Structural Anatomy of the Posterior Pelvic Compartment As It Relates to Rectocele

Structural anatomy of the posterior pelvic compartment as it relates to rectocele

John O.L. DeLancey, MD Ann Arbor, Michigan

OBJECTIVE: This study was undertaken to define posterior compartment structural anatomy relevant to rectocele. STUDY DESIGN: Dissection of 42 fresh and 22 fixed cadavers was supplemented by examination of histologic (n = 3) and macroscopic (n = 5) serial sections. RESULTS: Distal posterior compartment support involves connection of the halves of the perineal membrane (urogenital diaphragm) through the perineal body, preventing downward protrusion of the lower rectum. Above this level the posterior vaginal wall is held in place by sheets of bilateral endopelvic fascia that attach each side of the posterior vaginal wall to the pelvic diaphragm. Most of these fascial fibers attach to the vaginal wall and a few fibers unite in the midline. Pelvic floor closure by the levator ani muscles relieves pressure-induced stress on the midvaginal fascial supports. CONCLUSIONS: Midline perineal membrane union supports the distal posterior compartment and a fascial connection between the pelvic diaphragm and vagina supports the mid vagina. Muscular pelvic floor closure helps to relieve fascial stress. (Am J Obstet Gynecol 1999;180:815-23.)

Key words: Endopelvic fascia, levator ani muscles, pelvic organ prolapse, rectocele

One in every 9 American women requires surgery for Table I. Anatomic specimens problems related to defective pelvic organ support, and No. examined Age range among these women 1 in every 4 needs a second operation.1 Among women with documented prolapse, 76% Dissections had defects found in support of the posterior compart- Fixed cadavers Embalmed cadavers 17 30–104 y 1 ment. Despite the common occurrence of rectocele, the Immersed cadavers 5 26–56 y structural defects responsible for its formation remain Fresh (unfixed) cadavers 42 7 mo–87 y poorly understood. Serial sections Histologic sections 3 0–29 y The pelvic organ support system has an anterior com- Macroscopic sections 5 14–64 y partment containing the urethra and bladder and a pos- TOTAL 72 0–104 y terior compartment containing the anus and rectum. The vagina, the uterus, and the endopelvic fascia that attaches them to the pelvic walls separate these 2 compart- specific sites of anatomic defects could be identified to ments and prevent their contents from protruding down- guide more precise scientific and clinical research. ward through the urogenital hiatus in the levator ani muscles. Material and methods The study described here was carried out to define the The anatomic materials used in this study are summa- detailed anatomy of the normal structural supports of the rized in Table I. Among the specimens 14 cadavers were posterior compartment that prevent anterior protrusion nulliparous, 40 were parous, and parity could not reliably of the rectal wall. This research was necessary so that the be established for 18, either from a review of hospital records when these were available or by physical examination. Six cadavers were African American and the re- From the Department of Obstetrics and Gynecology, University of maining 66 were white. Twelve had undergone hysterec- Michigan Medical Center. tomy, 1 had undergone a supracervical hysterectomy, Presented at the Seventeenth Annual Meeting of The American Gynecological and Obstetrical Society, Hot Springs, Virginia, September and 47 had intact uteri. Early in the study, during the first 3-5, 1998. 12 dissections, hysterectomy status was not noted. Reprint requests: John O.L. DeLancey, MD, L4100 Women’s Hospital, Anatomic findings reported are confined to those struc- 1500 E Medical Center Dr, Ann Arbor, MI 48109-0276. Copyright © 1999 by Mosby, Inc. tures and relationships that could be documented on dis- 0002-9378/99 $8.00 + 0 6/6/97081 section of both fresh and embalmed cadavers and verified

815 816 DeLancey April 1999 Am J Obstet Gynecol

AB Fig 1. A, Peripheral attachments of perineal membrane to ischiopubic rami and direction of tension on fibers uniting through perineal body (arrows). B, These fibers have been transected, exposing unsupported distal rectum. Note level II fibers are not shown. U-shaped contour of posterior vaginal wall in level III and W-shaped contour in level II are shown.

