Bedside electromyography for clinical assessment of sacral motor and reflex activity adapted for patients hospitalized with acute neurological conditions: a pilot study

Abstract

Study design

Pilot cohort study.

Objective

To develop and implement a sacral electromyographic (sEMG) technique at bedside to ascertain sparing of sacral motor activity and reflexes in patients hospitalized for acute neurological conditions.

Setting

Hôpital du Sacré-Coeur de Montréal a Canadian Level-1 university trauma center specialized in SCI care.

Methods

Nine patients underwent digital rectal examination (DRE) and sEMG, assessing voluntary anal contraction and sacral spinal reflexes (bulbocavernosus reflex and the anal wink). Our sEMG technique utilized surface recording electrodes and tactile elicitation of reflexes. EMG signal was acquired at bedside through the Noraxon MR3 system.

Results

It was quick, well accepted and did no harm. We found that contrary to the DRE, sEMG detected subclinical sacral motor activity and reflexes in 20% of cases for voluntary anal contraction and 40% of cases for the anal wink.

Conclusion

We believe our sEMG technique is a powerful tool able to enhance management of patients suffering from acute neurological impairments and requiring sacral function assessment.

Introduction

The digital rectal examination (DRE) is essential to assess neurological impairment as described in the International Standards for Neurological Classification of spinal cord injury (ISNCSCI) [1]. A precise evaluation is important following spinal cord injury (SCI), and cauda equina syndrome (CES) to proceed with early diagnosis, determine injury severity, and facilitate clinical management and rehabilitation planning [2, 3].

The DRE assesses neurologic impairment at the S2–5 levels, in part, by evaluating motor sparing (voluntary anal contraction; VAC) and sacral reflexes (bulbocavernosus reflex; BCR, and anal wink; AW) [1, 4]. It is intimate and invasive for patients, has low sensitivity [5, 6], and yields qualitative results that are highly dependent on the examiner’s experience.

Sacral electromyography (sEMG) is an objective approach to sacral sparing evaluation. This technique measures the electrical activity of the pelvic muscles, (levator ani and bulbocavernosus muscles), responsible for VAC, AW, and BCR [7, 8] in lieu of manually detecting contraction. EMG detection was found superior to manual detection to ascertain the presence of the BCR and AW, in neurologically intact participants and patients suffering from chronic central nervous system lesions [8, 9]. In healthy and chronically injured individuals, EMG techniques could detect, as well as quantify, voluntary contraction of the external anal sphincter [10–14]. Even so, no substantial data pertaining to EMG evaluation of VAC is available in the acute phase of injury, despite VAC being integral in grading the severity of SCI as part of the ISNCSCI [1]. Furthermore, in the context of acute injury, the ability of sEMG and DRE to detect VAC has never been compared.

Sacral reflexes can be elicited through both tactile and electrical stimulation. The latter requires more equipment and can be uncomfortable for patients [11, 15] whereas the former is easier to implement but has been showed to less effectively elicit reflexes [8, 16]. When composing an EMG technique, recording electrodes need to be chosen between surface or more invasive methods. More invasive electrodes have higher precision and sensitivity, but the bigger the burden of their installation will be for the clinician [17].

No EMG technique evaluating sacral motor and reflex functions has been performed or could be easily carried out at bedside due to their complexity and bulkiness. We believe the sEMG has high potential to enhance patient care and management in acute settings. Hence why, this study aimed to develop a bedside sEMG for patients hospitalized with acute neurological conditions. We considered the inevitable trade-off between precision and swiftness of execution of the examination in the context of acute bedside care. As such, we developed a sEMG which utilized tactile ellicitation of sacral reflexes and recording through surface electrodes to detect sparing of motor activity and reflexes in the most caudal parts of the spinal cord.

Methods

We conducted a pilot study at the Hôpital du Sacré-Coeur de Montréal, a Level-1 university trauma center specialized in SCI care. A total of 9 patients were recruited from July to August 2022. Ethics clearance by the institutional review board and written informed consent were duly obtained. Patients were included if they were ≥18 years old and presented with an SCI or CES of any severity and etiology. Patients were excluded if they had associated pelvic injury, were pregnant, or were unable to consent by themselves.

Patients were placed in lateral decubitus and two surface electrodes (Dual-electrodes, gel covered, Ag/AgCl bipolar electrodes, 27 × 40 mm; Noraxon, Scottsdale, Arizona, USA) were positioned above the bulbocavernosus, and the levator ani muscles on the right or left side (see Fig. 1) after the area was shaved and cleaned with an alcohol wipe.

