Endometriosis is a disease which continues to be poorly treated. Drugs are still used to treat advanced disease when there is absolutely no evidence for their efficacy. Hysterectomy is still performed for a disease that by definition is “extra-uterine”. Myths and false concepts are carried from one textbook to the next.

Endometriosis is considered to be a modern disease, and does not rate a mention in any contemporary textbook of medical history. It has even been suggested that endometriosis is the result of industrialisation, yet Daniel Shroen1 described endometriosis in great and graphic detail in 1690. He described it as a disease affecting the pelvic surfaces, the broad ligament, the bladder and the bowel. He described its progressive nature, and its tendency to form adhesions.

He spoke about its vascularity, or the recently rediscovered importance of angiogenesis. He described endometriosis as a disease with onset at puberty, and adolescent endometriosis is still denied as a real entity by many modern gynaecologists. Finally he claimed no originality for his work and spoke about endometriosis in the tone of a commonplace disease.

There has been much debate about the aetiology of endometriosis in general, and specifically about deep rectovaginal endometriosis. Sampson’s theory (1927) of retrograde menstruation and cell implantation continues to be regarded as the basis for most peritoneal endometriosis and there is much circumstantial evidence in support of this. In 1919, Meyer suggested that coelomic metaplasia or transformation of pluri-potential mesenchymal cells from one cell type to another might account for some endometriosis.

Indeed it is otherwise impossible to explain the occasional reports of endometriosis occasioning in male subjects. Halban (1925) suggested that lymphatic dissemination of endometrial cells may occur, and Sampson (1927) once again acknowledged that vascular dissemination probably accounts for occasional reports of widely disseminated endometriosis. Iatrogenic transplantation to surgical wounds in the abdominal wall has also been described.

More recently, researchers have explored a range of molecular cellular defects that are involved in the development of endometriosis. The evidence for an immunological basis to the disease continues to be explored, and some environmental toxins, notably dioxin, have been implicated – either through a direct effect upon steroid hormones or their receptors, or an indirect effect upon immune system function. Peritoneal responsiveness has not been greatly studied, but may be contributory to this disease.

Without much doubt, there is a genetic basis to endometriosis, and many researchers around the world are attempting to identify the genetic aberrations responsible for this disease. Allow me to return to this concept towards the end of this dissertation, as I do believe that this will prove to be the common thread in unravelling the complex aetiology of endometriosis.

The cause of deep recto-vaginal endometriosis is a subject of particular debate. In 1908, Cullen proposed that recto-vaginal disease occurred as a result of direct extension of lower uterine adenomyosis into the rectovaginal septum. While this may occasionally be true, it is unlikely to explain most cases of a common disease. Incredibly, in 1899, Russel talked about “embryonic cell rest activation” which I interpret to be the same as the current theory of “mullerian duct remnant metaplasia”.

Nissole and Donnez2 believe that embryonic remnants in the recto-vaginal septum undergo metaplastic change to “endometrial-like” tissue, and by proliferation become surrounded by hyperplastic smooth muscle, representing a typical “adenomyotic nodule” deep in the recto-vaginal septum. Koninckx3 supports this view with evidence of altered oestrogen and progesterone receptor activity in these nodules compared with typical peritoneal endometriosis.

Vercellini4 approached this debate by measuring the Pouch of Douglas (POD) depth and volume of women with and without recto-vaginal  endometriosis. He found that women with deep endometriosis had POD depth measurements on average 1/3 less than those without deep disease, and therefore believed that this was a primary peritoneal disease, extending into a pseudo-recto-vaginal septum caused by adhesion formation between the anterior rectal wall and anterior POD peritoneum.

My own observation is that rectal endometriosis often arises from infiltrative utero-sacral disease, with involvement of the lateral rectal wall. It seems equally likely that peritoneal POD disease causes the majority of anterior rectal disease with rearrangement of POD anatomy by adhesion formation. Whatever the cause, the principles of management remain the same.