Fig 3. Sagittal histologic section of posterior vaginal wall and anorectum in 29-year-old cadaver showing thickening of connective tissue (outlined with dots) between vaginal wall and rectum in perineal body. Upper extent of this tissue is approxi- mately 2- to 3-cm area above hymenal ring. RVSp, Rectovaginal space; LMR, longitudinal muscle of rectum; IAS, internal anal Fig 2. Coronal macroscopic section of left side of 33-year-old sphincter. specimen showing connection at level of distal posterior vagina to ischiopubic ramus (IPR) by perineal membrane (PERIN MEMB) between vestibular bulb (VB) and levator ani muscle The overall arrangement of pelvic floor structures was (LAM). Note connection between levator muscle and perineal studied in specimens fixed by injection embalming. membrane. OI, Obturator internus muscle. Because these specimens are known to exhibit distorted spatial relationships,4 additional specimens were spe- with macroscopic and histologic serial cross sections. The cially fixed by an immersion technique that produces term vaginal wall is used to include the vaginal mucosa, specimens with topographic relationships that corre- submucosa, and muscularis. The term endopelvic fascia is spond to data available from living women.5 This used to denote those tissues between the vaginal muscu- process involves floating the specimen in formalin to laris and adjacent organs or the pelvic walls. Description of avoid gravity-induced sagging caused by loss of muscle different levels of support corresponds to those in an ear- tone. lier publication.2 Anatomic terms used conform to the Resistance of posterior vaginal wall support was stud- Nomina Anatomica.3 ied in fresh cadavers in a state not altered by fixation. Volume 180, Number 4 DeLancey 817 Am J Obstet Gynecol

A B

Fig 4. A, Macroscopic section of 14-year-old nulliparous cadaver. B, Histologic slide of Mallory trichrome–stained section of left half of 1-year-old infant. Note that most fibers of endopelvic fascia (outlined by dots) attach to lateral sulcus of posterior vaginal wall (VAG WALL) with only a small proportion of fibers connecting with fibers of contralateral side (asterisk). In B note origin of endopelvic fascia from superior fascia of levator ani muscle (LAM). OI, Obturator internus muscle; URETH, urethra.

During selected dissections (n = 6) in which rigor mortis and to elucidate the nature of attachments between vari- (as assessed by resistance to forearm extension) main- ous structures. tained levator ani muscle tone, downward force was applied to the pelvic floor. This was done first with a finger Results in the rectum to localize applied force in a specific direc- The distal rectum abuts against the dense connective tion and then with an air-filled transparent bag to apply tissue of the perineal body (level III). The perineal body pressure to the pelvis as a whole. In addition, in 3 of the represents the central connection between the halves of specimens fixed by immersion a string was threaded the perineal membrane (urogenital diaphragm). When from the origin of the puborectalis muscle on the inner the distal rectum is subjected to increased force directed surface of the pubic bones, parallel to the muscle fibers caudally, the fibers of the perineal membrane become passing dorsal to the anorectal junction and returning to tight and resist further displacement. These fibers derive the pubic bone on the opposite side. Tension on the their lateral support from their attachment to the pelvic string was adjusted to achieve normal urogenital hiatus bones at the ischiopubic rami (Fig 1, A, and Fig 2). This size6 and perineal body position.7 This technique served layer’s ability to resist downward displacement depends as a surrogate for puborectalis muscle tone and allowed on the structural continuity between the right and left interactions of connective tissue and muscle to be simu- sides of the perineal membrane. Transection of these lated and observed. fibers in the cadaver leaves the rectum exposed and al- The in situ relationships of the organs and their sup- lows the distal rectum (Fig 1, B) to prolapse downward. portive structures were studied in serial cross sections to The connection between the halves of the perineal supplement the observations mentioned previously be- membrane extends cranially for a distance of approxi- cause dissection by its very nature distorts the specimen. mately 2 to 3 cm above the hymenal ring (Fig 3). It is Macroscopic whole-pelvis cross sections were cut from thickest and densest in the distal perineal body, becom- frozen immersion-fixed specimens at intervals ranging ing progressively thinner toward its cranial margin. This from 5 to 20 mm. region (level III) of the posterior vagina is characterized Serial histologic sections were made available to me by by a U-shaped contour, in contrast to the W-shaped con- Dr Thomas Oelrich; their preparation has previously tour seen in level II (Fig 1, A). The upper extent of the been described elsewhere.8 These were used to confirm perineal body becomes confluent with the supportive the histologic nature of the tissues present, to determine structures of level II, and these 2 levels are structurally in- the specific direction of various fibers within the pelvis, terdependent. 818 DeLancey April 1999 Am J Obstet Gynecol

