• The first laparoscopic image above shows a very early stage of rectovaginal endometriosis. There is a vertical ridge in the middle of the pouch of Douglas between the two uterosacral ligaments. This signifies the rectum being pulled up towards the upper part of the vagina and the back of the cervix. Note how the pelvis otherwise looks clean. Deep endometriosis is better felt than seen.

• The second image is a more advanced stage of rectal endometriosis than the first image. The rectal wall itself was invovled as shown with the dark deep area pulled up. A more proximal part shows superficial involvment of the rectal serosa with red endometrioisis. This patient had severe dyspareunia and dyskesia.
  
  

Urinay tract symptoms of endometriosis

The urinary tract was estimated to be involved with endometriosis in 20% of the cases with the bladder being the most affected site. This figure is equal to the risk of endometriomas yet it is not usually addressed by gynaecologists. It usually involves the peritoneal and subperitoneal tissues in the uterovesical pouch leading to bladder symptoms without any abnormal findings at cystoscopy. Accordingly, inspection of the uterovesical pouch during laparoscopy for that purpose is very important. It could show as actual endometriotic lesions or scarring of the peritoneal fold making it less stretchable when pulled up with a pair of grasping forceps. Symptoms related to bladder involvement include: 

• Lower abdominal or suprapubic pain felt continuously or in relation to menstruation
• Increased frequency of micturition due to irritation of the bladder wall by endometriosis
• Abdominal pain when the bladder is full
• Lower abdominal pain during urination
• Suprapubic pain during intercourse due to mechanical irritation of the bladder bed
• Rarely haematuria would be a late sequel following deep invasion of the bladder wall.

The first laparoscopic image shown below illustrates scarring of the uterovesical pouch with endometriosis before being excised, as shown in the second image. The diagnosis was confirmed histologically in this case. Sometimes only few red or black lesions can be seen on the surface. Pulling the uterovesical peritoneum with a grasper can reveal the extent of the subperitoneal lesions. This may cause difficulty during dissection of the involved area because of subperitoneal tissue scarring.

The ureter may be involved with endometriosis, but this is usually superficial on the overlying peritoneum or outer serosal layer. Deeper involvement can lead to stricture formation, following severe scarring of the neighbouring tissues. This can be released after excision of such tissues, without operating on the ureter itself. Furthermore, such scarring can put the ureters at risk of being damaged during gynaecological surgery. The lower part of the ureter is especially prone to be involved with endometriosis of the uterosacral ligaments. On the other hand, rectovaginal nodules larger than 3 cm in size are also prone to involve the ipsilateral ureter. Most of the blood supply feeding the ureters comes from small blood vessels on their lateral sides. Accordingly, dissection of the ureters may be vascularly safer, if done from their medial aspect.

The first two laparoscopic pictures above show involvement of the right and left ureters with endometriosis, respectively. Both ureters were seen only at their proximal parts, near the pelvic brim. The third image shows the left ureter after being cleared. It was a very difficult procedure. Diagnosis was confirmed histopathologically.

The above shown laparoscopic image depicts the right pelvic sidewall, with the right ureter kinked in its distal part, before being crossed by the right uterine artery. This patient had no related symptoms to this findings, but future complications leading to proximal hydro-ureter and right kidney problems were most likely, had the kink was not discovered and relieved.

As a general rule, ureterolysis should always be started at a healthy area, usually near the pelvic brim. Care should be taken not to devitalise the ureter during the procedure by over zealous pealing of the ureteric sheath, and hence its blood supply. Ureteric peristalsis is a reassuring sign of ureteric wall integrity following such surgery.

A special association between endometriosis and interstitial cystitis (IC) is becoming more evident over the last few years, without the bladder itself being involved with any endometriotic lesions. More information will be given about this subject later on during this website.

Other symptoms related endometriosis

It is not unusual for patients with endometriosis to present with other symptoms on top of their chronic pelvic pain and abnormal uterine bleeding. This could add to their sufferings and affect their quality of life. It could be useful to take note of the following symptoms and deal with them in a total illness management plan:

·   A recent study showed women with endometriosis to be twice at risk to suffer from migraine headaches (38.3% vs. 15.1%) at a younger age of onset compared to a control group [Ferrero et al, 2004]. However, the severity and frequency of these attacks were not different between the 2 groups. A different study by Tietjen et al in 2007 looked at women with migraine headache as the study group and documented its correlation to endometriosis as well as other medical problems. They found higher prevalence of endometriosis in women with migraine than in non headache controls. Furthermore they found that women with migraine and endometriosis had more frequent and disabling headaches than women who had migraine but no endometriosis. Additionally, patients with migraine and endometriosis had more menorrhagia, dysmenorrhoea and infertility compared to other patients with migraine only and normal controls. The same was also valid for depression, anxiety, irritable bowel syndrome, interstitial cystitis and chronic fatigue syndrome. 

