High-Density Surface Electromyographic Assessment of Pelvic Floor Hypertonicity in IC/BPS Patients: A Pilot Study

Abstract

Objective:

To assess the feasibility of objectively assessing pelvic floor hypertonicity (PFH) in women with interstitial cystitis/bladder pain syndrome (IC/BPS) using an intra-vaginal high-density surface electromyography (HD-sEMG) probe.

Methods:

Seven female subjects (mean age 44±13 yr.) with a prior diagnosis of IC/BPS were recruited. A full digital pelvic exam was administered to identify hypertonic muscles. Intra-vaginal HD-sEMG was acquired during rest. Root-mean-squared (RMS) amplitude during rest was calculated for each channel to define a hypertonicity index and hypertonic zone. Innervation zones (IZs) were identified from the bipolar mapping of decomposed HD-sEMG signals, and summarized into an IZ distribution mapping.

Results:

Of the 7 subjects recruited, 5 had normal pelvic floor muscle tone and 2 exhibited hypertonicity upon muscle palpation. Subjects with PFH demonstrated a higher hypertonicity index (12.6±3.5 vs. 4.5±1.2) in sessions 1 and 2. The hypertonic zone defined by the 64-channel RMS mapping coincided with the digital pelvic exam findings. The corresponding IZs were localized for each motor unit. The hypertonicity indices between two consecutive sessions were well correlated (CC=.95).

Conclusions:

This study represents the first effort to employ intra-vaginal HD-sEMG to assess PFH in women with IC/BPS. Our results demonstrate the feasibility of HD-sEMG to provide a quantitative diagnosis of PFH and the precise localization of hypertonic muscles and IZs. The proposed HD-sEMG based techniques provide promising tools for clinical diagnosis and treatment of PFH, such as the personalized guidance of BoNT injections.

Brief Summary:

In this pilot study, a novel intravaginal high-density EMG technique was developed to objectively assess resting EMG activity in women with IC/BPS. Potential applications are discussed.

INTRODUCTION

Pelvic floor hypertonicity (PFH) is characterized by an increase in resting pelvic floor muscle tone resulting from either increased contractile activity (neurogenic hypertonicity) and/or passive stiffness (non-neurogenic hypertonicity) [1]. PFH presents in up to 87% of women with interstitial cystitis or bladder pain syndrome (IC/BPS) [2]. The etiology of PFH in IC/BPS patients is postulated to be that silent afferents within the bladder become activated and this results in a constant barrage of noxious stimuli that can then result in wind-up of the dorsal horn. This leads to the development of an abnormal viceromuscular reflex that triggers muscular hyperalgesia and allodynia. These neuromuscular changes result in hypertonicity [3]. PFH is often identified through digital pelvic palpation, which largely depends on the examiner’s subjective perception of an abnormality, and therefore results can vary between examiners. Currently, there lacks a proper objective technique that is clinically accepted for diagnosing PFH.

Furthermore, approximately 30% of patients fail conservative therapy [4] and require more aggressive treatment. Botulinum neurotoxin (BoNT) has recently received growing interest in the management of pelvic pain secondary to PFH,[5] with a well-evidenced therapeutic benefit [6]. BoNT relaxes the pelvic floor muscles (PFM) and nociceptive signaling by blocking the release of acetylcholine at neuromuscular junctions. The injection location modulates the efficacy of BoNT. The therapeutic effect of BoNT is suppressed by approximately half when injected 1 cm away from the innervation zone (IZ) [7]. The IZ represents the location where the alpha-motor neuron forms synapses with target muscle fibers and neuromuscular junctions are distributed densely [8]. To reach an ideal therapeutic effect, some practitioners may inject a higher dosage of BoNT; however, BoNT may induce dose-dependent side effects such as urinary and fecal incontinence.

The current clinical standard for BoNT injection involves the manual digital palpation of a hypertonic muscle, followed by an injection towards the palpating finger [9]. This injection protocol, like the assessment of PFH itself, is subjective and inaccurate, and may not provide a consistent therapeutic effect due to non-targeted injections.