Fig 5. Lateral view of pelvic organs after removal of left ischial bone and ischial tuberosity. Bladder, vagina, and cervix have been cut in sagittal plane to reveal their lumens. Rectum has been left intact. A strip of posterior and lateral vaginal wall and its attached endopelvic fascia are shown, indicating their position relative to levator ani muscle and this fascia’s course and attachment. Both portions of levator ani muscle (puborectalis and iliococcygeus) are visible. Ischial spine and intact sacrospinous ligament (SSL) are above level of removed ischial tuberosity. Left half of perineal membrane (urogenital diaphragm) is shown just caudal to puborectalis portion of levator ani muscle after its detachment from inferior pubic ramus that has been removed. EAS, External anal sphincter.

The lateral margin of the perineal body contains the brane but also by the connection of the upper vaginal termination of the bulbocavernosus muscle. Caudally di- wall to the level II attachments that help hold the top of rected force does not put this muscle under tension, the perineal body (level III) in place. however, because of its sole anterior insertion onto the The directions of the connective tissue fibers that pro- clitoris. A few wispy striated muscle fibers, named the su- vide support in level III and level II are different. The perficial transverse muscles of perineum, lie along the poste- predominant direction of fiber flow in level III is from rior margin of the perineal membrane. side to side, spanning the gap between the ischiopubic The middle portion of the posterior vaginal wall (level rami. In level II the fibers change to a dorsal and cranial II) is attached on either side of the rectum to the inner direction, where they pull upward in a parasagittal plane. surface of the pelvic diaphragm by a sheet of endopelvic The upper portion of the posterior vaginal wall in level fascia. These fascial sheets attach to the posterior lateral I is attached to the pelvic wall by the sheetlike mesentery vaginal wall, where the dorsally directed tension results of the paracolpium (Fig 6). This anatomy has been de- in a posterior vaginal sulcus on each side of the rectum scribed elsewhere2 and will not be recounted here. (Fig 4). These endopelvic fascial sheets prevent the ven- When the dynamics of posterior compartment support tral movement of the posterior vaginal wall (Fig 5) and are examined with the puborectalis portion of the levator create the W shape of the posterior vaginal wall charac- ani muscle in a simulated state of contraction, the poste- teristic of level II. rior vaginal wall was in contact with the anterior wall (Fig Most of the endopelvic fascia fibers attach to the vagi- 7, A). In this situation, the pressures in the anterior and nal wall, with only a few fibers passing from one side to posterior compartments are balanced and there is little the other (Fig 4, B). Cutting the fibers of the endopelvic or no stress on Level II fascial supports. The force is car- fascia between the vaginal sulcus and pelvic diaphragm ried by the levator ani muscles and perineal body. When results in significant destabilization of the posterior the muscle is relaxed or damaged, the vaginal canal vagina in level II, whereas transection of midline fascial opens (Fig 7, B) and support is required from the con- fibers results in a much less dramatic change in support. nective tissues in level II (Fig 7, C), indicating an interac- The level II and level III supports are continuous with tion between the muscular and fascial supports. one another. Force applied to the anterior rectal wall in It is not possible to completely separate the effects of level II is resisted by the posterior vaginal wall and its at- muscular action from those of connective tissue support. tachments to the inner surface of the pelvic diaphragm. The levator ani muscles are directly connected to the Pressure applied to the perineal body in a caudal direc- upper surface of the perineal membrane in level III (Fig tion in level III is resisted not only by the perineal mem- 2). Loss of connection between the left and right halves Volume 180, Number 4 DeLancey 819 Am J Obstet Gynecol

