Introduction

Continence is a variable of the ‘trifecta’ outcome following radical prostatectomy (RP), and its recovery is a crucial issue postoperatively. The urethral sphincter (US) plays a critical role in urinary continence after radical prostate cancer surgery [1]. Like any other musculature, its optimal function requires maintenance of supporting structures and functional muscular mass, as well as proper angulation and neural coordination for periodic relaxation and prolonged constriction. While the US anatomy has been studied extensively, its nerve supply and coordinated function is far from clear, as there are long-lasting conflicting hypotheses about the nature and course of its nerves [2–4].

Some researchers reconcile different views by suggesting that the US may be supplied by a combination of somatic and autonomic innervation or that possible neural communications may optimize and coordinate functioning [5]. Some have gone on to propose triple innervation involving somatic, parasympathetic and sympathetic supply [6]. It has even been suggested that somatic innervation would travel through the pelvic plexus, thereby facilitating a non-pudendal nerve course for somatic fibers [7]. In addition to the source of innervation, lack of detailed knowledge of the course of the relevant nerves remains a stumbling block for prostate cancer surgeons trying to improve functional outcomes after RP.

Regardless of the surgical approach—open [8, 9]/laparoscopic [10, 11]/robotic [12, 13]—or of the year of publication, the controversy about the effect of nerve sparing on continence continues and remains a topic of much debate. Reeves et al. have recently reviewed 13,749 RP in 27 studies and reported that nerve sparing was improving continence only in the first 6 months [14]. However, a recent prospective and multicenter study has reported an association between the degree of bundle preservation and urinary incontinence 1 yr after open or robotic surgery [15].

It is likely that return of continence is a multifactorial process involving coordinated muscular contraction of the sphincteric complex, a compliant and capacious detrusor reservoir, optimal angulation and support of the vesico-urethral junction, vascularity and suppleness of the vesico-urethral anastomosis, length of the sphincteric tube, type and duration of the drainage, and the patient’s age. Herein, we review the important clinical anatomy of the neural supply to the urethral sphincter and discuss the implications for continence recovery during radical prostatectomy.

Methods

To comprehensively review the neuroanatomy of the urethral sphincter, we searched the PubMed database for articles published in English from January 1985 to September 2015 using each of the following keyword: ‘urethral sphincter,’ ‘urethral rhabdosphincter,’ ‘urinary continence’ in combination (Boolean operator [AND]) with each of the following terms: ‘innervation,’ ‘nerve supply,’ ‘neuroanatomy,’ ‘nerve sparing.’ No other PubMed filter than English language was used. The selection was then made by the first and last authors. First, a negative selection was made by reading the abstracts: animal studies, non-anatomical methodology, surgical series and case reports were excluded from the review. Second, a positive selection was made by reading the articles: dissections of several adult bodies, histological analyses or reconstructions of several adult bodies or fetuses, intraoperative stimulations studies were included in the review. Third, another similar selection process was made on: the first 5 ‘similar articles’ suggested by PubMed for each of the formerly selected articles, and on each of its references published between January 1985 and September 2015. To ensure a comprehensive review, the following data were noted for each of the selected articles: anatomical methodology, number and gender of bodies examined, data, type of figures, relevant facts and text. Relevance was appraised in consensus between first and last authors and defined as: anatomical facts in relation with the innervation of the US. The information was grouped into somatic and autonomic innervations. For each pathway, a semiquantitative (from 0 to +++++) evaluation of quality of the reviewed data was performed for: number of subjects (none; 0–10; 10–30; 30–90; 90–200; > 200), illustrations (none; low-resolution image in 1 study; low-resolution images in several studies; high-resolution image in 1 study; high-resolution images in several studies); histology (none; tissue staining; solely an aspecific neural antibody (PS100); 1 autonomic antibody; several autonomic antibodies); electronic microscopy (none; axonal morphology criteria in 1 study; axonal morphology criteria in several studies; neuron or tissue analysis associated in 1 study; neuron or tissue analysis associated in several studies). Presented herein are the findings of this comprehensive review.