There is no existing technique that can guide the injection of BoNT to the IZ in the PFMs. Endovaginal ultrasound provides anatomical information, but does not specify myoelectric information. A 4D ultrasound technique was recently developed to provide visual feedback to the operator, greatly improving the precision and repeatability of the injection [10]. Ultrasound, however, is unable to localize the IZ or determine which muscle is overactive, permitting potential improvements to the injection process. Intramuscular electromyography (iEMG) has been employed to localize hypertonicity for BoNT injection;[11 ] however, iEMG is limited by its invasive nature, poor spatial resolution, and is incapable of characterizing the IZ distribution. A competent and objective tool is necessary for the reliable diagnosis and treatment of PFH.

High-density surface electromyography (HD-sEMG) allows for the non-invasive quantification of global muscle activity, as opposed to local activity assessed by iEMG. The motor unit action potential (MUAP) initiates at the IZ and propagates along muscle fibers innervated by that motor unit in opposing directions. HD-sEMG has proven to be successful in capturing this signal propagation, and IZs can be determined by observing the phase reversal using differential signal analysis or amplification [12].

In this study, we present the first effort to assess neurogenic PFH via an intra-vaginal HD-sEMG probe in women with IC/BPS, and localize the IZ nearest to the hypertonic region to allow for the targeted injection of BoNT using HD-sEMG signal decomposition. We hope to provide a reliable tool for the characterization of pelvic floor dysfunction in IC/BPS patients, and possibly improve the efficacy of treatment by directing the injection of BoNT to the IZs of the hypertonic region.

MATERIALS AND METHODS

Participants

Female subjects (n=7, mean age 44±13 yr., range (27-68)) with a prior diagnosis of IC/BPS were recruited via a voluntary sample from the Urology Clinic at Baylor College of Medicine in Houston, Texas. Inclusion criteria were females between the ages of 18 and 70, a diagnosis of IC/BPS, and an unpleasant sensation in the PFM in the absence of infection or an immediately identifiable cause. Subjects were excluded from participation if they were pregnant, breast feeding or were later found not to meet inclusion criteria. All subjects were informed of any associated risks and gave written informed consent. The institutional review boards of the Baylor College of Medicine and the University of Houston approved the experimental protocol.

Study Protocol

Urinary and pain symptoms associated with IC/BPS was assessed with the interstitial cystitis symptom and problem index. Pain symptoms were assessed with the McGill pain questionnaire, pain quality assessment scale, and Numeric Pain Rating Scale. Participants were asked mark their baseline pain on a 0-10 numeric pain rating scale (0 is no pain, 10 is worst possible pain). Subjects were asked to consider their pain in the preceding week. Subjects were then situated in the dorsal lithotomy position, and digital pelvic exams were administered by the study urologist (C.P.S). Pain upon palpating the left and right sides of the obturator internus, pubococcygeus, and puborectalis was graded on a visual analog scale. Muscle tone was assessed binarily and noted as normal (−) or hypertonic (+).

A 64-channel (8x8) high-density surface electromyography (HD-sEMG) probe (length 175 mm, diameter 22.7 mm, inter-electrode distance 8.5 mm) as shown in Figure 2(A), was lubricated with conductive gel and introduced into the vaginal space. A fully soaked Velcro strap was affixed to the wrist as the ground, and a single reference electrode was placed to the subjecťs thigh. Subjects were instructed to relax, and resting EMG activity was simultaneously recorded for 60 seconds. All HD-sEMG recordings were sampled at 2048 Hz with a Refa 136-channel amplifier (TMSi, Enschede, Netherlands). Signal quality was verified in real-time via a display monitor. To evaluate the repeatability of HD-sEMG recording, a 5-10-minute rest period was given, and then same recording protocol was repeated by the same examiner to give session 2. Consistent probe depth and orientation was achieved by orienting the company trademark in the anterior direction. Care was taken by the physician to keep the probe in a fixed location, with no longitudinal movement or rotation of the probe during testing to maintain a fixed spatial reference by holding the probe during the recording session.

Figure 2.