Fig 6. Histologic slide from right side of 1-year-old cadaver showing endopelvic fascia comprising paracolpium (outlined with dots) attaching lateral vagina to pelvic wall in level I. Note differ- ence in this level from level II shown in Fig 4. LAM, Levator ani muscle. of the perineal membrane allows the levator ani muscles to move apart, and reuniting the separated ends of the perineal membrane restores the muscles to their normal positions. In addition the endopelvic fascia in level II arises from the superior fascia of the levator ani muscles (Fig 4, B). When muscle contraction is simulated the Fig 7. Concept diagram showing mechanics of support. A, connective tissue is elevated, indicating that these tissues Closure of pelvic floor by puborectalis muscle (large arrow) that compresses posterior vaginal wall against anterior wall. Increases work together in the pelvic floor as they do in the rest of in abdominal pressure result in balanced pressure on anterior the body. It should be noted that the levator ani muscles and posterior vaginal walls (arrows) so that no net force on sup- are lateral to the vagina and rectum at all levels and do port results. Caudally, however, there is no balancing pressure, not come to the midline (Fig 4) and that the muscle and force results (dashed arrow) that must be resisted by the fibers themselves do not form a layer that directly influ- fibers of the perineal membrane (shaded area) of perineal body. B, Absence of levator-mediated closure of pelvic floor. Increases ences the rectovaginal support. in rectal pressure are unopposed and force on posterior vaginal wall results (arrow). C, Level II supports oppose force shown in B Comment (dashed arrow) by their upward dorsal tension (arrows attached to This study reveals that posterior compartment support posterior vaginal wall and endopelvic fascia). is multifaceted. Rectal position is maintained by the in- terrelated actions of connective tissue (endopelvic fascia and perineal membrane) and striated muscle (levator port the structures when the levator ani muscles relax or ani muscle). Furthermore, connective tissue support dif- lose power through neuromuscular damage. fers at different levels. In the distal vagina the connective The arrangement of structures in the posterior com- fibers of the perineal membrane lie in a transverse plane partment reveals how the levator ani muscles contribute spanning the anterior triangle between the ischiopubic to posterior wall support. When the muscles have their rami, whereas in the mid vagina support comes from par- normal resting tone9 pressures in the anterior and poste- allel fascial sheets in the parasagittal plane connecting to rior compartment are balanced and no stress occurs on the inner surface of the pelvic diaphragm. Contraction the level II supports. It is the distal vagina that does not of the levator ani muscles, especially the puborectalis benefit from levator ani closure, and this explains the portions, closes the vagina, relieving connective tissue of dense connective tissue seen in the midline union of the constant load, whereas the connective tissue must sup- perineal membranes through the perineal body. In this 820 DeLancey April 1999 Am J Obstet Gynecol way the muscular and fascial supports are closely interre- between prolapse and levator ani muscle dysfunction.14, 15 lated. Second, level II insertion of the endopelvic fascia into Experienced surgeons recognize different types of rec- the vaginal wall helps to clarify the nature of interaction tocele and have emphasized the need for individually de- between the vaginal wall, pelvic diaphragm and en- signed treatment.10 The anatomic findings of this study dopelvic fascia. Finally, the relationships among the per- correlate with the different types of rectocele encoun- ineal membrane, levator ani muscles, and perineal body tered. Defects in midvaginal support can give rise to a help to define the nature of distal defects. rectocele, which may occur in the middle of the vagina There has been a long-standing debate as to whether despite normal levator function and an intact perineal there is a “surgically useful fascia” separating the rectum body. On the other hand a perineal rectocele11 may from the vagina.16, 17 This debate has existed because sur- occur below this region. This condition must be ad- geons find that they can dissect a layer between the dressed not by plicating the rectovaginal fascia but by re- vagina and rectum but histologic observations disclose trieving and reuniting the separated fibers of the per- little if any tissue between the