·   Other symptoms include bloating sensation, recurrent vaginal candida infections, allergies and food sensitivities and history of infectious mononucleosis. This is over and above an increased familial tendency to allergic manifestations [Lamb and Nichols, 1986]. This could qualify endometriosis as a pelvic disease with systemic manifestations. A review by Agic et al in 2006 revealed that patients with endometriosis had increased blood levels of C - reactive protein, TNF-alpha, IL-6, IL-8 as well as CD44+ and CD14+ monocytes with reduced levels of C3+ T lymphocytes and CD20+ B lymphocytes. This might explain some of the general manifestations frequently associated with endometriosis and the reason why it is considered a systemic inflammatory condition.

·    Endometriosis may involve the upper abdominal cavity, and has been seen under the diaphragm. In such cases, it may cause pain in the affected area. Shoulder pain has also been reported due to irritation of the diaphragm with endometriosis. The liver itself may be affected with endometriosis. This is a rare complication reported by Nezhat et al, in 2005. Endometriomas in the ovary and liver have been described in the same patient before [Rovati et al, 1990]. Such lesions were reported in a woman, long time after she had hysterectomy and bilateral salpingo-oophorectomy for pelvic endometriosis [Bohra and Diamond, 2001].

·    Endometriosis can also involve the lungs, but this is not common. In such cases patients may present with chest pain, shortness of breath and haemoptysis, or pneumothorax during menstrual time. The right lung is usually more affected than the left one [Suginami H, 1991]. Catamenial pneumothorax was thought to be due to:

1. Gaps in the diaphragm which allows air to pass into the lungs from the abdominal cavity during menstruation,
2. Pleural lesions which have damaged the lung cortex leading to direct communication between the lung and the pleural space [Breton et al, 1990].

·   Pancreatic involvement with endometriosis has also been reported after ultrasonic diagnosis of a cystic structure in the pancreas suspected to be a post inflammatory pseudocyst. Histological assessment revealed a diagnosis of endometrioma [Verbeke et al, 1996]. 

·  Occasionally, the prolonged course of the disease and its disruptive effect on a patient's life might lead to depression, emotional problems and low self esteem.

The two laparoscopic images above show endometriotic lesions a little lower than the dome of the diaphragm, and at the dome itself respectively. Both patients had extensive pelvic endometriosis, but neither of them had related upper abdominal quadrant symptoms.

Diagnosis of endometriosis

As there are no reliable screening or conclusive diagnostic tests for endometriosis, a high level of clinical suspicion is necessary. This is especially so in women with adult onset pelvic pain, secondary dysmenorrhoea and deep dyspareunia especially if they had family history of endometriosis or other autoimmune problems. However, a recent report described some potential for serum antiendometrial antibodies as a screening test for endometriosis in patients with dysmenorrhoea, chronic pelvic pain and infertility. Randal et al (2007) reported a positive predictive value of 88%, negative predictive value of 86%, sensitivity of 87% and specificity of 87%.  At the same time Florio et al (2007) used urocortin, which is a neuropeptide from the corticotrophin-releasing hormone group, to help with the diagnosis of endometriomas. Elevated plasma urocortin levels were detected in 88% of the cases with endometrioma with 90% specificity. The corresponding figures reported after using CA 125 in the same group of the patients were 65% and 90%.

As mentioned previously, a structured protocol should be used to help with the clinical diagnosis of endometriosis including: 

• The pain questionnaire should address all systems reviews especially psychological issues and previous history of abuse. Other related concerns should be included as well. It is always useful for patients to fill the questionnaire at home without the intrusion of any other person. They should not exaggerate or belittle any of their symptoms or concerns. It is also important that they air their expectations and what level of pain they might be able to live with.