These principles depend upon an understanding of the pathophysiology of progressive endometriosis.  Early peritoneal endometriosis must establish its own blood supply, and in fact the pelvic distribution of endometriotic disease may have some dependence upon the ability of disease to induce angiogenesis. Once established, in a proportion of patients, disease is progressive. Growth cycles induce recurrent inflammatory response and ultimately fibrosis.

Some forms of disease are marked by extreme fibrotic response so that a fibrotic capsule comes to enclose implants. Even surface peritoneal disease often involves much thicker fibrotic peritoneum than first assessment would suggest. Inflamed peritoneal surfaces lead to adhesion formation in some patients. Peri-ovarian adhesions, colo-rectal adhesions and adhesions between pelvic organs, omentum and small bowel predominate. Surprisingly, direct tubal involvement is relatively uncommon.

The clinical presentation of classical endometriosis is well known to gynaecologists. Dysmenorrhoea, pelvic pain at other times of the cycle, abnormal uterine bleeding and particularly premenstrual spotting, and deep dyspareunia are common symptoms. Alternatively about 30 – 40% of patients with endometriosis may present with infertility – and sometimes these patients are otherwise relatively asymptomatic.

About 10% of patients with endometriosis will have intestinal involvement. They present symptoms of deep rectal pain, often worsened with bowel movements or exacerbated at the time of menstruation. Disturbance of bowel habit manifesting as alternating diarrhoea and constipation is common, and with more advanced disease tenesmus or a feeling of incomplete rectal emptying may occur.

Rectal bleeding may be noticed, and while commonly occurring at the time of menstruation, this is not always the case. Mucosal involvement with endometriosis is rare. I believe that most bleeding results from trauma to the rectal mucosa overlying   endometriotic deposits, or from venous congestion in association with an endometriotic nodule.

Patients with symptoms consistent with inflammatory bowel disease or with rectal bleeding should undergo colonoscopy. Very rarely will endometriosis be diagnosed at colonoscopy. Rather, this is to exclude alternative diagnoses, or to identify those patients with annular luminal constriction secondary to endometriotic fibrosis.

Diagnosis may only be made by laparoscopy. Even then deep rectovaginal endometriosis may easily be missed when the cul de sac is obliterated. One needs to introduce an instrument into the posterior vaginal fornix when its outline should be clearly visible. A sponge holder is ideal for this purpose and the isolated cul de sac obliteration will become apparent. Deep fibrotic nodules will also be palpable on vaginal or combined rectal / vaginal examination.

Furthermore palpation of the utero-sacral ligaments with a laparoscopic instrument will sometimes allow nodules arising from deep within these structures to become apparent. Assessment of rectal tethering can be enhanced by the introduction of a rectal probe or sponge forceps into the lumen of the bowel. The extent of muscularis involvement of serosal sigmoid deposits can also be assessed with the aid of a probe, as the muscular layers of the bowel wall will be seen to move independently of deposits limited to the serosa.

Treatment approaches to advanced endometriosis in recent times have been most illogical. No amount of progestogen and no degree of oestrogen suppression will permanently eradicate the endometrium, so how could we expect such hormonal approaches to eradicate endometriosis. Temporary suppression is the only outcome, and while symptomatic improvement may occur, recurrence of symptoms following cessation of treatment is almost invariable.

No effect on the associated fibrosis occurs, and I know of no drug of any class that will eradicate scar. This probably accounts for the fact that a percentage of women with deep fibrotic endometriosis fail to experience even temporary symptom relief with hormonal suppressive therapies.

The development of operative laparoscopy saw many gynaecologists turn to surface ablation of peritoneal disease. Many modalities have been used, including laser energy delivered to peritoneal surfaces, or electrosurgical energy delivered by direct electrode application or via Argon or Helium beams. While this may effectively eradicate the earliest of surface disease, any deeper disease will be inadequately treated.

Furthermore it is not possible to assess the depth of disease merely by surface visualization. Some delivery systems are marketed on the basis of their limited depth of penetration, and hence their safety regarding underlying critical structures. This is inconsistent with eradication of a disease that is invasive in all but the earliest of cases.