vaginal muscularis and the ineal body. If the perineal body has separated from the rectal wall. This analysis of serial sections reveals that level II supports, they must be reunited. most of the endopelvic fascial fibers attach to the vaginal Scientific studies to determine how the type of fascial wall and that only a relatively few cross the midline. I be- and muscular damage influences objective treatment lieve that the plane dissected surgically and labeled as fas- outcome have not yet been carried out. Richardson12 has cia includes portions of the vaginal muscularis. Studies begun to make progress in this direction by calling atten- are ongoing in our unit to address this issue. Breaking tion to different sites of localized fascial disruption in- strength tests, which measure the amount of force volving the rectovaginal septum. The findings of this needed to overcome connective tissue supports, should study expand on these observations to describe the leva- be performed with and without separation of the midline tor ani muscle’s role in posterior compartment abnor- fascial fibers to determine the relative contributions of malities. This study also clarifies the perineal mem- the endopelvic fascia and the posterior vaginal wall to brane’s role in distal posterior compartment support. structural support. This anatomic picture provides an expanded list of struc- Concepts of the perineal membrane have undergone tures that must be assessed in determining the structural significant revision in recent years. This membrane has defects present when studying posterior compartment been shown to be a primarily fibrous structure and the problems. muscles previously presumed to be associated with it Classification of the numbers and types of anatomic (deep transverse perineus) have been revised and re- defects present in individual women should help in named as the compressor urethrae and urethrovaginal sphinc- studying surgical outcome. For example, the success ter. An excellent discussion of current concepts of this re- rates of rectocele repair in women with normal and de- gion can be found in Oelrich.8 fective levator ani muscles can be compared. I believe Anatomic examination of pelvic floor structure in ca- that it will be possible to develop magnetic resonance davers has limitations. I sought to minimize these inher- imaging scans capable of displaying specific sites of dam- ent problems. Examination of both fresh and embalmed age on the basis of the anatomic observations in this material lessens the likelihood of making errors as a re- study. This objective morphologic information, com- sult of embalming artifact, and studying cross-sectional bined with functional measures of levator strength and anatomy avoids distortions created by dissection. Loss of connective tissue properties, should permit more system- muscle tone after death has been addressed by studying atic study of surgical strategies and outcomes. Such stud- some cadavers during the phase of rigor mortis and cor- ies would allow the anatomic and physiologic factors as- recting for sagging caused by loss of muscle function sociated with operative failure and success to be clarified. through the use of flotation fixation. The greatest chal- This study adds to previous descriptions of the recto- lenge comes in assessing the role of the levator ani mus- cele anatomy that have emphasized a layer of tissue sepa- cles. The techniques discussed in the Methods section rating the vagina and rectum (rectovaginal septum).12, 13 are not an exact recreation of normal muscular function. These studies nicely describe the normal rectovaginal They are useful, however, in gaining general insights into septum and its connections to the pelvic wall and per- the critical interaction between muscle and connective ineal body. In addition they describe how detachment be- tissue. Now that some general idea about this anatomy tween the septum and perineal body results in specific has been established, further evaluations in living women types of rectocele. My findings highlight additional con- with magnetic resonance imaging and measurements cepts important to posterior compartment support. First, made with and without striated muscle paralysis can clar- the interaction between the levator ani muscles and con- ify the nature of interactions between muscle and connective tissue, wherein levator ani muscle tone relieves nective tissue supports. connective tissue stress, helps to explain the relationship Great strides have been made in gynecologic surgery Volume 180, Number 4 DeLancey 821 Am J Obstet Gynecol