• A well-kept pain calendar is most valuable in revealing the timing, intensity and duration of pain. It can also reveal bowl, urinary or musculoskeletal symptoms and their relation to menstruation. This gives a clearer picture of all the systems involved and the level of disruption in the social or professional activities of the patient at different times of the month, especially in relation to menstruation. Describing such association from memory may not be very accurate. 

• Pain mapping using pictures would show the site of maximum pain and tenderness and any referral pattern as well as extra pelvic sites.

These diagnostic aids do not cancel the need for a thorough clinical interview as mentioned before. This should be followed by pelvic examination during menstruation which is the best time to examine for nodularity and tenderness of the uterosacral ligaments or the rectovaginal septum. Both vaginal and rectal examinations would be necessary. Pelvic tenderness simulating the patient’s symptoms and pain during defecation or intercourse is a very helpful diagnostic finding and usually indicates local pelvic pathology. With a tender nodular pelvis and negative or positive chlamydia antibodies tests, endometriosis could be seen in 80% and 65% of the patients respectively. Furthermore, lateral axis deviation of the cervix with asymmetry in the size of the lateral fornices and narrowing of the posterior vaginal fornix are occasionally seen during speculum examination. This might follow unilateral scarring pulling the uterus to the same side or an adnexal mass pushing the uterus to the opposite side.

Transvaginal scan examination  

Transvaginal ultrasound scanning may reveal no pelvic abnormality but can show signs of pelvic adhesions in few cases. Furthermore, the probe can be used to elicit site-specific tenderness during examination conducted during menstruation. In 15-20% of the cases an endometrioma can be seen as a cystic mass with thick wall, homogeneous low-level internal echoes and occasional wall calcifications. This pattern is not specific and may be seen in other adnexal masses including dermoid and hemorrhagic cysts, tubo-ovarian abscess and ectopic pregnancies. Furthermore, anechoic cysts or cysts with some internal debris or septae and solid-looking appearances have been described in histologically proven cases of endometriosis. In this respect breakdown of the blood components within an endometriotic cyst may follow the same resolution pattern similar to any other haematoma. Accordingly, the final picture depends on the age of the patient and the endometrioma itself. The picture in postmenopausal women was described as one of heterogeneous internal echoes with central calcification [Asch and Levin, 2007]. Accordingly, there is no appearance exclusively specific to an endometrioma and the clinical picture should help in the final interpretation of the result. This confirms the importance of pelvic scanning being conducted by gynaecologists who can use the scan probe to extend their clinical judgement for the benefit of that particular case. It can also be used to elicit site-specific tenderness, and to test for pelvic adhesions. Once an endometrioma has been diagnosed ultrasonically, other parts of the pelvis are mostly affected by endometriosis in more than 90% of the cases. On the other hand, such a diagnosis can not be excluded if an endometrioma is not seen, as such ovarian cysts are usually seen only in 15-20% of patients with proven pelvic endometriosis.

• The first image below shows an endometrial cyst with low internal echogenicity and a small crescent like ovarian tissue extending between 6 and 10 o'clock. This is the more usual pattern seen with endometriomas, but by no means the only one.
• The second ultrasound image below shows an ovarian cyst with fluid level. This pattern is more common with dermoid cysts. In this case the cyst proved histologically to be endometriotic in nature.

• The first image shows two cysts of different echogenicity. Initially the cyst on the left was described as endometriotic, while the one in the right ovary had all the characteristics of a simple cyst. The left side cyst was confirmed histologically to endometriotic. Interesting, the right side cyst was reported as a serous cystadenoma with superficial layer of endometriosis .

• The second colour image below shows three different histologically confirmed endometriomas in the same ovary. Note the difference between the dark looking one and the other two. Colour was used in this case to emphasise the different patterns of endometriomas even in the same ovary.

Surgical diagnosis difficulty

It has often been stated that laparoscopy with tissue biopsy are the gold standards to diagnose endometriosis. However, certain criteria have to be adopted for patient’s selection to prevent misuse of this diagnostic procedure. Following a thorough clinical assessment, the following indications could be used:

• There is a strong clinical suspicion after using a structured protocol. 
• When the symptoms affect the quality of life. 
• Examination during menstruation elicited pelvic nodularity or site-specific tenderness similar to the patient's complaint.
• When ultrasound scan examinations revealed a pelvic abnormality.
• When an initial treatment with oral contraceptives and NSAID has failed.
• When new symptoms occur in a patient already diagnosed with endometriosis.
• Planned second look for completion of treatment.
• Infertility investigations especially in patients with chronic pelvic pain or those who had pelvic surgery in the past.

