Stein, Tamara A. MA1,3; DeLancey, John O. L. MD2,3
Pelvic floor dysfunction is a common problem that leads to significant suffering among women.1 Its origin during vaginal birth and surgical correction later in life unite the two primary elements of our specialty: obstetrics and gynecology. Understanding the nature of pelvic floor dysfunction must begin with an accurate anatomical appreciation of the structures that make up the pelvic floor. The perineal membrane (formerly called the urogenital diaphragm) is often discussed in both the delivery room and descriptions of gynecologic operations. In the process of examining pelvic organ supports and levator ani muscles,2–4 it became evident that existing descriptions of this region were at variance with the anatomy seen in human cadavers.
In his defining paper on the striated urogenital sphincter muscle in women,5 Dr. Thomas Oelrich provides a detailed history of articles, published from the mid-1800s to 1960, that investigate the urogenital sphincter complex, the adjacent “urogenital diaphragm,” and the deep perineal space. In his paper he disproved the concept of a urogenital diaphragm that consists of a layer of striated muscle with superior and inferior fascial coverings. He introduced the term “perineal membrane,” which is now the accepted anatomical term by Terminologia Anatomica.6 A recent review article summarizes articles on the same structures from 1960 to the present.7 Although some of the recent research reassesses the descriptions of the perineal membrane and/or the urogenital diaphragm, current knowledge of the structure is based primarily on historical descriptions and conceptual diagrams copied and recopied in anatomical and surgical texts.
The traditional description of the perineal membrane describes a trilaminar musculofascial structure that spans the ischiopubic rami and connects to the perineal body, through which the urethra and vagina perforate.8 More recent descriptions,7,9,10 although variable, do not agree with this three-layered model, clearly stating that the urogenital sphincter musculature is not sandwiched between superior and inferior fascial sheets. However, these more recent descriptions describe the inferior fascial layer of this “historical structure” as the perineal membrane (or triangular ligament): 1) extending between ischiopubic rami, 2) creating a specific boundary between superficial and deep perineal spaces, and 3) with the urethra and vagina piercing this fascial sheet.7,9,10 This does not correlate well with the anatomy seen in cadaver dissection, cross-sectional anatomical specimens or imaging studies of living women. Our objective was to directly examine the anatomy of this area in serial cadaver sections and dissection to see whether the concept of a sheet of fibromuscular tissue pierced by the urethra and vagina holds true or whether the perineal membrane is part of a complex of several interconnected structures.
MATERIALS AND METHODS
Serial trichrome-stained histologic sections of five female pelvic specimens (0–37 years old) were examined. Specimens included the urethra, vagina, perineal erectile tissue and muscles, and surrounding structures described in previous work of the senior author,2 as well as in Dr. Oelrich's original paper.5 The histologic sections were used to define tissue types (smooth muscle, connective tissue, striated muscle), specific fiber direction and connection, and the detailed structural spatial relationships. In addition, serial macroscopic whole-pelvis cross-sections of three adult females (28–56 years old) were studied. The pelves were fixed by flotation to avoid embalming artifact and then prepared by freezing the lower body and upper thighs in dry ice and cutting them on a band saw at 5-mm intervals in axial, sagittal and coronal planes.4 In addition, to correlate cross-sectional observations, gross dissection of the perineal region was conducted on a convenience sample of six Caucasian female cadavers, ranging in age from 48 to 90 years. These observations were supplemented with the senior author's experience in dissecting approximately 100 unembalmed cadavers while teaching pelvic anatomy to gynecologic surgeons.
The perineal membrane and its interconnected structures form a complex apparatus that is connected to many structures (Fig. 1). It is a three-dimensional mass of tissue with several attachments. There are two distinct regions: a dorsal portion lateral to the perineal body and a ventral portion lateral to the urethra. The dorsal portion consists of bilateral transverse fibrous bands of connective tissue that attach the perineal body and lateral wall of the vagina to the ischiopubic rami. This portion is bounded above by ischiorectal fossa fat and below by the structures of the perineum including the vestibular bulb, clitoral crus, and their investing muscles, the bulbospongiosus and ischiocavernosus muscles. In the ventral region, the membrane is part of a solid three-dimensional tissue mass in which several structures are embedded. It is continuous with the paraurethral and paravaginal connective tissues and contains the compressor urethrae and urethrovaginal sphincter muscles of the distal urethra. The ventral margin of this mass is continuous with the insertion of the arcus tendineus fascia pelvis into the pubic bone. The levator ani muscles are attached to the cranial surface of the perineal membrane complex, while the vestibular bulbs and clitoral crus are fused with the caudal surface. Medially, these structures fuse with the walls of the urethra and vagina.