Fig. 1. Schematic representation of electrode placement.

Placement of surface electrodes and position of stimulation for sacral reflex in (A) males and (B) females. AW anal wink, BCR bulbocavernosus reflex.

Digital rectal examination

A DRE was performed by testing VAC, AW, and BCR. Each was repeated twice and scored as absent (−) if the examinator did not detect contraction of the external anal sphincter (EAS), or present (+) if they did. The AW was done by pricking the anal mucosa, on the left and right sides, and visually determining if there was contraction of the EAS [8]. VAC was performed by inserting a finger into the rectum and asking the patient to contract as if to hold back a bowel movement [1]. The BCR was then performed by keeping the finger inserted into the rectum and slightly tugging on the patient’s indwelling catheter or asking the patient to squeeze its glans penis or clitoris [4, 18]. These three methods, share anatomical pathways through the pudendal nerve but could be viewed as separate reflexes (respectively the vesico-anal, penilo-anal, and clitorido-anal reflexes), even so they are all part of the pudendo-anal reflexes and will hereby all be referred to as the BCR. If feces were present in the rectal ampulla, examination was postponed.

Bedside electromyographic assessment

Bedside sEMG assessment of the VAC, AW and BCR was performed with a portable computer. The EMG signal was acquired through the MR3 software (Noraxon, USA, sample rate: 4000 Hz, High pass filter: 5 Hz, Low pass filter: 1500 Hz, Analog output gain: x500 [5 V–10,000 µV]). The sEMG was carried out the same way as the DRE, except no finger was inserted in the rectum. First, EMG activity at rest (baseline) was recorded. To synchronize the EMG activity and the examination, the timestamp (sec.) at which either signal for VAC was given or AW and BCR were performed was noted and used as a synchronization event. To determine if VAC, AW, and BCR were absent (−) or present (+) the EMG signal was then treated and analyzed (Matlab 2019b, signal analyzer application). The raw signal was imported (see Fig. 2A) and then rectified. The envelope of the signal was computed using a Root mean square method (window size: 200; see Fig. 2B). To be categorized as present (+), an increase in muscle activity had to be observed following the synchronization event. If no increase was observed, it was scored as absent (−). To eliminate artefactual EMG signal when eliciting the AW, only signal from the contralateral side to the tactile stimulation was considered.

Fig. 2. Example of electromyographic signal processing.

A Raw EMG data recorded while at rest or performing the voluntary anal contraction, anal wink, or bulbocavernosus reflex. B The EMG data was processed by rectifying the signal and then computing the signal envelope through the Root mean square method (window size: 200) using the Matlab 2019b signal analyzer application. Vertical red line: synchronization event. LA levator ani, BC bulbocavernosus.

For every EMG recording, after each synchronization event, the beginning and end of muscle response was hand-tagged (Matlab 2019b, signal analyzer application), creating a region of interest. For each region, the variation in amplitude was calculated by subtracting the minimum amplitude to the maximum amplitude attained (Max Amp – Min Amp). For each patient, this variation of amplitude was averaged for every repetition of VAC, AW, and BCR to obtain a separate, mean strength of contraction for each component.

The DRE and sEMG were performed once in all patients, except one for whom they were performed twice (14-day delay). A total of 10 DRE and EMG assessments were conducted by one physiatrist specialized in SCI.

Results

Our cohort comprised 9 patients (F:2/M:7), with a median age of 56 years old (20–83 years). Seven had a SCI and two CES (see Table 1). The median number of days between the beginning of neurological impairment and examination was 13 (4–26 days).

Table 1.

Characteristics of the study population.

Age, median (range)	56 (20–83)
Sex
Male	7
Female	2
Nature of injury
Spinal cord injury	7
Cauda equina syndrome	2
AIS grade at admission
A	2
B	2
C	2
D	1
NLI at admission
C1–C4	3
C5–C8	1
T1–T11	3
L1–L5	2
Days between injury and study participation, Median (range)	13 (4–26)

AIS ASIA [American Spinal Injury Association] Impairment Scale, NLI neurological level of injury.

