Acute dorsal genital nerve stimulation increases subjective arousal in women with and without spinal cord injury

Introduction

Female sexual function has been historically understudied, leading to limited treatment options for the approximately 40-50% of women who suffer from symptoms associated with female sexual dysfunction (FSD)¹. Existing treatment options for FSD, such as bremelanotide² and flibanserin³, primarily target hypoactive sexual desire disorder (HSDD). There is a lack of treatments that target challenges with the physiological aspects of sexual function, such as lubrication or arousal. Sildenafil, a successful pharmaceutical for treating male sexual arousal dysfunction, was pursued for FSD but ultimately abandoned due to its low efficacy rate and high incidence of adverse events⁴.

FSD can have a variety of etiologies, one of which is spinal cord injury (SCI). People with SCI report sexual function as one of their top priorities to regain^5,6 and sexual function is an important factor in quality of life for all adults⁷. Location and severity of the injury often determine which aspects of sexual function are impacted (e.g., arousal, desire). Psychogenic, but not reflexogenic, arousal is often possible in women with sacral level injuries while reflexogenic arousal is generally retained in women with injuries above the lumbar level⁸. These two examples demonstrate the heterogeneity in FSD symptoms among women with SCI, an underserved population that would particularly benefit from FSD treatments developed with pathophysiology taken into consideration.

Neuromodulation, or electrical stimulation of neural targets, has shown some promise in treating FSD in non-neurogenic women. Clinical trials using sacral neuromodulation to treat women with bladder dysfunction found that their sexual function, as evaluated by the female sexual function index (FSFI)⁹, improved as an unanticipated benefit^10–12. Other bladder dysfunction neuromodulation targets have been investigated as treatments for FSD, including tibial nerve stimulation (TNS)^13,14 and dorsal genital nerve stimulation (DGNS)¹⁴. Although the mechanisms of these interventions are not fully understood, we theorize that DGNS and TNS can improve FSD by increasing genital arousal. Genital arousal, often measured by vaginal blood flow, has shown to have short-term increases during peripheral nerve stimulation in preclinical models. Animal studies using TNS have shown increases in vaginal blood flow^15,16 and we hypothesize that the underlying mechanisms involve a spinal reflex pathway. Similarly, animal studies using pudendal nerve stimulation, the proximal source of the dorsal genital nerve, have shown increases in vaginal blood flow^17,18. These studies suggest that pudendal nerve stimulation activates spinal pathways that in turn activate the pelvic efferents that modulate vaginal blood flow. We hypothesize that increased blood flow contributes, at least in part, to improved FSFI scores for women with FSD.

Clinical studies on female sexual function often measure genital arousal, however this is the first study to measure genital arousal in response to transcutaneous neuromodulation. We sought to investigate if acute neuromodulation can modulate vaginal blood flow in women with SCI, able-bodied women with non-neurogenic FSD, and able-bodied women without FSD as healthy controls. Our goal was to assess if acute transcutaneous neuromodulation of the dorsal genital or tibial nerve can modulate genital and subjective arousal. We chose these three groups of participants to assess which treatments were best at evoking a blood flow response and subjective arousal given the presence of SCI or FSD.

Materials and Methods

All study activities were approved by the University of Michigan Institutional Review Board (HUM00148746) prior to initiation and all data was collected at Michigan Medicine between November 2020 and March 2022. We recruited participants via physician referral, flyers placed in relevant clinics in the local area, and online through a University of Michigan health research portal. The study is registered at clinicaltrials.gov under identifier NCT04384172.

This study consists of a randomized crossover design with three groups: women with SCI (SCI), women with non-neurogenic FSD (FSD), and women with No Dysfunction and who are Able-Bodied (NDAB). Participants were screened for eligibility with a clinical study coordinator prior to enrollment. All participants were over 18 years old, biologically female, and sexually active at least once a month. SCI participants could be interested in sexual activity if not sexually active. To be included in the SCI arm, participants had to have a clinically diagnosed spinal cord injury at grade AIS (American Spinal Injury Association Impairment Scale) A-C at a level within C6-S1 at least six months prior to enrollment and a short-form FSFI¹⁹ score below 19. Women with FSD were neurologically intact with a short-form FSFI score below 19 and an FSFI lubrication sub-score below or equal to 3. Women without FSD were neurologically intact with a short-form FSFI score above or equal to 19 and FSFI lubrication sub-score above 4. Exclusion criteria for all participants were as follows: (1) pregnant, (2) clinically diagnosed bladder dysfunction, pelvic pain, or other pelvic organ symptoms, (3) active infection or active pressure sores in the perineal region, (4) epilepsy, and (5) implanted pacemaker or defibrillator. Additional exclusion criteria for SCI participants included worsening in motor or sensory function in the last month. NDAB and FSD participants were also excluded if they had clinically diagnosed bladder dysfunction, pelvic pain, or other pelvic organ symptom.