Figure 2 demonstrates the changing nature of the perineal membrane in coronal section. Panel A shows the arcuate pubic ligament that spans the distance between the ischiopubic rami ventral to the first appearance of the perineal membrane complex seen in panel B. Here, the ventral portion of the perineal membrane complex is defined by the location of its surrounding structures. It lies cephalad to the vestibular bulb and the clitoral crus and caudal to the levator ani muscles. Medially, a portion of the wall of the urethra can be seen, as well as the paraurethral portion of the vagina. It is evident in these images that the ventral portion of the perineal membrane is an integral component of these interrelated structures, not a freestanding structure.
Panel C lies at the level of the vaginal lumen where the transition from the ventral to the dorsal portion of the perineal membrane complex occurs. At this point the structure becomes a distinct layer of connective tissue between the lateral vagina and the ischiopubic ramus, although still maintaining its close relationship to the levator ani muscle. Like the traditional paradigm, this dorsal portion is distinct from its surrounding structures, creating a “deep space”; that is, ischio-anal fossa. The ischio-anal fossa defines the upper margin of this dorsal portion, and the erectile structures (vestibular bulb and clitoral crus) its caudal margin.
Panels D and E show the dorsal portion of the perineal membrane complex. The fibromuscular layer of the posterior vaginal wall (panel D) and the perineal body (panel E) are attached to the ischiopubic ramus via this fibrous dorsal portion of the perineal membrane. The distance between the midline viscera and the lateral ischiopubic rami has widened in this region compared with panels A–C, so that the perineal membrane is a distinct structure with less interconnectedness between surrounding structures.
Reviewing the panels from B to E reveals the constant anatomical relationships. The viscera are medial, the ischiopubic rami are lateral, the erectile structures are caudal, and the levator ani muscles cranial in the ventral sections, and the ischio-anal fossa cranial in the dorsal sections.
Figure 3 demonstrates the histologic detail in coronal section of the ventral portion of the perineal membrane. Panel A shows the ventral-most portion of the striated urethral sphincter muscle and the compressor urethrae/urethrovaginal sphincter muscle complex. Panel A indicates, via shading, the location of the pubic bone that was removed to facilitate cutting the histologic sections. Note the attachment of the arcus tendineus fascia pelvis to the pubic bone indicated by the asterisk. There is no evidence of any perineal membrane structures in this section.
Panels B and C show the ventral margin of the perineal membrane complex developing. Notice that the superior fascia of the levator ani muscles is continuous with the insertion of the arcus tendineus fascia pelvis (panel B), which is continuous with the paraurethral connective tissue supports (panel C), which is, in turn, continuous with the ventral portion of the perineal membrane complex (panels D and E). As the distance between the midline viscera and the pubic bone increases (panels C, D, E), the connective tissue mass of this ventral portion of the perineal membrane increases. Also, notice the pudendal neurovascular bundle embedded in this tissue (panel C).
Panels D and E show the continued development of the 3-dimensional mass of tissue; the perineal membrane. As seen in previous ventral panels, this mass of tissue is continuous with the superior fascia of the levator ani muscles and with the paraurethral connective tissue or endopelvic fascia. The compressor urethrae and urethrovaginal sphincter muscles are embedded into the medial portion of this tissue mass, while the pudendal neurovascular bundle is embedded into the lateral portion near the periosteum of the pubic bone. Also, notice that the clitoral crus and the vestibular bulb and their respective muscles are fused to the caudal surface of the ventral portion of the perineal membrane complex.