by evaluating the scientific basis of such pelvic diseases as article describes anatomic findings only and does not dis- stress urinary incontinence and fecal incontinence. By cuss how this anatomy can be used in surgical manage- applying these same principles to rectocele, progress can ment of the rectocele. be made in understanding the pathologic anatomy re- I must admit that throughout my professional career I sponsible for the development of this common condi- believed—and taught students and residents—that a uro- tion. Once a more clear understanding of the nature of genital diaphragm existed that supported the anterior individual defects has been achieved it will be possible to pelvis and was formed by deep transverse perineal mus- study the relationships among clinical rectocele appear- cle with its anterior and posterior fascia. Dr DeLancey has shown us that the deep transverse perineal muscle ance, symptoms, and the underlying anatomic defects does not exist and that the previously recognized uro- present. Such knowledge will allow specific surgical genital diaphragm is almost all connective tissue, with strategies intended to correct anatomically localized de- only a few stray muscle fibers. He goes even further and fects to be carried out successfully. eliminates the term urogenital diaphragm and replaces it with a new name, the perineal membrane. He even further REFERENCES destroys my concept of perineal anatomy by stating that 1. Olsen AL, Smith VJ, Bergstrom JO, Colling JC, Clark AL. there is not any muscle in the perineal body between the Epidemiology of surgically managed pelvic organ prolapse and vagina and rectum, only condensed connective tissue. He urinary incontinence. Obstet Gynecol 1997;89:501-6. also states that the superior endopelvic fascia that I be- 2. DeLancey JO. Anatomic aspects of vaginal eversion after hys- lieved to exist between the vagina and rectum primarily terectomy. Am J Obstet Gynecol 1992;166:1717-28. 3. International Anatomical Nomenclature Committee. Nomina attaches to the posterior lateral vaginal wall, with only a anatomica. 5th ed. Baltimore (MD): Williams and Wilkins; 1983. few fibers united between the vagina and rectum. His pic- 4. Richter K. Lebendige Anatomie der Vagina. Geburtshilfe tures nicely illustrate these findings. I have certainly re- Frauenheilkd 1966;26:1213-23. vised my opinion of the pelvic support, and I wonder why 5. DeLancey JO. Standing anatomy of the pelvic floor. J Pelv Surg 1996;2:260-3. the older anatomists were so far off in their descriptions. 6. DeLancey JO, Hurd WW. Size of the urogenital hiatus in the le- By default I have become the pelvic reconstructive vator ani muscles in normal women and women with pelvic surgeon at the University of Virginia and perform several organ prolapse. Obstet Gynecol 1998;91:364-8. such repairs every month. Although the posterior repair 7. Henry MM, Parks AG, Swash M. The pelvic floor musculature in the descending perineum syndrome. Br J Surg 1982;69:470-2. is a common operation, it has not been my experience 8. Oelrich TM. The striated urogenital sphincter muscle in the fe- that 1 in 9 elderly women require this surgical proce- male. Anat Rec 1983;205:223-32. dure. Such an incidence would probably make it the 9. Parks AG, Porter NH, Melzak J. Experimental study of the reflex most common operation in elderly women. I do agree mechanism controlling muscles of the pelvic floor. Dis Colon Rectum 1962;5:407-14. that many women require a second repair, but I do not 10. Nichols DH, Randall CL. Vaginal surgery. 4th ed. Baltimore believe that this is because of tissue weakness or lack of a (MD): Williams and Wilkins; 1996. p. 257-89. known proper surgical repair procedure. I believe that it 11. Richardson AC. The anatomic defects in rectocele and entero- is because gynecologists who have not been adequately cele. J Pelv Surg 1995;1:214-21. 12. Richardson AC. The rectovaginal septum revisited: its relation- trained in technique or rarely perform it are doing the ship to rectocele and its importance in rectocele repair. Clin procedure. I am extremely concerned that decreased Obstet Gynecol 1993;36:976-83. surgical training in some residency training programs 13. Milley PS, Nichols DH. A correlative investigation of the human and the decreased financial incentive to refer patients rectovaginal septum. Anat Rec 1969;163:443-51. because of managed health care plans will only aggravate 14. Berglas B, Rubin IC. Study of the supportive structures of the uterus by levator myography. Surg Gynecol Obstet 1953;97:677- this problem in the future. Without question there is an 92. art to performing a lasting posterior repair. Long-term 15. Smith AR, Hosker GL, Warrell DW. The role of partial denerva- experience and gray hair helps. As a card-carrying gyne- tion of the pelvic floor in the aetiology of genitourinary pro- cologic oncologist, I am proud to say that most of us are lapse and stress incontinence of urine: a neurophysiological study. Br J Obstet Gynaecol 1989;96:24-8. superb pelvic surgeons; however, I am embarrassed that 16. Ricci JV, Thom CH. The myth of a surgically useful fascia in vagi- most are not trained to perform reconstructive pelvic nal plastic reconstructions. Q Rev Surg Obstet Gynecol surgery. In my opinion this big void in our oncology 1954;11:253-61. training programs has resulted in a loss of a pool of su- 17. Uhlenhuth E, Nolley GW. Vaginal fascia, a myth? Obstet Gynecol 1957;10:349-58. perbly trained gynecologic surgeons performing this po- tentially technical difficult surgical procedure. I am fear- ful that as we older self-taught reconstructive pelvic Discussion surgeons fade away there will be hiatus before adequate DR PAUL B. UNDERWOOD, JR, Charlottesville, Virginia. numbers of young reconstructive pelvic surgeons are This study involved the anatomic dissection of the poste- trained. I sincerely hope that I am wrong because there is rior lower pelvis of 64 fresh or fixed cadavers with macro- an enormous need as the number of older women grows, scopic and microscopic sections to document the find- their life expectancy increases, and their desire to feel ings. In addition, Dr DeLancey evaluated the effect of normal and function normally is paramount. intraluminal rectal pressure and simulated levator ani I have 2 questions to ask Dr DeLancey. (1) In light of muscle contraction on the posterior pelvic support. The the fact that the anatomic findings illustrate that the su- 822 DeLancey April 1999 Am J Obstet Gynecol