After obtaining informed consent, we instructed participants to submit demographic information and complete five clinically validated surveys: the American Urological Association Symptom Index (AUASI) bladder symptom index²⁰, the female sexual function index (FSFI)⁹, the fecal incontinence severity index (FISI)²¹, the patient assessment of constipation-symptoms (PAC-SYM)²², and the short-form qualify of life survey (SF-36)²³. The surveys were collected online in REDCap, a standard clinical tool for survey data collection²⁴. Participants completed one or two study sessions corresponding to two stimulation targets: the dorsal genital nerve and tibial nerve. We used block randomization, with block sizes of 10 for each group, to determine which nerve target was used in the first study session. Study team members and participants were not blinded. Participant’s second study session was one to five months after their first. Participants filled out pelvic function surveys that asked them about their bladder, bowel, and sexual function for a given day. The surveys were filled out daily from two days prior to two days after each study session to monitor any carryover effects from neuromodulation.

At each study session, participants were asked to sit, partially reclined in a comfortable position. A vaginal photoplethysmography transducer (TSD204A, Biopac Systems Inc., Goleta, CA) was placed in the vaginal canal to monitor vaginal pulse amplitude (VPA), a measurement of relative vaginal blood flow²⁵. A clinician placed two round surface electrodes (1.25 inch diameter, ValuTrode Neurostimulation Electrodes CF3200, Axelgaard Manufacturing Co. Ltd., Fallbrook, CA) on either side of the clitoris²⁶ for DGNS study sessions and above the malleolus and on the bottom of the foot²⁷ for TNS study sessions. Stimulation was delivered with a transcutaneous electrical nerve stimulation (TENS) device (Empi Select 199584, Medi-Stim Inc., Wabasha, MN, USA). The amplitude was determined by slowly increasing it from 0 mA until a maximum comfortable level or 60 mA was reached, whichever was lower. We recorded VPA at a sampling rate of 200 Hz during a 5-minute baseline period, 20 minutes of 20 Hz nerve stimulation at the pre-determined amplitude, and a 5-minute post-stimulation period for a total of 30 minutes. We asked participants to rate their level of subjective arousal on a 5-point Likert-style scale at four times throughout the recording trial: before baseline, before stimulation, after stimulation, and after the washout period. After the trial, we asked participants their opinion of the TENS device, if it elicited any genital sensations, and if they would consider using it.

All data analysis was performed in MATLAB (Mathworks, Natick, MA, USA). VPA signals for each participant session were processed before subsequent analysis and statistics across participants. We bandpass filtered the raw VPA signal from 0.5 to 30 Hz and identified peaks and troughs using MATLAB’s findpeaks function. We visually inspected the peaks and troughs for artifact removal. We removed obvious artifacts if they did not conform to the typical sawtooth shape²⁸ or had an trough-to-peak amplitude that had over a 100% increase from the previous waveform²⁹. The average percentage of data points removed was 14.3%. We calculated trough to peak amplitude and binned the data into 10 second intervals³⁰. Binned values were averaged for three time periods: 5-minute baseline (VPA_Baseline), 20-minutes of stimulation (VPA_Stim), and 5-minute recovery (VPA_Recovery). We made comparisons between these three periods within each participant with a one-way Analysis of variance (ANOVA) followed by post-hoc pairwise Tukey HSD tests. We calculated the percent change for each participant between each of VPA_Baseline, VPA_Stim, and VPA_Recovery, and made comparisons across participants for VPA_Change (VPA_Stim – VPA_Baseline) for each stimulation location with a paired t-test. We compared subjective arousal scores between each timepoint with a paired t-test. We compared baseline heart rate and mean arterial blood pressure to the last heart rate and blood pressure recorded during stimulation with paired t-tests. We compared the survey scores (SF-36, AUASI, PAC-SYM, FISI, and FSFI) across participants from the three different groups with pairwise t-tests. All statistical analysis used alpha = 0.05 to determine significance.

Results

We screened 101 participants for eligibility over the phone, of which 92 were either excluded, declined to participate, or were lost to follow-up. Ultimately, 3 participants in each group completed at least one study session. Participants in each group were lost to follow-up or became ineligible after the first session, leading to a total of five participants who completed two sessions. Figure 1 provides a breakdown of participant recruitment and retention. Demographics for all three groups of participants can be found in Table 1. Survey results averaged across each participant group can be found in Table 2. The SCI group and FSD group reported lower FSFI lubrication sub-scores than the NDAB group (p < 0.05). The FSD group reported significantly lower total FSFI scores than both NDAB and SCI groups (p < 0.005 and p < 0.05). All other survey scores were not significantly different between participant groups. The average stimulation amplitude for DGNS and TNS sessions was 28.8 ± 26.8 mA and 33.4 ± 24.4 mA respectively.