Results

Seventeen articles have been selected for review. Articles solely based on per-operative data, articles based on less than three anatomical subjects and data acquired on women or female fetuses were excluded from the review. Results were organized into a comprehensive anatomical presentation based on the main possible neural pathways that supply the US: somatic, autonomic and communicating branches (Fig. 1). For each, we will discuss the possible course of nerves according to the current literature.

Fig. 1
figure 1

Potential neural pathways to urethral sphincter (US) innervation

Full size image

Somatic innervation

The somatic motor innervation of the US arises from the cell bodies of sacral spinal cord segments S2–S4 and is organized in Onuf’s nucleus. The somatic branches travel to the target structures either through branches of the pudendal nerve or through nerves that travel with branches of the pelvic plexus. The nerves to the male US can be classified as either extrapelvic branches of the pudendal nerve, intrapelvic branches of the pudendal nerve or branches of the inferior hypogastric plexus (IHP).

Pudendal pathways: extrapelvic branches

Eight anatomical studies based on 108 conventional dissections of adult male bodies and the histological examinations of 7 fetuses have described the extrapelvic pudendal pathway (Table 1; Fig. 2) [16–23]. This is the most established neural pathway for the extrinsic US (EUS). The PuN, after its origin, travels through the greater sciatic notch to course anteriorly within the pudendal canal (of alcock). The ischial spine is a common anatomical landmark in all anatomical studies. PuN branches provide the inferior rectal, perineal, posterior scrotal and sphincteric nerve fibers. A first set of sphincteric fibers commonly enters the prostatic urethra at the 9–12 o’clock and 1–3 o’clock positions, less than 10 mm away from the prostatic apex, making them at risk for damage during apical dissection. A second set of sphincteric fibers branches from the dorsal nerve of the penis and reaches the sphincter retrogradely, in an anterolateral position, making them at risk for damage during the dorsal vein complex stitching.

Table 1 Somatic pudendal extrapelvic nerve pathways to the US
Full size table
Fig. 2
figure 2

Pudendal pathways—extrapelvic branches. Branches of the pudendal nerve in the perineum after removal of the coxal bone. Lateral aspect. The branch to the urethral sphincter region from the dorsal nerve of the penis is shown (black dot), and the branches to the deep transverse perinei are observed (white star). Bl bladder, dp dorsal nerve of the penis, Dtp deep transverse perinei muscle, La levator ani muscle, pd pudendal nerve, Pn penis, Pr prostate. Exact caption from the original article by Akita et al. used with permission [16]

Full size image

Pudendal pathways: intrapelvic branches

Four anatomical studies based on 81 conventional dissections of adult male bodies have described the intrapelvic pudendal pathway (Table 2; Fig. 3) [16, 19, 20, 24]. The intrapelvic branches exit, while the PuN is in the pudendal canal, and they travel through the levator ani muscle (LAM) toward the EUS near the prostatic apex. Intrapelvic branches of the PuN enter the EUS at a 5 or 7 o’clock positions. In the largest studies, this pathway is highly variable and only observed in 30–40 % of the dissections whereas the most constant (60 % under the operating microscope) pathway remains extrapelvic.

Table 2 Somatic pudendal intrapelvic nerve pathways to the US
Full size table
Fig. 3
figure 3

Pudendal pathways—intrapelvic branches. Pelvic nerve and intrapelvic branches of pudendal nerve (right). Rc rectum, Pr prostate, Pni intrapelvic branches of pudendal nerve, Pen pelvic nerve, La levator ani. Exact caption from the original article by Song et al. [20] used with permission

Full size image

Non-pudendal pathways

Five anatomical studies based on 88 subjects were reviewed. Three studies with 53 conventional dissections in adult bodies have illustrated that some of the somatic sacral 2, 3 and 4 branches arise from the lowermost root of the pelvic splanchnic nerve and travel with the autonomic fibers of the pelvic plexus and innervate the US (Table 3; Fig. 4) [16, 18, 20, 23, 24]. In two studies based on the examination of 35 fetuses, Karam et al. [25] and Takenaka et al. [26] observed and distinguished both myelinated and unmyelinated nerve fibers traveling along the bladder neck to the urethra, suggesting dual innervation of these structures. All fibers were at the posterior face of the bladder neck and followed the same course, penetrating the US: from its anterolateral surface for the myelinated fibers classically considered as somatic; from its posterolateral surface for the unmyelinated fibers classically considered as autonomic. The non-pudendal somatic pathway to the US has been conceptualized by Akita [16].