A) Recording electrode with 64-channel mapping overlaid and intra-vaginal HDsEMG probe; B) 64-channel mappings for each subject in the non-hypertonicity group (Subjects 1-3, 5-6) generated from session 1; C) 64-channel mappings from each subject in the hypertonicity group (subjects 4 and 7) from session 1. Red indicates increased muscle RMS amplitude at that location. *Mappings were interpolated by a factor of 4 for display purposes

Signal processing

Resting HD-sEMG signals were bandpass filtered between 10 and 500 Hz with a second-order Butterworth bandpass filter. Time periods with movement artifacts or poor electrode contact were segmented out from further analysis. Power line contamination was removed with a 60 Hz second-order Butterworth notch filter. Root mean squared amplitude (RMS) was calculated for each channel in 0.5 second intervals. The resulting resting RMS values for each 0.5 second interval were then averaged to give a 64-channel RMS mapping. The average resting RMS of the 16 channels with the highest RMS amplitude were averaged, and termed the “hypertonicity index” (HI), and were compared between groups, and sessions, as shown in Figure 3(C). All offline processing was performed post-exam in MATLAB R2018 (Mathworks, Natick, MA).

Figure 3.

A) Bar chart displaying hypertonicity index for both sessions; B) Comparison of hypertonicity index between sessions; C) Bar-plot for both groups and sessions. A marked difference between groups was found in sessions 1 and 2. An apparent reduction in hypertonicity index was found between sessions 1 and 2 for the hypertonicity group, but not the non-hypertonic group.

Monopolar HD-sEMG recordings during rest were decomposed into constitutive MUAP spike trains using a blind source separation approach, as described in previous publications [13]. The 64-channel MUAP profiles were then constructed using by averaging the HD-sEMG signals at the time of each decomposed MUAP’s firing. The bipolar mappings of each 64-channel MUAP profile was obtained by subtracting the signal of each channel from that of its neighboring channel in a clockwise direction around the probe. IZs were visually identified by observing the phase reversal from the bipolar mappings [8,14,15], as shown in Figure 1. The clock position in axial view, depth from the introitus, and MUAP propagation length were noted and visualized for each motor unit, as shown in Figures 4 and 5. MUAP propagation length was determined by the most distant channels with visible MUAP waveforms (> 10 μV) from the defined IZ.

Figure 1.

Top) Filtered HD-sEMG signals from one subject, Bottom) 64-channel motor unit action potential mapping overlaid on hypertonicity mapping result for one subject. Red box marks the IZ location.

Figure 4.

shows the IZ maps for the two hypertonic subjects overlaid on their respective 64-channel mappings during session 1. *Note 64-channel mappings were interpolated by a factor of 4 for display purposes. The red dot represents an innervation zone near the hypertonic region, the blue dot represents an innervation zone away from the hypertonic region. Black lines represent the muscle fiber propagation of the recorded motor unit.

Figure 5.

A) Axial view of IZ mappings in subjects 2 and 6. B) Axial view of IZ mappings in subjects 4 and 7. The red dot represents an innervation zone near the hypertonic region, the blue dot represents an innervation zone away from the hypertonic region, as defined by the hypertonicity mapping shown in Figure 4. The numbers next to the dots represent the depth from the introitus.

Statistical analysis

Subjects with confirmed PFH were grouped to the hypertonicity group and the remaining subjects were grouped to the non-hypertonicity group. The repeatability of the HD-sEMG assessment was evaluated by the correlation coefficient (CC) between the hypertonicity indexes in two consecutive sessions.

RESULTS

The average interstitial cystitis symptom and problem scores were 10±1.8 (range: 8-14) and 9.4±2.2 (7-13), respectively. The average self-reported baseline numerical pelvic pain score of pain felt in the preceding week for all subjects was 5.6±1.7 (3-8). All seven subjects described their pain as “intense”, “cramping”, and reported a sensation of “heat” in the PFM muscles, as described in the pain quality assessment scale. 3 of the 7 subjects reported numbness or the sensation of electrical shocks in the PFM. PFM pain and muscle hypertonicity upon palpitation was assessed all subjects, and summarized in Table I. An averaged cumulative pain score, defined as the sum of pain felt in each muscle when palpated, of 16.3±6.9 (7-27) was found. PFH was observed in two of the seven subjects. Average cumulative pain upon palpation for the women without PFH was 14.6±7.4 and 20.5±2.5 for women with PFH. The average baseline pain scores were 4.8±1.3 and 7.5±0.5, for women without and with PFH, respectively.