• Download figure
• Open in new tab

Figure 1.

Participant screening and recruitment.

In DGNS trials, we found significant increases in subjective arousal from before the trial to after the stimulation period (p < 0.05) across all participants (Figure 2a). These changes were also significant (p < 0.05) from before the trial until after the washout period. Participants in each of the three groups reported increased arousal during DGNS. There were no significant changes in subjective arousal across TNS trials (Figure 2b). There were no significant differences between heart rate or mean arterial blood pressure between baseline and stimulation in DGNS or TNS trials. The daily pelvic function surveys indicated that most participant’s bladder, bowel, and sexual function were stable and no participants reported carry-over effects from the stimulation session.

• Download figure
• Open in new tab

Figure 2.

Subjective arousal scores for dorsal genital nerve stimulation (DGNS) and tibial nerve stimulation (TNS) sessions. Icons represent the same individual participants. Icon color corresponds to group (indigo = able bodied, non-dysfunction, teal = SCI, rose = FSD). Solid black line denotes average values. Subjective arousal increased significantly (p < 0.05) in DGNS trials, but not TNS trials.

Most participants reported genital sensations in response to DGNS (7/8). Comments about receiving DGNS were mostly neutral and participants indicated that they would be willing to use a TENS device at the dorsal genital nerve. Half of participants (3/6) reported genital sensations during TNS and seemed just as willing to use TENS at the tibial nerve if they knew it would help with their sexual dysfunction. Table 3 contains all participant feedback.

VPA recordings from one participant in the NDAB group had corrupted data from both of their study sessions. All participants except one (FSD-3) had significant differences in their peak-to-peak VPA between at least two time periods in both DGNS and TNS sessions. These changes were not consistent within or across groups. Across DGNS sessions, two participants had an overall increase from VPA_Baseline to VPA_Recovery, four had an overall decrease, and one participant had no changes (Figure 3). Across TNS sessions, 3 participants had an overall increase from VPA_Baseline to VPA_Recovery, 1 participant had an overall decrease, and 1 participant had a decrease from VPA_Baseline to VPA_Stim that returned to baseline values in VPA_Recovery (Figure 4). The average VPA_Change was +3.5 ± 26.7% and +3.5 ± 14.0% for DGNS and TNS sessions respectively, which were not significantly different from zero. Average VPA_Baseline, VPA_Stim, and VPA_Recovery, as well as percent changes between each of the time periods for each participant, can be found in Tables 4 & 5. Subjective arousal and VPA data are publicly accessible online³¹.

• Download figure
• Open in new tab

Figure 3.

Boxplots of peak-to-peak VPA across different time periods (Baseline, Stim, and Recovery) in DGNS sessions. Boxplot central lines give the median, edges indicate the interquartile range, and dots represent outliers. Color and icons correspond to participants and groups as per Figure 2. * denotes significant difference (p < 0.05).

• Download figure
• Open in new tab

Figure 4.

Boxplots of peak-to-peak VPA across different time periods (Baseline, Stim, and Recovery) in TNS sessions. Color and icons correspond to participants and groups as per Figure 2. * denotes significant difference (p < 0.05).

Discussion

In this study, we sought to understand the effect of one-time transcutaneous DGNS and TNS on genital arousal, measured as vaginal pulse amplitude. We also investigated the effect of DGNS and TNS on subjective arousal. This is the first study of its kind to measure genital arousal in response to neuromodulation in women. We observed a significant increase in subjective arousal during DGNS but not TNS across patients, and varying effects on VPA across stimulation sessions and patient groups. Women in all three participant groups gave positive comments to receiving acute neuromodulation and women with complete SCI experienced genital sensations in response to DGNS and TNS.

Subjective arousal increased following stimulation across all groups of participants in DGNS sessions. As expected, the FSD group had lower subjective arousal than their NDAB and SCI counterparts after stimulation. Sexual function involves complex coordination of the parasympathetic, sympathetic, and central nervous system. There are several different proposed models of the sexual response cycle^32–34. Most models recognize that deficits in physiological areas (e.g., arousal) can influence the psychological (e.g., desire), making it difficult to pin-point the reason for lower subjective arousal in the participants with FSD. It is possible that the awareness of their FSD status in an isolated, clinical setting made them feel uncomfortable and led to lower subjective arousal. Participants with FSD experienced the fewest genital sensations and one participant (FSD-2) indicated that they might feel more receptive to using DGNS in a non-clinical setting (Table 3). Unexpectedly, some participants in DGNS sessions (3/8) had increases in subjective arousal before stimulation was turned on. This could perhaps be due to the presence of the VPA probe or anticipation of increased arousal.