Table 3 Somatic extrapudendal nerve pathways to the US
Full size table
Fig. 4
figure 4

Non-pudendal pathways. Branches from the sacral roots to the levator ani running inside the pelvis—cadaveric dissection. Exact caption from the original article by Zvara et al. [23] used with permission

Full size image

Autonomic innervation

Thirteen studies involving 219 subjects (adult bodies, deceased fetuses or living patients) have, over the years, progressively raised the level of evidence to support the role of autonomic nerve fibers in EUS innervation (Table 4; Fig. 5) [16, 17, 20, 21, 24–32]. Conventional dissections have been the first method to identify nerve fibers reaching the EUS from the neurovascular bundles (NVBs). These fibers were observed at the posterolateral aspects of the prostate apex, along the rectum covered by the LAF, as they originate from the most caudal edge of the inferior hypogastric plexus (IHP) or from the neurovascular bundle (NVB) itself. Basic histological studies have then established the unmyelinated feature of these fibers, which was an argument in support of their autonomic nature. With more refined immunolabelings and electron microscopy, it has been possible to characterize sympathetic and parasympathetic fibers to the EUS. Few intraoperative electrophysiological studies have also given credit to the hypothesis that the NVB contains sphincteric fibers, as the intraurethral pressure increases in response to electrical stimulation.

Table 4 Autonomic nerve pathways to the US
Full size table
Fig. 5
figure 5

Autonomic innervation. a Macroscopic dissection of so-called right NVB in fixed cadaver. NVB contains many nerve fibers to cavernous tissue (arrowhead), urethral sphincter (arrow) and bottom of levator ani muscle (star). b Frontal histological section around so-called right NVB stained with hematoxylin and eosin. Some nerve fibers go to cavernous tissues and urethral sphincter between membranous urethra and levator ani muscle fascia. H&E stain. Exact caption from the original article by Takenaka et al. [32] used with permission

Full size image

Somatic–autonomic communications

Five of the previously reviewed studies have illustrated and classified topographical communications between the somatic (infralevator) and the autonomic (supralevator) pathways in 95 adult bodies or fetuses (Table 5; Fig. 6) [16, 17, 20, 24, 26]. The more proximal one follows these fibers, the more communications were constant (94 %); the more distal, the less communications were constant (10 %). A common trunk to the LAN and the PSN was the first and the most consistent shared pathway by different nerve fibers. The levator ani muscle (LAM) also appeared to be a crucial anatomical structure as it is pierced by fibers from the autonomic and somatic systems, and it could thus have an underestimated role in urinary continence. The urethral sphincter was described as uniquely innervated through both autonomic and somatic innervations that may coordinate the complex function of urinary control and voiding during a wide variety of stress and bladder filling [17].

Table 5 Nerve communications supplying innervation to the US
Full size table
Fig. 6
figure 6

Somatic–autonomic communications. Communicating branches between pudendal nerve, pelvic splanchnic nerves and nerves to the levator ani nerve (left). Bl bladder, La levator ani, Lan levator ani nerve, Psn pelvic splanchnic nerves, Cb communicating branches. Exact caption from the original article by Song et al. [20] used with permission

Full size image

Discussion

The US is assimilated to a muscular complex with smooth and skeletal muscle fibers involved in passive and active continence [33]. Specifically, the external US is said to be more amenable to prolonged contractions. It envelops the prostate and urethra from the vesical orifice base to the membranous urethra. This crescent-shaped structure above the verumontanum transitions into a horseshoe shape wherein the muscle is thickest ventrally [34]. The rectourethralis muscle and external US are in juxtaposition, and along with the levator ani muscle and the bulbospongiosus muscle, converge along the median fibrous raphe. Though they are very close in proximity, a distinct layer of connective tissue separates the levator ani muscles and external urinary sphincter muscle [35]. US innervation has long been reduced to a somatic pudendal extrapelvic nerve supply, and the present review rather pleads for a complex or a variable innervation. With as many descriptions as reviewed articles, consensus regarding the urethral sphincter’s innervation has not been reached over the last three decades. We have comprehensively reviewed the anatomy and evidence from all relevant studies in this article. The elegant reconstruction by Stolzenburg’s team (Fig. 7) and the informative schema by Akita et al. (Fig. 8) have enabled an anatomical-surgical conception of the US innervation pathways (Fig. 9) [16, 36, 37].