Table I:

Digital VAS pain scores and hypertonicity assessment (+/−) for all subjects upon palpitation. In which: NPS-Baseline Numerical pain score without palpitation; ROI - Right Obturator Internus; RPC- Right Pubococceygeus; RPR –Right Puborectalis; LOI - Left Obturator Internus; LPC – Left Pubococceygeus; LPR – Left Puborectalis; CP: cumulative pain; HI: hypertonicity index

Sub	NPS	ROI	LOI	RPC	LPC	RPR	LPR	CP	HI
1	7	3 (−)	3 (−)	2 (−)	2 (−)	0 (−)	0 (−)	10	3.0±1.0
2	5	5 (−)	9 (−)	6 (−)	7 (−)	0 (−)	0 (−)	27	5.8±0.8
3	4	0 (−)	0 (−)	2 (−)	3 (−)	1 (−)	4 (−)	10	5.6±0.3
4	8	3 (−)	2 (−)	6 (+)	5 (+)	4 (−)	3 (+)	23	15.1±3.4
5	3	0 (−)	1 (−)	2 (−)	2 (−)	1 (−)	1 (−)	7	3.6±1.8
6	5	0 (−)	4 (−)	4 (−)	5 (−)	3 (−)	3 (−)	19	4.7±1.2
7	7	3 (−)	3 (−)	4 (−)	2 (−)	3 (+)	3 (−)	18	10.2±3.2

HD-sEMG signals were successfully acquired from all subjects, and 64-channel high-density RMS mappings were calculated for the subjects without hypertonicity Figure 2(B) and with hypertonicity Figure 2(C). Average hypertonicity indexes of 4.5±1.2 (3.0-5.7) and 12.6±3.5 (10.2-15.1) were obtained for subjects without hypertonicity and with hypertonicity respectively, as shown in Table 1 and Figure 3(A). Subjects with PFH demonstrated a higher hypertonicity index in both sessions 1 and 2. The proposed hypertonicity index was repeatable with a correlation coefficient between sessions 1 and 2 of 0.95, as shown in Figure 3(B). The hypertonicity index represents the myoelectric output of the muscles nearest the 16 channels with the highest amplitude. In both subjects with hypertonicity, the RMS mapping “hotspot” appeared on the ipsilateral side with respect to the hypertonic muscle found on the digital muscle tone exam. These two subjects exhibited an average hypertonicity index more than 5 standard deviations higher than the mean non-hypertonicity subjects, as shown in Figures 2(C) and 3(C). Hypertonicity index was not associated with increased pelvic floor muscle pain upon palpation in each muscle. Subjects with PFH, however, reported higher average baseline pain (4.8±1.3 for the non-hypertonicity group, 7.5±0.5 for the hypertonicity group), and slightly higher cumulative pain upon palpation (14.6±7.4 for the non-hypertonicity group and 20.5±2.5 for the hypertonicity group).

IZ locations were successfully localized from the decomposed HD-sEMG signals, as exemplified in Figures 1, 4, and 5. Hypertonic zones were most commonly found near the right puborectalis (RPC) and right puborectalis (RPR), agreeing with our digital palpation findings. An axial depiction of decomposed motor units innervating the PFM was derived, as shown in Figures 5(A) and 5(B).

The linear relationship between either baseline pain, or cumulative pain, and average hypertonicity index was assessed as shown in Figure 6, and strong correlation was found for baseline pain and hypertonicity index (R=0.67). Cumulative pain upon palpation and hypertonicity index were found to be moderately correlated (R=0.44).

Figure 6.

Left) Relationship between average hypertonicity index and baseline pain. Right) Relationship between average hypertonicity index and cumulative pain upon palpation.