Panel F is a section through the vaginal lumen, at which point the transition from the ventral to the dorsal perineal membrane complex occurs. The distance between the midline viscera and the lateral pubic bone has increased, and therefore, the connective tissue mass has increased. It is continuous with the superior fascia of the levator ani muscles, the paravaginal connective tissue, and, likely, the periosteum of the pubic bone as well. The levator ani muscles are inserting into this connective tissue mass cranially, while the clitoral crura and vestibular bulbs, with their respective muscles, are inserting into or are fused with the tissue mass caudally.
Figure 4 demonstrates in sagittal section the insertion of the levator ani muscle fibers into the portion of the perineal membrane lateral to the vagina. Panel A is a parasagittal section through the vaginal lumen, perineal body, and the ventral portion of the internal anal sphincter muscle. Ventral to the vagina, the lateral portion of the striated urethral sphincter and the ventral midline portion of the compressor urethrae and urethrovaginal sphincter muscle complex can be seen.
Panels B and C are sections lateral to that in panel A and show the vaginal lumen becoming the lateral vaginal wall and its connective tissue (panel C). Similarly, the perineal body also becomes connective tissue as the sections move laterally (panel C). This connective tissue lateral to the vagina and perineal body, within which the compressor urethrae and urethrovaginal sphincter muscles are contained, is the perineal membrane structure.
Panels C and D show the levator ani muscle fibers inserting into the cranial surface of the perineal membrane complex. These sections demonstrate that the structures in this region are not separate entities but are interconnected parts of a single tissue mass.
Figure 5 shows the axial sections of the same cadaver whose coronal sections are shown in Figure 2. A comparison of the coronal and axial sections from the same cadaver demonstrates that axial section is not favorable to defining the relationships of the complex three-dimensional structure of the perineal membrane, its attachments to the pubic bone, perineal body, and pelvic viscera. Note the lower vagina, midurethra, levator ani, and obturator internus muscles. Panel B shows the ventral portions of the perineal membrane lateral to the urethra and vagina. The superficial transverse perineal muscle marks the dorsal edge of the perineal membrane. Although this muscle is not part of the complex, it lies on the caudal surface of the dorsal edge of the perineal membrane where fibrous connective tissue bands connect the perineal body to the ischia. The section in panel C is caudal to the perineal membrane.
Although the perineal membrane has been described for well over a century, opinions have differed widely regarding its structure and relevance to pelvic organ support. In this study, rather than commenting on terminology or comparing previous descriptions, we have directly examined the anatomy of the perineal membrane and its surrounding tissues. The importance of this structure as a surgical landmark and its functional role in posterior vaginal wall support make it imperative that an accurate description be available. Examination of serial cross-sections revealed that the structure called the perineal membrane (previously referred to as the urogenital diaphragm) is a complex structure that is only one component of a larger interconnected support apparatus. In addition, this study reveals that the perineal membrane has two distinct parts, a dorsal portion and a ventral portion, and that the levator ani muscle is intimately connected with this structure.
Because the perineal membrane is attached to so many surrounding structures, reconnecting the two sides, as occurs during posterior colporrhaphy, can realign these structures into their normal configuration. The dorsal portion consists of bilateral connective tissue fibrous bands, reaching from the lateral vagina and perineal body to the dorsal portions of the ischiopubic rami, that can become separated after childbirth. In this region, the levator ani muscles are directly connected to the perineal membrane, the perineal body, and the bulbospongiosus muscles. The levator muscle fibers insert directly into the perineal membrane connective tissue bands, and, ventrally, the superior fascia of the levator ani muscles is continuous with the perineal membrane tissue mass. The attachment of the levator ani muscles to the perineal membrane and perineal body means that disruption to the midline connection between the perineal membranes of each side through the perineal body4 allows loss of perineal body support and also lateral displacement of the perineal membrane. Because the levator is fused with this area, this would result in widening of the urogenital hiatus seen in women with prolapse.4 This provides an anatomical support for the idea that posterior repair, by bringing the separated perineal body and membrane attachments back to the midline, can restore the normal alignment of the levator ani muscles.