Fig. 7
figure 7

Autonomic nerves branching from cavernosal nerves to the urethra (3D visualization). The cavernosal nerves of the penis emerge from the neurovascular bundles and divide into medial and lateral branches after penetration of the muscular pelvis. The medial branches innervate smooth muscle component of the external urethral sphincter; the lateral branches continue to enter the cavernosal bodies. Exact caption from the book chapter by Schwalenberg et al. [36] (book by Stolzenburg et al.) used with permission

Full size image
Fig. 8
figure 8

Schematic presentation of the positional relationship between the nervous branches to the rhabdosphincter (black dot) and the levator ani. The somatic nervous branches to the rhabdosphincter might have two routes to reach the muscle (dashed line): one is a course from the pudendal nerve, and the other from the pelvic plexus. The pelvic splanchnic nerve (black square) forms a common trunk (asterisk) with the nerve to the levator ani (black triangle). A connecting branch between the nerve to the levator ani and the pudendal nerve (black star) and a connecting branch between the pudendal nerve and the branch to the rhabdosphincter from the pelvic plexus (double black stars) are sometimes observed. Co coccygeus, dp dorsal nerve of the penis, hg hypogastric nerve, La levator ani, pd pudendal nerve, px pelvic plexus, Rc rectum, Rs rhabdosphincter, Ut urethra. Exact caption from the original article by Akita et al. [16] used with permission

Full size image
Fig. 9
figure 9

Integrated representation of the possible US innervation pathways. Surgeon’s (AKT) conception (bottom left frame) with anatomical (TB) reproduction. Somatic pathway in dark blue; neurovascular bundle in green; pelvic plexus in red; communications in light blue. Co communicating branches, CN cavernous nerve, CS, colliculus seminale, CT common trunk of the LAN and PuN, DNP dorsal nerve of the penis, EDs ejaculatory ducts, EUS external urethral sphincter, HN hypogastric nerve, LAF fascia of levator ani, LAM levator ani muscle, LAN levator ani nerve, NVB neurovascular bundle, P prostate, PF pelvic fascia, PPx pelvic plexus, PuN pudendal nerve, Re recurrent branches of the DNP, SoPPx somatic pelvic plexus, SN spongious nerves, LSN lesser sciatic notch, SV seminal vesicle, TLA translevator ani branch, U urethra

Full size image

By comparing all of the selected studies, we were able to estimate different degrees of certainty regarding each anatomical claim. Credibility was estimated by the number of subjects and studies, the extent of the dissections, the quality of the illustrations, the study design, the methodologies and the antibodies’ reliability at the time of publication, the logical reasoning and even the vocabulary used by authors in their demonstration, which we summarized in a gray-scale reading template (Table 6). In particular, great attention was paid to the histological protocols. Conclusions drawn on multicontrolled immunolabels [17] were considered more reliable than simple stainings [19, 25, 26], whole-pelvic studies were rated higher than restricted regions of interest on the EUS [25, 30], and intraoperative stimulations [18, 31] were more valuable than postoperative stimulations [27].

Table 6 Level of evidence of the nerve pathways to the US
Full size table

In the controversy of EUS innervation, anatomical facts add up and strengthen the possibility of a somatic and autonomic co-innervation. In the 1980s, little credit was given to an autonomic component of the EUS innervation because anatomical evidence was missing, but it was not factually excluded. In the 2010s, the debate is still ongoing but, with cumulative evidence for nerve supply to the EUS through both autonomic and somatic innervation, it seems the burden of proof has changed. In the meantime, the anatomical concept of a sole innervation for a striated muscle is also changing for the LAM [29]. An autonomic participation to the LAM innervation is now described and a coordinated role of the LAM and the EUS, both eventually being mixed autonomic and somatic-innervated, becomes consistent. Autonomic innervation originates from the hypogastric and pelvic plexuses, whereas somatic innervation arises through both pudendal and non-pudendal branches. These branches may play a role in preventing muscle fatigue and thereby facilitate the early return of urinary continence [29]. These branches supply not only the sphincter but also supporting muscles such as the levator ani and rectourethralis, which can contribute to continence indirectly.