In order to avoid treating just the “tip of the iceberg”, the only logical approach must surely be surgical removal or excision of disease. Disease overlying the ureter dictates that the ureter be dissected free and retracted from the surgical site.  Disease overlying or involving the bowel dictates that this organ be mobilized, and dissected free from the disease. Herein lies a problem. Certainly in my country, the surgical training of most Gynaecologists has not equipped them for such dissections.

So patients continue to be treated with surface ablation for deep disease, avoiding the danger sites, with resulting treatment haphazard at best. Even worse they are subjected to multiple courses of hormonal therapies that are known to be ineffective, interspersed with multiple laparoscopies to “see how there disease is going”. These are about the only patients who do not need a laparoscopy to diagnose endometriosis.

In desperation, hysterectomy and bilateral oophorectomy is often recommended to relatively young women with advanced endometriosis. This is often performed leaving deep disease remaining. Not surprisingly such women often remain symptomatic.

The requirement for hysterectomy should be determined by the presence of uterine disease. Coexistent adenomyosis, uterine fibroids or extensive serosal involvement with endometriosis may make hysterectomy advisable, but removal of the normal uterus is to be avoided. Similarly, ovarian involvement with endometriosis and its extent should dictate the need for oophorectomy.

I now wish to discuss the surgical principles involved in the conservative excision of deep endometriosis.  This is almost always possible to achieve as a laparoscopic procedure, with the advantages of faster patient recovery, less post-operative adhesion formation, and most importantly improved identification and access to deep disease. The disadvantage of course is the more prolonged operating time. Furthermore there is no doubt that this is complex and difficult surgery, and certainly not without risk of serious operative morbidity.

Let us assume a patient with pelvic sidewall disease, invasive uterosacral disease and a deep rectovaginal nodule with bowel tethered centrally to the posterior cervix. There is no ovarian involvement.

The patient has been thoroughly bowel-prepped preoperatively, insufflation to 12mm Hg has been established and an intra-umbilical 5mm endoscope has been introduced. Three further 5mm ports have been placed, one in each iliac fossa, and one suprapubically in the mid-line.

I have two laparoscopic grasping forceps (one toothed), a pair of laparoscopic scissors through which unipolar current can be delivered, a suction irrigation probe with an irrigation pressure generator connected via wide bore tubing and a bipolar coagulating forceps. We also have available, if required, a 5mm clip applicator and laparoscopic needle holders.

The first principle is to commence dissection in an area of normality. The second principle is to use constant tissue traction during dissection and the third is to use a combination of high power density cut (unmodulated) current and sharp scissor dissection. For this reason I prefer to deliver electrosurgical current through the laparoscopic scissors.

The electrosurgical generator is set at 100W pure unmodulated current and coagulation effect can be achieved by varying power density by altering the amount of electrode in contact with the tissue. The bipolar output is set to 35W, and similarly delivers unmodulated current automatically.

The ureter is identified through the peritoneum high on the pelvic sidewall and the peritoneum opened linearly above it. The ureter can then be swept off the peritoneum in an area where it is healthy and non-adherent. The peritoneum is then retracted medially while dissection around all areas of disease occurs with about a 0.5cm margin. Particular care need be taken where overlying peritoneum is adherent to the ureter, but it is virtually always possible to shave the ureter free from overlying disease.

The peritoneum is always found to be thicker and more fibrotic than expected, reinforcing my belief that excision of disease is the only logical approach. As one proceeds towards the side of the uterus, vascular injury becomes of increasing concern. The uterine artery will usually become apparent in its tortuous course over the top of the ureter, and the uterine veins lateral to the insertion of the utero-sacral ligaments become particularly prone to injury. The ureter at this point lies more laterally and is less susceptible to injury.

Next the para-rectal spaces most be opened. An uninvolved area of recto-sigmoid is chosen, and the peritoneum opened on the medial side of the utero-sacral ligament. This space may be safely opened quite deeply although bleeding from para-rectal vessels will always be encountered. This bleeding can be controlled with the application of electrosurgical energy through bipolar forceps. Bleeding from the rectal sidewall can also be controlled with bipolar energy although it is frequently prudent to use “liga-clips” to minimise inadvertent thermal injury to the rectal wall.

