EMG of the pelvic floor muscles (PFM) has been reported as a promising tool to aid in patient selection, lead placement, lead programming and troubleshooting [1]. Nevertheless, PFM EMG is not embedded in the current treatment regimen of sacral neuromodulation patients. A possible explanation for the latter is that none of the EMG tools used in the aforementioned studies was validated.

This study assesses the validity of a probe with 24 electrodes to register PFM EMG as a tool to measure the motor response upon lead stimulation during lead placement and programming. 
Requirements for validity: (#1) the EMG signals measured are repeatable and (#2) the EMG signals are strongly correlated with visual observation of the motor response and at least as sensitive. 
Requirement for use during programming: (#3) the EMG signals occur before or on the sensory threshold (as patients are stimulated subsensory and stimulation above the current perceived threshold is often painful).

This is a single tertiary center, prospective study conducted from 17/10/2017 to 20/02/2019 including patients with overactive bladder and non-obstructive urinary retention, implanted using the standardized tined lead placement technique [2]. 

EMG of the PFM was recorded using the multiple array probe (MAPLe ®), placed intravaginally [3]. 
Lead stimulation was performed using Medtronic’s stimulation equipment (square wave pulses, pulse width: 210 µsec, frequency: 14 Hz). The most important characteristics of each EMG signal (amplitude and latency) were determined. 

The assessment of the repeatability of the EMG signals (requirement #1) as well as the correlation and association between motor and sensory response threshold (requirement #3) were done with the patient awake. Five lead electrode configurations with bipolar stimulation (3+1-/3+2-/3+1-/0+3-/1+3-) were tested upon lead stimulation up to the sensory threshold (ST) and when the first EMG motor response was noted. After this procedure was performed once, the neuromodulator was turned off for 10 minutes. Thereafter , the same 5 lead electrode configurations were stimulated at the exact same stimulation amplitude. 

The assessment of the association between EMG and visual motor response threshold (requirement #2) occurred during tined lead placement. All 4 lead electrodes were stimulated with increasing stimulation amplitude (1-2-3-5-7-10 mA). EMG motor response was recorded by EMG and visual motor response was recorded using a camera. Afterwards the thresholds for both were determined.

Requirement #1: (figure 1)
For assessment of the repeatability of the EMG signals, 13 consecutive female patients were recruited and included. In total 61 lead electrode configurations were included and stimulated twice (with at least 10 minutes between two measurement). The repeatability between both assessments of the EMG responses at the sensory threshold was excellent for the amplitude (ICC: 0.994, 95% CI 0.992-0.995, p<0.001) and latency (ICC: 0.927, 95% CI 0.911-0.941, p<0.001). This is shown in figure 1.

Requirement #2:
For the assessment of the association between EMG and visual motor response 19 patients and 76 lead electrodes were included. Weighted kappa showed a perfect association between EMG response and visual motor response (κ=0.86, 95% CI 0.86; 0.86). McNemar-Bowker test showed none of both is more sensitive in relation to the other (p=0.92).

Requirement #3: (figure 2)
For assessment of the relation between the EMG motor response and sensory response threshold, 33 female patients and 154 electrode configurations were included. Pearson correlation coefficient showed a strong positive relationship between the motor and sensory threshold (r² = 0.826, p<0.001). A significant difference between the mean sensory response threshold (1.77 +/- 1.11 mA) and mean EMG motor response threshold (1.52 +/- 1.02 mA) was noted, with the mean EMG motor response threshold being on average 0.25 +/- 0.051 mA (95% CI: 0.15; 0.35) lower (paired t-test: p<0.001). Linear regression, with sensory response as dependent variable and EMG motor response as independent variable, showed that a one-unit increase of motor leads to an average increase of 0.90 mA, with a standard error of 0.049 (p<0.0001). The 95% confidence interval around this slope ranges from 0.80 to 0.99. Since this confidence interval for the slope does not include the value of one, one could argue that the increase in EMG motor response threshold is associated with a marginally smaller increase in sensory response threshold. This is in line with the observation that the motor threshold is usually reached before the sensory response threshold. Of the 154 lead electrode configurations eliciting an EMG motor response, the EMG motor response was present before the sensory response threshold in 113 (73.4%) patients, on the sensory response threshold in 16 (10.4%) patients and above the sensory response threshold in 25 (16.2%) patients. Finally, in each of the 33 patients, stimulation of at least one lead electrode configuration showed an EMG motor response threshold which occurred before the sensory response threshold. This is shown in figure 2.

Firstly, as EMG motor response is almost perfectly correlated with visual observation of the motor response by the naked eye, PFM EMG can be used as an objective tool to measure and quantify contractions of the PFM. Therefore, PFM EMG can differentiate the strength of the PFM contraction elicited by stimulation of different locations within one subjects thereby allowing an objective comparison.

Secondly, although placement of the lead in sacral neuromodulation patients is mostly done based upon assessment of the motor response, as a positive motor response is shown to be a superior predictor for a successful test procedure in regards to a positive sensory response, the use of the motor response is abandoned during lead programming and troubleshooting, as the current stimulations needed to elicit a visual motor response are often too high leading to an uncomfortable sensation in an awake patient. The use of PFM EMG solves this problem as our results show an EMG motor response can be measured before or on the sensory threshold in all patients for at least one lead electrode configuration and for 83.8 % of all lead electrode configurations stimulated.

Lastly, PFM EMG allows to examine the effect of the motor response elicited by stimulation of the sacral spinal nerves on the clinical efficacy through objective quantifiable data.

Motor response upon lead stimulation can be reliably measured by PFM EMG in sacral neuromodulation patients under anaesthesia and awake. Therefore, EMG can be used as a valid tool to assist in lead placement and programming.

Noblett KL. Neuromodulation and the role of electrodiagnostic techniques. Int Urogynecol J. 2010;21 Suppl 2:S461-466.
Matzel KE, Chartier-Kastler E, Knowles CH, et al. Sacral Neuromodulation: Standardized Electrode Placement Technique. Neuromodulation. 2017.
Voorham-van der Zalm PJ, Voorham JC, van den Bos TW, et al. Reliability and differentiation of pelvic floor muscle electromyography measurements in healthy volunteers using a new device: the Multiple Array Probe Leiden (MAPLe). Neurourology and urodynamics. 2013;32(4):341-348.