The different parts of the perineal membrane complex suggest different functions. It would appear that the dorsal portion is related to the support of the perineal body and lateral vaginal wall through its attachment to the ischiopubic rami.4 It seems likely that, during the second stage of labor, the dorsal portion, by attaching the vagina and perineal body, participates in holding these structures in place while the infant's head dilates the introitus. It is often stated that the perineal membrane supports structures during normal activities, but this is not logical since considerable descent of the perineal body during straining occurs in normal women when the levator ani muscles are relaxed.11 Therefore, it may be implied that the perineal body, via its attachment to the dorsal portion of the perineal membrane complex, limits downward motion when the levator ani muscles are relaxed, but that it is normal muscle tone that maintains perineal position.
The ventral portion is contiguous with the urethral supportive apparatus as previously discussed by Milley and Nichols.12 Whether or not the urethral supports and perineal membrane are considered different supports or the same structure is an arbitrary distinction about which experts will probably always have honest disagreements. The real need is for an accurate understanding of the structural mechanics of this region. This need is not addressed by arguments about nomenclature. In this region, it is not possible to understand structure and function under the historical paradigm that the perineal membrane is a sheet of tissue.
Although there has been some recent research and reassessment of the perineal membrane,7,10 the traditional, century-old paradigm persists in most anatomy and surgical textbooks.7 Recent research clearly states that there is no deep transverse perineal muscle in the female and that the urogenital diaphragm concept of urogenital sphincter muscles sandwiched between two layers of fascia is also incorrect.5,7,10 Our study agrees with these two points. However, these recent reassessments of the perineal membrane continue to describe a single sheet of fascia spanning the ischiopubic rami, denoting a boundary between superficial and deep perineal spaces. Oelrich5 describes a thin layer of perineal membrane separating the superficial compartment containing the erectile tissues and their muscles from the “urethrovaginal compartment,” which is continuous with the pelvic cavity. A more recent paper describes the perineal membrane as extending between the pubic arch and the ischiopubic rami, creating a boundary between the superficial and deep perineal spaces.7 Our study describes a ventral portion of the perineal membrane and emphasizes the fact that the perineal membrane is not a free-standing structure in this region. It is part of a multifaceted complex of interconnected tissues and structures. It does not create a deep perineal space in the ventral portion. It is difficult to visualize a structure that is a solid tissue mass intermingling with other tissues and structures when the structure is called a “membrane.” The fact that misleading terms can lead to misleading concepts has long been known, having been captured by the father of the scientific method, Francis Bacon, over 300 years ago: “Whereas the meaning ought to govern the term, the term in effect governeth the meaning.”13 This has certainly been true of the term urogenital diaphragm.
We should consider why this area, smaller than the palm of a hand, has been so confusing. Dissection in this area is difficult because of the dense connective tissue present and because there are not natural cleavage planes that allow structures to be easily separated from one another. Perhaps as important, however, has been the longstanding practice of considering the female a variant of male anatomy. In the male, the idea that the perineal membrane is a continuous sheet perforated by the urethra is not only plausible, but mechanically sound. In the female the situation is completely different because of the presence of the vagina. It is interesting that in the area of the human body that actually defines the anatomical differences between the sexes, the presumption that one sex is a variant of the other has been applied. The anatomy in the male and female must be separately described.
There are limitations to any anatomical study that must be considered in interpreting their results. Anatomical material, either embalmed or fresh, has topographic distortion because of fixation and loss of muscle tone.14 Magnetic resonance imaging scans show anatomy in normal women that is helpful in understanding topographic relations but are limited in resolution of fine detail and only show part of the structures involved. We have chosen to focus primarily on cross-sectional anatomy because it avoids the distortion that dissection creates and also avoids the “creative” portion of dissection where dissection is often carried out to make a structure look like what textbooks suggest. We recognize that examining cross-sections can be challenging for those who do not have prior experience in evaluating serial cross-sectional images, but cross-sectional images remain the gold standard for evaluating pelvic floor structure, as Lawson15 has previously pointed out. Careful study of these images allows the undisturbed anatomy to be studied. We have especially depended on histologic cross-section because it allows tissue type (striated muscle, smooth muscle, dense connective tissue, loose areolar connective tissue) to be characterized. This study has addressed the constant features seen in all specimens. There is considerable individual variation seen in this region. A larger study will be needed to address these variations and their relation to such issues as vaginal parity. For dissection, we have chosen adult cadavers in the age ranges relevant to the occurrence of pelvic organ prolapse.