Once the rectum has been mobilised laterally, the most difficult part of the surgery commences. It is necessary to mobilise the rectum from the posterior cervix until the areolar tissue of the normal recto-vaginal septum is reached. It may be possible to find a tissue plane between the nodule and the posterior cervix / vagina, and it may be possible to find a plane between the nodule and the rectum. More commonly the dissection proceeds through the nodule, leaving disease both on the posterior cervix / vagina, and on the anterior rectal wall.

With these spaces opened, it is now possible to excise uterosacral disease to whatever depth necessary to achieve eradication. Much reliance is placed upon trans-laparoscopic palpation of tissues to confirm that all uterosacral disease has been removed.

Residual rectal disease must now be removed. It is usually possible to shave such disease from the anterior rectal wall, but in cases of deep muscularis involvement a disc excision may be required. Occasionally, with the greatest of care the rectum may be inadvertently opened. Only when there is rectal stricture formation from fibrosis, or extensive endometriotic disease involving more than 1/3 of the rectal circumference would anterior segmental resection be considered. In this instance, complete mobilisation of the rectum, resection of the involved segment and either trans-anal staple anastomosis or mini-laparotomy for a hand-sewn anastomosis is undertaken.

Regarding repair of a rectal defect, any hole greater than 1cm should be closed in a transverse direction. This prevents hourglass stricturing of the rectum that could lead to significant functional disturbance.

Rectal defects may be closed by single layer interrupted sutures or larger defects may be closed quickly with an ENDO GIA multifire stapling device. The articulating head stapling devices are particularly suitable for this, as they facilitate the desired transverse closure.

Finally, any residual disease on the posterior cervix / vagina must be removed. This may be shaved laparoscopically, or with deeply invasive disease may be more efficiently removed via a vaginal approach. Excision of a segment of vaginal wall may be required with primary vaginal closure.

If a defect in the bowel wall has been repaired, I would leave a Penrose drain from the site of repair, exiting through one of the lower lateral port sites. This should be left for 5-7 days.

Before removing the ports, the pelvic and peritoneal cavity should be copiously irrigated with saline and haemostasis checked. It is wise to finally check haemostasis after deflation of the pneumoperitoneum for a period of time. The pneumoperitoneum pressure may otherwise tamponade venous bleeding sites that only become apparent after release of this pressure.

This outlines the approach to excision of advanced endometriotic disease. This is often long and complex surgery, but symptomatic response to this surgery is most rewarding. Redwine and Perez5 have reported a series of more than 500 cases of excisional surgery for advanced endometriosis. These patients have been followed for up to four years with significant and sustained improvement in symptoms.

The great challenge for the future lies in the prevention of new disease or the prevention of recurrence in treated disease. I predict that this will occur either by genetic manipulation or immune system modification. There is little doubt that endometriosis has a polygenic basis, with various genetic aberrations driving various metabolic or immunologic abnormalities. While such research proceeds, it is less likely that these approaches will offer hope for established, advanced fibrotic disease. Primary surgical approaches for this problem will continue to have a place for a long time in the future.


  1. Shroen D. Disputatio inauguralis medica de ulceribus uteri. Jena: Krebs, 1690:6-17
  2. Nisolle M, Donnez J. Peritoneal endometriosis, ovarian endometriosis, and adenomyotic nodules in the rectovaginal septum are three different entities. Fertil Steril, 1997; 68:585-596
  3. Koninckx PR
  4. Vercellini P, Aimi MD et al. Deep endometriosis conundrum: evidence in favour of a peritoneal origin Fertil Steril, 2000; 73:1043-1046
  5. Redwine DB, Perez JJ. Pelvic pain syndrome: endometriosis and midline dysmenorrhoea. In: Arregui ME, Fitzgibbons RJ, Katkhouda N, McKernan JB, Reich H, editors. Principles of Laparoscopic Surgery – Basic and Advanced Techniques. New York: Springer Verlag, 1995:545-558.