Detection of voluntary anal contraction

For VAC, the sEMG and DRE agreed in 80% of cases (see Table 2). However, in 20% of cases, the sEMG detected a contraction while the DRE did not (DRE−/EMG+; see Fig. 3A). Those two cases were, one T6 AIS (ASIA [American Spinal Injury Association] Impairment Scale) grade D SCI (strength of contraction: 3.31 µV), and one CES (strength of contraction: 6.09 µV), presenting a mean strength of contraction of 4.70 ± 1.39 µV (see Fig. 4). In both cases, EAS tonus and anal sensation were preserved. The SCI showed a 16 points motor score improvement through their stay in acute care, while the incomplete CES’s motor score remained stable at 97/100.

Table 2.

Agreement between the digital rectal examination and sacral EMG examination in the detection of voluntary anal contraction.

		Sacral EMG
		+	−
DRE	+	4	0
	−	2	4

+: VAC present, −: VAC absent.

DRE digital rectal examination, EMG electromyography.

Fig. 3. Individual electromyographic signal of voluntary anal contraction.

Recording of electromyographic activity during voluntary anal contraction in cases presenting voluntary anal contraction during (A) only the sacral EMG (DRE−/EMG+); (B) both the digital rectal examination and the sacral EMG (DRE+/EMG+); or (C) was absent during both examinations (DRE−/EMG−). s: second, Vertical red line: Moment at which patient was asked to perform voluntary anal contraction.

Fig. 4. Strength of contraction recorded during the sacral EMG examination.

Voluntary anal contraction was detected through both DRE and EMG (DRE+/EMG+) in 4 cases (4.02 ± 1.07 µV) and was detected only through the sacral EMG (DRE−/EMG+) in 2 cases (4.70 ± 1.39 µV). The anal wink was DRE+/EMG+ in 1 case (35.30 µV) and DRE−/EMG+ in 4 cases (23.43 ± 15.75 µV). The bulbocavernosus reflex was DRE+/EMG+ in 2 cases (6.97 ± 1.38 µV) and DRE−/EMG+ in 1 case (65.49 µV).

The 4 DRE+/EMG+ cases (see Fig. 3B) were SCIs AIS grade C or D with a neurological level of injury (NLI) between C3 and T11. The mean strength of contraction for those cases was of 4.02 ± 1.07 µV (see Fig. 4). Four cases presented with no VAC both at the DRE and EMG (DRE−/EMG-; see Fig. 3C). Three of them were SCIs AIS grade A or B with a NLI between C4 and T2 and one was a severe CES.

Detection of the anal wink

For AW, both exams agreed in 60% of cases (see Table 3). In the other 40%, the EMG detected contraction following pricking of the anal mucosa while the DRE did not. Those 4 DRE−/EMG+ cases included 2 cervical SCIs (AIS grade A and B), 1 thoracic SCI (AIS grade D), and 1 CES. The mean strength of contraction for all DRE−/EMG+ cases was of 23.43 ± 15.75 µV (see Fig. 4). The case reporting the highest strength of contraction (69.59 µV, see Table 4), was a C4 SCI AIS grade A 24 days after injury and presented with augmented patellar reflex.

Table 3.

Agreement between the digital rectal examination and sacral EMG examination in the detection of anal wink.

		Sacral EMG
		+	−
DRE	+	1	0
	−	4	5

+: AW present, −: AW absent.

DRE digital rectal examination, EMG electromyography.

Table 4.

Individual values of delta amplitude obtained through sacral EMG for each patient regarding VAC, BCR, and AW.

	Strength of contraction (µV)
Patient #	VAC	BCR	AW
1	0	65.49	69.59
2	2.36	0	0
3	7.14	0	0
4	0	0	1.76
5	0	0	0
6	3.31	0	5.05
7	6.09	0	17.33
8	0	0	0
9	3.71	5.58	35.3
10	3.04	8.35	0

VAC voluntary anal contraction, BCR bulbocavernosus reflex, AW anal wink reflex, EMG electromyography.

The only case where AW was DRE+/EMG+ was a C5 SCI AIS grade C. In this instance, the strength of contraction was of 35.30 µV (see Fig. 4). Other concordant cases (DRE−/EMG−) were 4 SCIs AIS grade A-D (C4–T11) and 1 CES.

Detection of the bulbocavernosus reflex

For the BCR, the examinations were concordant in 70% of cases and DRE+/EMG− in 20% (see Table 5). The DRE+/EMG− cases were 2 patients suffering from a cervical SCI AIS grade B and D. However, 10% of cases were DRE−/EMG+. This corresponded to a patient suffering from a C4 AIS grade A SCI presenting with augmented patellar reflex, 24 days after injury. In this instance, the strength of contraction was of 65.49 µV, the highest value obtained for the BCR (see Fig. 4 and Table 4).