DISCUSSION

This study presents the first effort to assess PFH and extract innervation information in female IC/BPS subjects using intra-vaginal HD-sEMG. The presented intra-vaginal probe can capture abundant myoelectric information from muscles located in the proximity of the vaginal wall, including puborectalis, pubococcygeus, and bulbospongiosus. In subjects with confirmed hypertonicity, the RMS mapping “hotspot” appeared on the ipsilateral side with respect to the digital muscle tone exam. A marked difference in the RMS mapping and defined hypertonicity index acquired at rest was found between IC/BPS subjects with and without hypertonicity, as shown in Figure 2(B) and Figure 2(C). An increase in hypertonicity index was found in the hypertonic group, when compared to the non-hypertonic group, as shown in Figure 3(C), indicating that the proposed HD-sEMG based technique can detect hypertonic muscles from resting EMG signals. A decrease in hypertonic index was found between sessions 1 and 2 for the hypertonic group, but not the non-hypertonic group. Our findings in the non-hypertonic group were similar to those reported in healthy subjects by Grape et al, with a mean hypertonicity index of 4.5±1.2 being comparable to their resting EMG amplitude findings of 5.0±1.1 in 17 healthy women [16].

Furthermore, the hypertonic zones visualized from the mapping match well with the hypertonic muscles defined by the digital pelvic exams. Our results suggest that HD-sEMG can be a promising tool to enhance the clinical diagnosis of PFH, including the localization of hypertonic muscles via RMS mapping and the objective estimation of severity.

We did not observe a relationship between hypertonicity index and individual muscle pain scores upon palpation. This observation suggests that severe pelvic pain can exist in the absence of a myofascial component. This may be explained by the etiology of PFH, where noxious stimuli (pain) results in upregulation of the dorsal horn, causing muscular instability and hyper contraction. In this case, the pain may exist without the associated muscle over activity. It should be mentioned that the patients with hypertonicity rated their baseline pelvic pain to be much higher than the patients with normal muscle tone, as shown in Figure 6. Furthermore, previous studies have estimated the prevalence of PFH in ICBPS patients to be as high as 87% [2]. The low rate (28.5%) of PFH we observed is likely explained by the low sample size of our study, and care should be taken not to extrapolate this prevalence to the general population.

IZs are commonly identified using a evenly-spaced linear electrode array that is placed along the muscle fiber direction [14]. Beretta Picolli et al. employed visual signal inspection for signal phase inversion to localize IZs in 43 superficial muscles, and reported excellent or good results in 76% of the muscles using this method [14]. Enck et al. were the first to assess external anal sphincter IZ distributions by employing visual identification of signal phase inversion in 52 healthy adults [17]. If IZs are unable to be discerned directly from the surface interference pattern, a valid alternative is a decomposition based IZ detection scheme. Our previous work has demonstrated the feasibility of characterizing IZ distribution in the pelvic muscles of healthy participants using pelvic HD-sEMG. Peng et. al. employed HD-sEMG decomposition to generate an IZ distribution of the PFMs of healthy young women [12]. Dias et. al. applied a similar technique to study the IZ distribution of the EAS in young and elderly healthy women [18]. HD-sEMG decomposition was performed in the present study to suppress signal crosstalk and generate a clean mapping for the estimation, and we successfully generated personalized IZ mappings for each subject. As shown in Figure 4, we were able to localize an IZ near the hypertonic zone for both subjects with hypertonicity from the decomposed surface interference pattern EMG data, suggesting the feasibility of an HD-sEMG guided injection protocol directed to the IZ. A marked inter-subject variation of the IZ maps was observed, as shown in Figures 4 and 5, which coincides with previous observations of varied IZ distribution in limb muscles [19]. This finding stresses the importance of a personalized injection strategy to maximize BoNT efficiency.