perior endopelvic fascia primarily attaches to the lateral they do not restore normal anatomy. I would agree that vaginal wall, rather than running between the vagina and this operative approach is not a particularly logical way to rectum, why does trimming away the mid vagina and do things on the basis of the normal anatomy, but there sewing the lateral vaginal wall together in the midline not are certain things about the rectocele anatomy that we correct a rectocele? We know that this does not work. cannot correct. I believe that much of the cause of recto- Would you please explain this discrepancy with your cele is muscular dysfunction, and that is uncorrectable at anatomic findings? (2) Because the perineal body does this time. So the strategy of compensating for defects is, I not contain any muscle but only condensed connective think, important in posterior wall support. tissue, is the new obstetric trend not to perform epi- The second question concerned the fact that the per- siotomies but rather to “stretch the perineum” going to ineal body does not contain muscle but rather connec- result in more posterior pelvic relaxations in the future? tive tissue. Dr Underwood asked about what happens DR ALFRED I. SHERMAN, Bloomfield, Michigan. The with the prolonged second stage of labor and attenua- causes of rectoceles, of course, are probably multifactor- tion of these tissues. The cervix goes from a diameter of 1 ial, and there are many reasons for them to develop. cm to 10 cm, and that connective tissue is able to re- Some women have a greater tendency than others. One assemble to form a normal cervix afterward. I think that area to which you are attending is that of the muscular as long as there is no mechanical disruption of the con- factors and the collagen factors, the support from below. nective tissue in the perineal body one could expect a I think that you also touched on the pressure from above, similar kind of recovery. the intra-abdominal pressure that exerts itself onto the If you look at the literature on connective tissue in- pelvic floor. I have always been interested in relating jury, injury arises from disruption of connective tissue. I these 2 factors with respect to the onset of rectoceles: the think that in this instance major disruption of this con- area of the pelvic inlet and the inclination of the pelvic nective tissue in a way that would prevent healing would inlet, which directs the forces from the abdominal con- be the mechanism of long-term obstetric damage. I think tents down into the pelvis. that more focused research in looking at the different DR FREDERICK B. STEHMAN, Indianapolis, Indiana. I mechanisms of connective tissue injury at the time of noticed that your cadaver specimens represented a broad vaginal birth certainly would be fruitful. range of ages. I wonder whether a range of parity was also Dr Underwood also talked about the term perineal represented. Can you comment on the differential im- membrane. This is, in fact, the current Nomina Anatomica pact of advancing age versus advancing parity in your term. It is not a term that I have coined. Its adoption re- anatomic specimens? flects the improvements in anatomic understanding of DR RICHARD C. BUMP, Durham, North Carolina. First, this area. you did not mention the relationship between the per- It is interesting that in the 1840s both correct anatomy ineal membrane and the posterior endopelvic fascia or fi- and incorrect anatomy were described at about the same bromuscular wall of the vagina. That tends to be among time. Unfortunately, the incorrect anatomy was perpetu- the areas on which we concentrate most in repair; do you ated for many years, but the Nomina Anatomica has now have any observations on that relationship? recognized the change in anatomy with the new term per- Second, please discuss the rectal wall itself and its im- ineal membrane. portance in the generation of rectoceles. We find a num- Dr Sherman pointed out the multifactorial nature of ber of patients on evacuation proctograms who have un- this mechanism that contains both muscle and connec- dergone successful posterior repairs but still have tive tissue. I agree that this is critically important to un- rectoceles either posteriorly or laterally or have internal derstanding the structural unit and the way that muscle mucosal prolapse of the rectal wall. and connective tissue interact. He asked specifically DR GREGORIO DELGADO, Pittsburgh, Pennsylvania. about the pressures on the pelvic floor. Pressure, as you When we do operations posteriorly, once we open the know, can be measured in pounds per square inch, so if perirectal space and the rectovaginal septum the first ele- you increase the number of square inches exposed to the ment that we confront is the rectal pillars, and right after same pressure you increase the force. I therefore believe that we encounter the uterosacral ligaments. What role that we will find that increases in pelvic size place greater do they play in the prevention of rectoceles? forces on the pelvic floor even though the pressures on DR DELANCEY (Closing). Dr Underwood’s first ques- the pelvic floor may be the same. I think that this will tion was why removal of the posterior vaginal wall is used prove an important factor in the genesis of pelvic organ in rectocele repairs if the tissues do not cross in the mid- prolapse. line. The approach described compensates for abnormal Dr Stehman asked about the issues of differences in anatomy rather than correcting it. Many of our opera- anatomy between the young cadavers and the old cadav- tions do not actually recreate normal anatomy but are ers and between the nulliparous cadavers and the multi- still effective. A familiar analogy is that stomach stapling parous cadavers. It is interesting that the neonatal and does not address the cause of obesity; it does, neverthe- young material is almost an ideal schema, a very clean less, correct overweight conditions. The operations that anatomy and very little distortion. What you see as you we use are often empirically derived, and the success of get into the middle-aged and older cadavers, in which these operations can be determined by outcome evalua- parity and age have intervened, is a series of different dis- tion. This is preferable to condemning them because tortions of that basic schema. We are in the process of Volume 180, Number 4 DeLancey 823 Am J Obstet Gynecol