All participants had significant changes in their peak-to-peak VPA between at least two periods (baseline, stim, or recovery) (Figure 3, Figure 4). Three participants (NDAB-3, FSD-1, and FSD-2) had opposing trends in their peak-to-peak VPA between DGNS and TNS sessions. It may be that not all women are responders to DGNS or TNS, as is common for neuromodulation therapies^35–37. There were no consistent VPA trends (decrease, increase, no change) within groups, which is not entirely unexpected given the small sample size. However, we did hypothesize that DGNS would increase VPA based on prior animal studies receiving stimulation at the pudendal nerve, the main trunk of the dorsal genital nerve. Cai et al¹⁷ theorized that pudendal nerve stimulation activates a spinal autonomic efferents via the pelvic nerve. Larger studies are needed to examine these trends further. For example, in women with SCI, it could be that one nerve target is more effective than the other depending on their injury level and severity.

As participants were not exposed to sexual stimuli during the study procedure, our VPA results are not directly comparable with prior VPA studies. It is possible that repeated neuromodulation sessions over time, as is common for percutaneous tibial nerve stimulation for bladder dysfunction³⁸, may yield consistent VPA changes as was found in a prior study that found repeated stimulation can increase FSFI scores¹⁴. Although animal studies have reported genital blood flow increases in response to tibial or pudendal nerve stimulation^15–18,39, those studies used anesthetized animals and directly stimulated the nerve, which limits direct comparisons to this study. An awake animal model study, such as the one by Zimmerman et. al in which they found increased sexual receptivity after 6 weeks of biweekly TNS⁴⁰, would provide a more analogous study paradigm to clinical studies.

Subjective arousal increased for one participant (SCI-2) in TNS sessions. It is likely that the mechanisms of TNS to modulate sexual function involve an indirect pathway to the pelvic organs, similar to the spinal reflexes proposed for bladder function⁴¹. Notably, this participant has a complete SCI (AIS-A) and reports that she does not experience any sensation below her level of injury (T5). It is possible that there are residual fibers in her spinal cord that are carrying afferents that were activated by DGNS or TNS. Another hypothesis is that afferents from the genitals during arousal, activated by DGNS or TNS, could be circumventing the spinal cord via the vagus nerve, leading to subjective arousal. One study in women with complete SCI asked participants to perform vaginal-cervical self-stimulation in a functional magnetic resonance imaging machine. Researchers found that the nucleus solitary tract (NTS), where the vagus nerve projects in the brain stem, was active during self-stimulation⁴². This is supported by an animal study that found neurons in the NTS that responded to vaginal distension and cervical stimulation⁴³.

Almost all women were willing to use DGNS or TNS outside of this study, provided it would help with their sexual dysfunction (if they had it). There was more hesitancy with DGNS as a therapy among the FSD group, perhaps due to the sensitive location of the electrodes. All participants who were lost to follow up after their first study session received DGNS, which may suggest that the method for delivering DGNS could be improved. Both nerve targets have distinct advantages. DGNS was able to increase subjective arousal but is applied at a sensitive location. Women may find they are more comfortable with stimulation at their ankle rather than their genitals and TNS is easier to administer at the ankle, however the mechanisms of how it improves sexual function are less clear.

Our study did not include any audio-visual materials, so we are limited in comparing our results to clinical studies that include videos with their interventions⁴⁴. We are also limited by our sample size (n = 3 per group), which prevented us from identifying trends within or across participants groups. We have now established a baseline for VPA response to DGNS and TNS in a one-time session, and although there were no trends in VPA across participants, individual participants had a VPA_Change as large as +70.7%. It is likely that the clinical environment with study team members present coupled with the lack of sexual stimuli contributed to an inhibition of sexual arousal, dampening genital arousal responses.

Conclusion

To our knowledge, this is the first clinical study to measure subjective and genital arousal in response to acute neuromodulation. We sought to understand the impact of two potential treatment modalities, transcutaneous DGNS and TNS, on subjective arousal and vaginal pulse amplitude during a one-time neuromodulation session. We found that DGNS, but not TNS, increased subjective arousal across all participants. We did not observe a consistent VPA response to DGNS or TNS across all participants or within participant groups. All SCI participants experienced genital sensations during DGNS and TNS sessions. Future studies may incorporate audio-visual materials or another type of sexual stimuli to better facilitate arousal. Studies with repeated stimulation sessions over time may find more clinically relevant improvements in sexual function.