Table 5.

Agreement between the digital rectal examination and sacral EMG examination in the detection of the bulbocavernosus reflex.

		Sacral EMG
		+	−
DRE	+	2	2
	−	1	5

+ BCR present, − BCR absent.

DRE digital rectal examination, EMG electromyography.

The 2 DRE+/EMG+ cases were high cervical SCIs AIS grade C and D, with a mean strength of contraction was of 6.97 ± 1.38 µV (see Fig. 4). Five cases, 2 CES and 3 thoracic SCIs (AIS grade A–C–D) were DRE−/EMG−.

Discussion

Our main objective was to develop and implement a sEMG examination performed at bedside during acute care. The sEMG was fast (~10–20 min), well accepted and did not harm patients. We carried out 10 sEMG and DRE in 9 patients suffering from acute neurological impairment. We also compared sEMG to the standard of care for sacral function examination, the DRE. When testing for VAC and AW, sEMG detected contraction while the DRE did not (DRE−/EMG+), respectively, in 20 and 40% of cases.

Detection of voluntary anal contraction

EMG evaluation of EAS function is most often performed using needle electrode recording, which can objectively detect and quantify voluntary and reflex contraction [10–12]. This technique is time consuming and uses invasive electrodes. This raises the question as to what trade-off would be acceptable between precision of the recording and patient experience. In the case of fecal incontinence, surface EMG was just as good as needle EMG to assess pelvic floor activity [13]. Our results also demonstrate that an invasive and tedious technique such as needle EMG is not necessary to objectively quantify activity of the EAS and that surface EMG recording is a valid and feasible method. Our sEMG was easily used at bedside, in patients in the intensive care unit, whereas needle EMG is not.

Not only did we show that sEMG could be used in acute care to quantify VAC, we also showed that it was more precise than the DRE. This is of high importance seeing as presence of VAC is linked to a more favorable prognosis for independent indoor walking post-injury [19]. In two cases, T6 AIS-D SCI and incomplete CES, the EMG was able to capture subclinical contraction of the EAS, which could reflect sacral motor preservation undetected with the DRE by a specialized clinician. Also, this case of AIS-D SCI had a motor score increase linked to significant clinical improvement [20], during their stay in acute care. Although VAC was clinically deemed absent, both patients retained sensation in the region, as such we hypothesize that due to possible discomfort during the DRE patients did not optimally perform VAC, prompting it to be undetected manually. Further studies would be needed to fully understand this phenomenon. VAC was not detected with our EMG technique in all patients with AIS grade A and B SCIs, which was expected as those injuries are deemed motor complete as per the ISNCSCI [1]. This reinforces that activity captured through the sEMG was VAC and not an artifact, even when undetected by the DRE. A subgroup of AIS-A SCI patients presents more favorable recovery trajectories [21], which could maybe be explained by misclassification through the DRE and could potentially be rectified with the sEMG.

Detection of sacral spinal reflexes

The sEMG detected contraction provoked by the AW, while the DRE did not (DRE−/EMG+) in 40% of cases. These results were expected, seeing as the AW is better detected by EMG in healthy individuals and patients with chronic central nervous system injury [8]. Moreover, cases of DRE−/EMG + AW presented with smaller mean strength of contraction than the DRE+/EMG+ cases. This could be expected and explain why contraction of the EAS was not detected during manual examination. Although further studies would be needed to understand the relationship between the reflex being DRE−/EMG+ and the strength of contraction exhibited. This is true when one case of this group is excluded. This case corresponds to the highest strength of contraction, which is exhibited by an AIS grade A SCI, 24 days post-injury (see Table 4). This patient had augmented patellar reflex, a high amplitude for the BCR, and undesirable spasticity [22]. We believe this patient was in the hyper-reflexic stage of spinal shock [23], which could explain the high values for both reflexes.

In determining presence of the BCR, 20% of cases were DRE+/EMG−. This was not expected, since in chronic injury the EMG was more sensitive than the DRE [9]. We postulate there could have been an effect of reflex exhaustion, but further investigation would be necessary. Since the force to elicit the BCR was not standardized, it might be that during the DRE it was enough to provoke a response but not when repeated for the sEMG. Suboptimal electrode positioning could also be partly responsible [24]. The BCR was elicited by either tugging the indwelling catheter [25] or squeezing the glans penis or clitoris [4, 9] and the response was always evaluated through contraction of the bulbocavernosus muscle. With some anatomical differences, these methods evaluate S2–S4 reflex arc integrity with a common efferent response: contraction of the EAS. Recording of the bulbocavernosus muscle instead of direct sphincter measurement could explain DRE and sEMG discrepancies, further studies would be needed. Of note, in a minority of healthy people, BCR and AW are absent [8, 9].