In the reviewed descriptions, different anatomical landmarks appeared of noteworthy surgical interest. Serving as a boundary between intra- and extrapelvic branches, the fascia of levator ani (FLA) has its lowest point at a mean distance of 5 mm from the sphincteric branch. The distance between the pelvic floor and the nerve entry point decreases with opening and exposure of the LAM fibers beneath the FLA, thus making sphincteric nerve fibers vulnerable to injury during pelvic floor surgery [22]. At the prostate apex, the extrapelvic somatic sphincteric branches enter the EUS at a distance of ranging from 4 to 11 mm, with risk for damage during apical dissection [20, 21, 23]. Steiner has refined surgical techniques to minimize stress manipulation of the EUS and to preserve these nerves in the name of continence preservation [38]. At the era of open prostatectomy, he recommended to avoid using a right-angle clamp to establish the plane between the posterior urethra and underlying rectourethralis and median fibrous raphe, and his concept remains timely for laparoscopic and robot-assisted RP.

The return of continence following RP is a multifactorial process [39] that involves the length (and bulk) of the urethral sphincter, integrity of supporting structures, appropriate angulation, proper functioning of the bladder as a reservoir, correct technique of the vesico-urethral anastomosis, disruption of the retropubic space and reperitonization of the bladder in its normal location. Over time, the body attempts to heal these anatomical changes, and continence recovers in the majority of cases. Technical refinements that target early return of continence vary. Techniques include preservation of the bladder neck, reconstruction of the posterior support, reconfiguration of the bladder opening, and realignment and reconstruction of supporting ligaments. Several authors have advocated anterior, posterior and total reconstruction to assist in early return of continence. It may be that like other factors, the contribution of nerve sparing to early return of continence is supportive, and by 12 months postoperatively, healing and compensation by other mechanisms facilitate recovery. In expert hands, this impact may become minimal as surgeons develop several compensatory techniques that may overcome the minimum neurological handicap that may occur due to nerve damage. We have summarized possible nerve routes to the sphincter that may help surgeons in visualizing their locations intraoperatively during radical prostatectomy (Fig. 2).

The possible locations of continence nerve injury are summarized in the flowchart that also discusses the possible mechanisms and strategies for avoidance even when performing non-nerve sparing surgeries (Fig. 10) [11, 13, 20, 38–42]. From the surgical technique standpoint, both somatic and autonomic nerves are at risk of damage, due to either wide excision at the level of the seminal vesicles or the peri-prostatic dissection, or at the level of the apex where convergence occurs and the nerves are within a few millimeters of the dissection plane and suture bites. The somatic and autonomic nerves travel within the layers of FLA and could sustain injury if the resection plane is sufficiently wide. Visualizing the existence and course of these nerves may help surgeons minimize the damage to these potentially important structures. These nerves could be damaged not only by surgical excision but also by cautery, clips or sutures that are used to control bleeding.

Fig. 10
figure 10

Technical refinements in nerve sparing to avoid incontinence

Full size image

Conclusion

It is important to understand the complex innervation of the US such that incontinence associated with radical prostate cancer surgery can be minimized. It appears that the complexity of the neuroanatomy of the US include significant cross-communication, redundancy and possibly neuroplasticity. It seems that both somatic and autonomic systems contribute to the innervation, and pelvic surgery poses risks to the integrity of many of these nerves. The autonomic nerves travel with the neurovascular bundles or travel separately or as communication with somatic nerves. The somatic nerves are shown to come from both pudendal and non-pudendal pathways. Specifically, nerves can get damaged at higher levels, during seminal vesicle dissection, or at lower levels, during periprostatic and apical dissections. Research that continues to elucidate the neuroanatomy of the US will enable surgeons to avoid destruction of these important structures and thereby minimize the incontinence associated with pelvic surgeries.