity. Frequency domain measured power spectra with calculation of the integral of 1) very low frequency (VLF; ≤0.04 Hz) 2) low frequency (LF; 0.04-0.15 Hz), which is the marker of primarily sympathetic cardiac modulation with some parasympathetic in uence, 3) HF (0.15-0.4 Hz), which is a marker of parasympathetic modulation, and 4) LF/HF ratio, which is an indicator of the autonomic balance between the sympathetic and parasympathetic nervous systems. 17 Statistical analysis Data are expressed as mean ± standard deviation. The Mann-Whitney U test was used in comparison of mean values as a nonparametric approach and the t-test was used for parametric variables. Coe cient of variation (CoV) (CoV = SD/mean) were calculated to measure the dispersion of a probability distribution. Changes of HRV during bladder lling were compared in each group using repeated measures analysis of variance (ANOVA). The compound symmetry assumption was studied with the Mauchly sphericity test. After observing the signi cance of the Mauchly sphericity test and seeing the lack of spherical or posite symmetry in repeated measurements of HRV parameters, the Pillai trace multivariate test was adopted.  Figure 1 Changes of heart rate variability during bladder lling phase in patients with spina bi da and control group * p<0.05 compared only with P0 by paired t-test, ** p<0.05 compared with P0 and P1 by paired ttest