In sacral neuromodulation (SNM), sacral spinal nerves are electrically stimulated. This improves bladder and bowel function in patients with overactive bladder, non-obstructive urinary retention or faecal incontinence. The mechanism of action and predictive factors of SNM are still not completely understood.
This study evaluates pelvic floor muscle activity as a factor that may reflect accurate lead placement. It also evaluates if SNM changes pelvic floor muscle activity and if these changes are predictive for SNM success.

A tined lead electrode was placed according to the standardized procedure for lead placement, aiming to obtain visually inward movement of the pelvic floor upon stimulation below 2V at 3 or 4 electrodes (1).
EMG of the pelvic floor muscles (PFM) was recorded using the MAPLe, placed intravaginally (2). The recording procedure consisted of consecutive stimulations (using a monophasic square wave of 210 µs at 14Hz) of the 4 electrodes of the tined lead with increasing amplitudes (1-2-3-5-7-10V). These recordings were made after electrode placement (m0) and after 3 weeks of continuous subsensory threshold electrical stimulation (m1). During the 3 weeks of SNM therapy, stimulation could be adapted after 1 week if no clinical improvement was felt by the patient. At the end of the 3 weeks, efficacy of SNM was evaluated by comparing changes in a 3-day voiding diary. More than 50% improvement was defined as a successful outcome of SNM stimulation (responders), otherwise the patient was considered a non-responder to treatment.
For statistical analysis the highest PFM EMG response was identified for each stimulation train (at each lead electrode and each stimulation amplitude) for each MAPLe electrode. EMG responses were averaged over all leads and voltages per MAPLe electrode.
Unpaired t-tests were used to detect differences of average EMG responses between responders and non-responders at the first measurement or at follow-up after three weeks stimulation. Paired t-tests were used to detect changes in EMG response after follow-up compared to initial measurement for non-responders and responders.

Fourteen patients were eligible for inclusion (9 responders and 5 non-responders). In total 8064 registrations (14 patients stimulated at 4 lead electrodes with 6 different voltages resulting in 336 stimulation trains on 24 MAPLe electrodes) were analyzed. 
Table 1 shows EMG responses averaged for all lead electrodes, stimulation amplitudes and all MAPLe electrodes. Global averaged EMG response was not significantly different for responders comparing follow-up to initial measurement, for non-responder it was significantly lower. Both the initial and follow-up measurement were significantly higher for responders compared to non-responders.
Figure 1 shows the differentiated average EMG responses for the responders and non-responders during the initial and 3-week follow-up measurements (figures a and b for responders and figures d and e for non-responders respectively).  In addition, the difference in EMG activity between the initial measurement and the one after 3 weeks as well as the difference between responders and non-responders is illustrated. A distinction is made between the different sides: for anterior, ipsilateral (side where lead is placed), posterior and contralateral sides and the different depths of the PFM. The overall maximum EMG response is found at the follow-up measurement for the Responder-group, Ipsilateral at the second deepest MAPLe electrode represented as black or darkest colour in Figure 1 b). 
For responders, after 3 weeks of stimulation, the average EMG response was significantly higher on in the deep and mid parts at the ipsilateral side of the PFM and at the superficial layer at the contralateral side of the PFM. Average EMG response was significantly lower in all parts of the PFM (except one mid part) at the anterior side of the PFM, at the superficial part on the ipsilateral side of the PFM and at the deepest part at posterior and contralateral sides of the PFM (figure c).
For non-responders, the average EMG response was not significantly higher on any depth or side of the PFM. Average EMG response was significantly lower on all depths of the PFM anterior and all but the deepest part at the posterior side of the PFM, at the superficial  part at the contralateral side of the PFM and at deep and middle parts at the ipsilateral side of the PFM (figure f).
Comparing the initial measurement (m0) for responders versus non-responders the average EMG response was significantly higher at the deepest part of the ipsilateral, anterior and posterior side of the PFM, at the middle parts of the ipsilateral and contralateral side of the PFM and the superficial parts at contralateral and anterior side of the PFM. Average EMG response was significantly lower at the superficial part of the PFM at the posterior side and the middle part at the anterior side of the PFM (figure h).

Baseline pelvic floor muscle activity upon stimulation of the different lead electrodes appears to be different in SNM responders compared to non-responders. Responders have a significantly higher EMG response in the deepest at the anterior, ipsilateral and posterior sides of the PFM, middle part at the ipsilateral side of the PFM and at the middle and superficial parts at the contralateral side of the PFM than non-responders. Responders seem to have significantly higher PFM EMG responses at almost all parts and sides compared to non-responders at the end of the test stimulation phase. Whether these differences are due to differences in lead placement or patient anatomy cannot be deducted from this study.
Subsensory threshold electrical stimulation of the sacral spinal nerves, such as occurs in SNM treatment, changes pelvic floor muscle activity. Responders seem to have significantly higher PFM EMG responses nearest to the simulation site (ipsilateral deep and mid parts of PFM) after 3 weeks of stimulation. Whether this neuromodulation of pelvic floor activity is related to clinical outcome of SNM treatment needs to be studied in a larger group.
The current recording method with the MAPLe enables to evaluate different parts of the PFM separately during the same stimulation sessions. In responder’s pelvic floor muscle activity increases primarily at the deep and middle parts at the ipsilateral side of the PFM (the side where the tined lead is placed). No increase in activity was noted in the non-responders.

Subsensory electrical stimulation of the sacral spinal nerves such as during SNM changes pelvic floor activity. Changes are different between responders and non-responders of treatment. Good or high EMG response during the initial measurement in the deep parts of the PFM and at the mid and superficial part at the contralateral side of the PFM may be indicators if the patient is going to respond to sacral neuromodulation. This statement needs to be explored in a larger group of patients.

Matzel K.E., Chartier-Kastler E., Knowles C.H., Lehur P.A., Munoz-Duyos A., Ratto C., Rydningen M.B., Sørensen M., ~ van Kerrebroeck P., de Wachter S. 2017. Sacral Neuromodulation: Standardized Electrode Placement Technique. Neuromodulation 2017; 20: 816–824
Voorham-van der Zalm PJ, Voorham JC, van den Bos TW, Ouwerkerk TJ, Putter H, Wasser MN. Reliability and differentiation of pelvic floor muscle electromyography measurements in healthy volunteers using a new device: the Multiple Array Probe Leiden (MAPLe). Neurourol Urodyn. 2013;32:341–348