starting to study those distortions to see which are nor- supports these areas. I think that these will be productive mal age-related changes and which are exceptions to the areas of investigation in starting to define the many dif- normal age-related changes. I think that the normal vari- ferent types of rectocele. When you have one morpho- ation is an important point into which we are just starting logic picture of a rectocele, you have to answer the ques- to get some insight. tion of the anatomic reason that has given rise to that Dr Bump made the important observation that there change in shape. I think that magnetic resonance imag- are some lateral and posterior rectoceles. I have been ing will be the tool to provide that answer. confining my remarks to the anterior rectocele at pre- Dr Delgado called attention to the rectal pillars that sent. The lateral margins of the rectum abut against the are seen at the time of a radical vaginal hysterectomy. I levator ani muscles, as does the posterior part; changes in think that these are the upper parts of vaginal support, rectal shape therefore come from changes in the rectal which go off of the vaginal wall on either side. I also think wall, which have been studied by the colorectal surgeons that these are areas of connective tissue that, with appro- to some extent, or from changes in the levator ani that priate traction, become visible in the operating room.

Receive the Journal’s table of contents each month by e-mail

To receive the tables of contents by e-mail, send an e-mail message to

[email protected]

Leave the subject line blank, and type the following as the body of your message:

Subscribe ajog_toc

You can also sign up through our website at http://www.mosby.com/ajog.

You will receive an e-mail to confirm that you have been added to the mailing list. Note that TOC e-mails will be sent out when a new isssue is posted to the web site.