In neurologically intact men, previous data has shown a mean amplitude of 16.53 ± 12.21 µV when the BCR was evoked through electrical stimulation and recorded with surface electrodes over the bulbocavernosus muscle [7]. Our data is in line with this, for all EMG+ cases, except one which showed an augmented strength of contraction, possibly reflecting and hyper-reflexic state.

In acute SCI, cutaneous spinal reflexes have been described as some of the earliest to reappear when the spinal cord emerges from spinal shock [23, 26]. Detailed evaluation of sacral reflexes through sEMG could potentially help monitor a patient’s evolution through spinal shock. Testing spinal reflexes can also help distinguishing between upper and lower motoneuron lesions [27, 28].

Limitations

We chose tactile elicitation of the AW and BCR to develop a technique better suited to the needs of bedside care. The experience and time required to evaluate the BCR by electrical stimulation is a major reason preventing the implementation of previous sEMG techniques in clinical settings [29–31]. Our proposed method circumvents this limitation. However, similarly to the DRE, the manual force used to elicit the BCR is not standardized and may explain some unexpected results. To improve the validity and comparability of the results, we suggest that future studies explore standardization of the manual stimulus, keeping in mind that it should not hinder the ease of the technique and its clinical use. Despite this limitation, this study is the first to propose an accessible method to quantify the magnitude of the BCR’s motor response at bedside.

The positioning of the surface electrodes influences the strength of contraction measured through the EMG signal [24]. As such, it would be important to study the effect of electrode positioning on the EMG response during spinal reflexes and VAC. It would also be necessary to standardize electrode placement, by measuring the distance between the electrodes and a set point, to facilitate inter- and intra-subject comparison. Furthermore, in SCI and CES, deficits can be asymmetric [32]. As such, bilateral sEMG recording should be explored, as opposed to unilateral recording as was done in this study.

This study focused on the assembly of the sEMG technique and the feasibility of its implementation in bedside care. As such, we did not evaluate the inter-rater reliability of the technique since a single experienced clinician performed the sEMG, supporting the validity of this pilot study. To this end, inter-rater reliability of the technique and external validity must be completed in a future study, assessing the identified limitations prior to its widespread use in the clinic.

Following the sEMG, quantification of EMG activity was performed off-line. To facilitate its clinical implementation, it would be important for sEMG to quantify activity directly at bedside through the acquisition software.

Conclusion

The current standard of care to assess possible neurological impairments of the sacral segments is the DRE. Although it has low sensitivity, this examination is integral to the ISNCSCI, where it serves in grading the severity of SCI and sparing at the sacral levels. Multiple studies have shown that detection of sacral reflexes is possible through EMG and is better than the DRE in the context of chronic injury. We demonstrated it is also the case, for voluntary anal contraction and anal wink, in acute care where the diagnosis, estimation of the prognosis, and the rehabilitation plan are made.

Further research is needed to fully evaluate the efficacy and reach of our technique. Even so, we believe this study shows how powerful of a tool our sacral EMG could be to enhance patient management in acute care for individuals requiring sacral neurological examination.

Author contributions

MD contributed to patient recruitment, data collection, extraction, and analysis, interpreting results, redaction of the first draft of the manuscript, and editing. JMMT contributed in designing the protocol, result interpretation, and manuscript editing. ARD participated in designing the protocol, data collection, result interpretation, and manuscript editing.

Funding

This work was supported by the Chaire de recherche Medtronic en traumatologie spinale and funds from the Réseau provincial de recherche en adaptation-réadaptation des Fonds de recherche du Québec—Santé. MD was supported by a scholarship from the Premier fund from the University of Montreal.

Data availability

Upon reasonable request, data generated throughout this study as well as the simple MATLAB code used to analyze the EMG signal and draw the figures from this data are available from the lead contact.

Ethics approval and consent to participate

Ethics clearance by the institutional review board of the Hôpital du Sacré-Coeur de Montréal and written informed consent from each patient were duly obtained for this study.

Competing interests

The authors declare no competing interests.