Since Charles-Pierre Denonvilliers first discovered the firm, membranous structure between the rectum and prostate or bladder in 1836, now called “Denonvilliers’ fascia” (DF), the origin of DF has remained controversial. The most widely accepted theory is that DF represents fusion of the embryonic peritoneum of the retrovesical cul-de-sac [1,2]. Disorderly, loose connective tissue is present between the cul-de-sac and the rectourethralis muscle (RUM), with a tight, thick membrane that includes smooth muscle fibers between the cul-de-sac and the posterior aspect of the prostate near the base of the seminal vesicle [3]. DF is often fused with the prostatic capsule at the center of the posterior prostatic surface [4-6]. DF is not well adhered to the prostatic capsule toward the posterolateral aspect of the prostate. The space between DF and this capsule is filled by areolar tissue and the neurovascular bundle (NVB) [5-7]. Costello et al. [8] demonstrated plexus- or mesh-like nerves extending along the posterior aspect of the prostate after removal of DF using cadaveric dissection. Laterally, DF becomes continuous with the “pararectal fascia” posteriorly and the “lateral pelvic fascia” (LPF) anteriorly.

DF has been used as a surgical landmark during intrapelvic surgery for many years, and is one of the major areas of interest for urological surgeons for cancer control and functional preservation. Since the operative procedure for robot-assisted radical prostatectomy (RARP) was established by Menon et al. [9], it has widely expanded is use. RARP provides a precise, highly magnified view and less venous oozing, leading to a better anatomical understanding of fine membranous structures.

To the best of our knowledge, site-dependent and interindividual histological differences in DF have not previously been investigated. Therefore, this study aimed to examine site-dependent and interindividual differences in DF among 25 donated male cadavers and to determine whether changes in the current approach to radical prostatectomy are warranted in light of these histological findings. In the present study, DF was defined as a structure between the rectum and prostate, bladder, or seminal vesicles.

Twenty-five donated male cadavers (mean age at death, 85 years; range, 72–95 years) were examined. None of the patients had undergone abdominal surgery based on the medical records and macroscopic observations after opening the abdominopelvic cavity. These cadavers had been donated to Sapporo Medical University for research and education on human anatomy. Their use for research was approved by Sapporo Medical University ethics committee and all the participants provided written informed consent in their lifetime. Each donated cadaver had been fixed by arterial perfusion of 10% v/v formalin solution and stored in 50% v/v ethanol solution for more than 3 months. From each cadaver, one large tissue block including the bladder neck, seminal vesicles, membranous urethra, prostate, and rectal anterior wall, as well as connective tissue around these viscera, was obtained. After trimming the tissue block, routine procedures for paraffin-embedded histological examination were performed. Horizontal sections (15 cadavers) or sagittal sections (10 cadavers) were prepared at 2- to 5-mm intervals for hematoxylin and eosin staining (HE) and immunohistochemistry. When the prostate was large, the superior and inferior halves were included in two separate paraffin blocks. From each cadaver, 20–80 HE-stained and unstained sections were prepared.

Based on observations of HE-stained sections, some sections were used for immunohistochemistry and elastic–Masson staining (a variation of Masson–Goldner staining) [10,11]. The primary antibodies that were used to identify connective tissues and nerves were mouse monoclonal anti-human alpha-smooth muscle actin (1:100, Dako M0851; Dako, Glostrup, Denmark) and mouse monoclonal anti-human S100 protein (1:200 dilution, Dako Z0311; Dako), respectively. The smooth muscle antibody yields positive results not only for any smooth muscle, but also for vascular endothelium [12].

The study was performed in accordance with the provisions of the Declaration of Helsinki 1995 (as revised in Edinburgh 2000).

The present study showed that the configuration of the firm membranous structure of DF consistently involved multiple leaves. Even when a monolayer configuration appeared, this was restricted to a small area and actually comprised a thick leaf and other thin leaves. As Bertrand et al. [13] suggested, DF variations appeared to depend on differences in mechanical stress from either or both the prostate and rectum. Site-dependent and interindividual variations in DF were also observed, including the following: 1) variations in configuration behind the seminal vesicles; 2) variations in the connection with the LPF at the posterolateral angle of the prostate; 3) fusion of all or most DF leaves with the prostatic capsule near the base of the seminal vesicles; and 4) fusion of some DF leaves with the prostatic capsule anteriorly and/or the fascia propria of the rectum posteriorly.

Between the seminal vesicles and rectum, some DF leaves were clearly evident, while others were unclear or fragmented. During RARP, after bladder neck transection and dissection of the seminal vesicles and vas deferens, the connective tissue is present in the dorsum of the seminal vesicles and vas deferens. When the seminal vesicles, vas deferens, and prostate are pulled ventrally, DF is recognized as a membranous structure between the prostate and the rectum by its tension, irrespective of whether the leaves of the DF are clear or unclear histologically (Figure 5A).