Future studies may explore whether the derived axial IZ mappings shown in figure can be used to optimize the treatment efficacy of BoNT for alleviating PFH. The efficacy of BoNT therapy can likely be potentiated by endplate targeted injections achieved by specifying the muscle(s) responsible for PFH, as well as the offending IZ. Lapatki et al. reported a 46% reduction in efficacy when BoNT was injected 1 cm away from the IZ, stressing the importance of an accurate injection guidance [7], however it has yet to be determined if these results can be generalized to the PFMs. Several studies have explored guided BoNT injections by employing ultrasound techniques to localize spastic muscles in both limb and face muscles [20,21]. Ultrasound, however, is unable to quantify neuromuscular activity and is best used as a source of real-time feedback during BoNT injection, rather than a tool to localize the source of muscle spasticity. Morrissey et al. utilized intramuscular EMG (iEMG) to specify the muscular origins of PFH in women, and reported significant improvements in pain scores upon BoNT injection [11]. However, iEMG does not provide an estimation of the global muscle activity of the PFM; rather, it is limited to the small uptake area of the intramuscular sensor. This limitation necessitates multiple needle insertions to assess the activity of the entire pelvic floor.

It is important to contrast myofascial trigger points and IZs. Myofascial trigger points are identified by a palpable band or nodule in the muscle that refer pain and elicit a twitch response when touched. A recent trial of BoNT injections into palpable myofascial trigger points in women with myofascial pain did not result in a significant improvement in myofascial pain when compared to placebo [22]. However, the accuracy of manual needle placement has been proved surprisingly low, and in most cases more than half (57%) injections fell outside the target muscle [23]. In addition, as BoNT acts at the neuromuscular junction, which is electromyographically indicated by the IZ, injection towards the IZs in hypertonic patients may produce better therapeutic outcomes than trigger point injections. In fact, the location of myofascial trigger points and IZ’s have been shown to not overlap. Barbero et al. used a IZ detection technique to compare the locations of IZs to myofascial trigger point locations in the upper trapezius muscle, and found that trigger points were proximally located to the IZ, but did not overlap (spaced 10.4mm apart on average) [8]. This finding suggests that BoNT injections targeted near trigger points may not reach the IZ.

In addition to personalizing injection therapies, the proposed hypertonicity mapping technique may help guide myofascial and/or Theile’s massage towards hypertonic muscles. Weiss et. al. used conventional surface EMG to show that women with myofascial pain demonstrated high resting EMG pre-treatment via manual therapy when compared to post-treatment sessions [4]. We believe that augmenting the myofascial massage procedure presented in Weiss et. al. to include patient-specific HD-sEMG guidance towards hypertonic muscles, rather than simply as a measure of global muscle activity may improve treatment outcomes.

Although HD-sEMG can reliably capture activities from multiple muscles, it is unlikely that vaginal EMG is successful in detecting signals from the obturator internus, or piriformis. Source imaging studies need to be completed to verify this assumption. Another limitation of this study is the lack of an objective measure of PFM tone, such as vaginal manometry. Therefore we can only use the subjective findings of the digital exam to validate the proposed technique. However, our results suggest that the hypertonicity index can stand as an objective measure of PFH severity and may provide new perspectives in the understanding and diagnosis of PFH. Follow-up studies with a larger sample of hypertonic subjects are necessary to confirm these findings. The location of the probe electrodes with respect to the surrounding anatomy needs to be confirmed for our probe in a future study. Magnetic resonance images have been published with a similarly dimensioned intravaginal probe in place, and it was able to clearly capture signals from both sides of the puborectalis, pubococcygeus, and iliococcygeus muscles [24]. Finally, a concurrent EMG-anatomical imaging study should take place to assess potential confounding factors introduced by anticipatory contraction of the PFMs prior to pelvic exam or intravaginal HD-sEMG.

CONCLUSION

HD-sEMG provides novel perspectives for assessing the neuromuscular health of the PFMs. We have found that this intra-vaginal HD-sEMG probe was successful in assessing PFH in women with IC/BPS. The innervation information of the PFM, whether hypertonic or not, can be extracted to provide a personalized IZ mapping that may greatly benefit the clinical management of PFH.

Abbreviations

PFH

Pelvic floor hypertonicity

IC/BPS

Interstitial cystitis or bladder pain syndrome

EMG

Electromyography

IZ

Innervation zone

MUAP

motor unit action potential

HD-sEMG

High-density surface electromyography

BoNT

Botulinum neurotoxin

PFM

Pelvic floor muscles

RPC

Right pubococcygeus

RPR

Right puborectalis