Clear variation in DF morphology was also observed in the connection with the LPF at the posterolateral angle of the prostate on the posterior side of the NVB, as either clear, unclear, or absent. When leaves connecting the DF and LPF were unclear or absent, the mesorectal loose tissue appeared continuous with the NVB. Periprostatic nerves were not restricted at the posterolateral corner of the prostate, but their distribution extended medio-posteriorly in the DF [7,14]. These findings might be useful for understanding that the NVB is not restricted and that colorectal surgeons should pay careful attention to the anterior dissection plane of the rectum during total mesorectal excision.

Our study showed that all or most DF leaves were fused with the prostatic capsule near the base of the seminal vesicles. Notably, when the fascial space between the prostate and rectum was narrower without interposed mesorectal loose tissues at or near the midsagittal line, all or most DF leaves were fused with the prostatic capsule. Therefore, the dissection plane between the prostatic capsule and DF does not exist at the fusion site, because of the fascial complex between DF leaves and the prostatic capsule. Subsequently, periprostatic nerves were embedded in the combined capsule in a small area. Villers et al. [5] reported that muscular bundles and collagenous fibers of DF behind the vas deferens blend with central zone stroma and the ejaculatory duct sheath at the junction of the base of the prostate with the seminal vesicles and vas deferens. Kiyoshima et al. [6] reported that, in 79 surgically obtained specimens, DF was fused with the prostatic capsule at the center of the prostatic posterior aspect in 97% of cases. Hisasue et al. [15] reported the distribution of neuronal nitric oxide synthase positive nerve fibers, which are candidates for parasympathetic pro-erectile nerves [16]. They found that the distribution of these nerve fibers at 5- to 6-o’clock positions on the base of the prostate was 13.4% in the same hemisphere slice around the prostate in 23 specimens from non-nerve-sparing radical prostatectomy. However, what these nerves that are embedded in the fusion site innervate remains unclear. During RARP, DF is recognized as a membranous structure between the prostate and rectum by tension when pulling the seminal vesicles, vas deferens, and prostate ventrally. Simultaneously, the fusion site with DF and the prostatic capsule near the seminal vesicle-prostate junction are visible. DF between the seminal vesicles and rectum should first be cut at the midline to avoid entering the prostatic capsule. After this incision, a mesh-like structure behind the posterior aspect of the prostate corresponds with loose connective tissue and multiple leaves, and a flexible approach to the apex is easily found [17,18]. In cases of intra- or interfascial dissection, the dissection plane should be as close to the prostate as possible to avoid iatrogenic positive margins. Therefore, double cutting of DF while avoiding the fusion site is necessary to perform nerve-sparing procedures [19]. In cases of extrafascial dissection, the dissection plane unfolds at the loose connective tissue of DF to the apex. If advanced cancer is present at the border of the posterior aspect of the prostate, the dissection plane should be as close as possible to the rectal wall. This is because the fascial space between the prostate and rectum is not interposed with mesorectal loose tissues in some cases (Figure 5B) [17].

Distal to the fusion site, some DF leaves fuse with the prostatic capsule and/or the fascia propria of the rectum. DF reportedly conglutinates to the prostatic capsule with considerable frequency [4,5], whereas conglutination to the fascia propria of the rectum has not yet been reported, possibly because of difficulty in identification of the latter. During RARP, surgeons should be careful to avoid rectal injury during pre-rectal dissection because leaves of DF sometimes conglutinate to the fascia propria of the rectum.

DF converges on the grossly and histologically demonstrable posterior median raphe of rhabdosphincter and fibers of the RUM extended anteriorly into this fibrous raphe or central tendon of the perineum [20]. Soga et al. [21] reported that DF ends at the rhabdosphincter and the apical portion of the RUM. In our study, at least three specimens were found in which parts of the DF ended at the RUM. Of course, interindividual variations exist in the rhabdosphincter and RUM, and the shape of the rhabdosphincter is not always circular, but can show an omega shape [22]. In older men, the posterior rhabdosphincter is thin or absent in most cases. We suggest that the type of termination of DF depends on the size and shape of the rhabdosphincter and RUM [3].

Some potential limitations to this study should be considered when interpreting our findings. Variations of DF, such as the area in square millimeters and the numbers of leaves, were not evaluated quantitatively, because semiserial sections were used. Similarly, neither prostate size nor volume was evaluated before preparation of specimens for histological examination. Furthermore, because specimens including the rectum were difficult to obtain from young cadavers, no controls for likely changes with age were available.

Site-dependent and interindividual variations in DF were observed in donated elderly male cadavers. All or most DF is fused with the prostatic capsule near the base of the seminal vesicles and some DF is fused with the fascia propria of the rectum posterior to the apex of prostate. Based on our results, surgeons should be aware of variations and search for them to create a suitable dissection plane to avoid iatrogenic positive margins and rectal injury.