Pelvic endometriosis may show as atypical clear, white, vesicular, flame-like, yellow and red lesions. These patterns are usually seen in adolescents and in patients with recent onset of the disease and may spontaneously regress or grow according to the ovarian hormonal cycle [Evers, 1987]. Accordingly, extra care should be taken to avoid missing them. Occasionally, peritoneal defects are seen on the pelvic sidewall. These are known as Allen-Master windows and can be related to endometriosis. They were first described by Sampson in 1927 and since then they have been described by different authors. They were found in 7% of patient with chronic pelvic pain [Chatman, 1981]. Histological assessment of these pockets was done by Vilos and Vilos in 1995. Endometriosis was found in 39%, chronic inflammation in 20%, endosalpingosis in 12%, calcification in 4% and no abnormality in 25% of the cases. The typical black or brown lesions historically associated with endometriosis usually take many years to develop and are usually seen in adults many years later than the atypical ones. They represent haem degradation products as they undergo haemorrhage and fibrosis. On the other hand, dormant or healed lesions may look white or as calcified areas, made of remnants of glands embedded in fibrous tissue [Nisolle et al, 1995].

The three above shown laparoscopic images demonstrated different looks of Allen Master Windows. Nevertheless, the histology reports in all three cases showed endometriotic tissues with some chronic inflammatory cells. Some reports showed that similar lesions may show only scar tissues, probably due to burnt out endometriosis. Alternatively, the sampling might not have be deep enough to reach the deep lesions.

Historically, visual identification in 33% of all supposed endometriotic lesions did not satisfy strict histopathological diagnostic criteria [Fielding, 2003]. Accordingly, the advise was to excise, or at least biopsy all suspected lesions for histological confirmation of the disease before destructive surgery. This is especially so, as old suture material, carbon deposits following previous surgery, old blood and such conditions as endosalpingosis may all give the same visual appearance. This is especially so for endosalpingosis which is made of ciliated tubal tissue without any stroma. It was found in 7.9% and 7.3% of women with or without pelvic pain respectively [Hessling and De Wilde, 2000]. Accordingly, it is considered to have only a minor role in this respect. Over diagnosis of endometriosis is as bad as under diagnosing it, as patients will be liable to have unnecessary medical or surgical treatment in the future for the alleged disease. Furthermore, no more efforts will be made to find the real cause of the patient’s symptoms.

On the other hand, histological diagnosis of endometriosis may be fraught with many difficulties related to the presence or absence of the glandular component. This is demonstrated by ‘stromal or gland-poor’ endometriotic lesions. Furthermore, infiltration of the stroma by histiocytes and smooth muscles as well as peritoneal fibrosis, metaplasia and decidual changes may interfere with the histological picture as reported by Clement PB in 2007. The same author also reported some histological diagnostic difficulties when dealing with cervical, ovarian surface, tubal, intestinal and colonic endometriosis. This information can explain the negative histopathological reports occasionally received from the laboratory after examination of biopsies removed from areas with characteristic endometriotic appearances, and patient's who had favourable response to down regulation with GnRH-analogues. In recent years, both the Royal College of Obstetricians and Gynaecologists and ESHRE recommended that diagnosis of endometriosis should rely on visual inspection by an experienced gynaecologist.

Anterior abdominal wall endometriomas

Occasionally patients may present with cyclical anterior abdominal wall pain during menstrual time. This is mostly felt along an abdominal scar following caesarean section or myomectomy. In such cases a tender mass will be felt and Carnett’s sign may be positive. High frequency ultrasound examination of the anterior abdominal wall may reveal a mass with different types of echoes. As with similar ovarian lesions, there is no specific characteristic picture for abdominal wall endometriomas. They may be echogenic, isoechoic or hypoechoic in comparison to the neighbouring tissues. However, the cyclical nature of the pain relative to menstruation with a previous history of pelvic surgery should raise the suspicion for making the diagnosis.

The differential diagnosis of such masses includes hernias, spontaneous haematomas, lipomas and desmoid tumours. However, the clinical history and ultrasound examination findings will be different for these pathologies. Anterior abdominal wall pain caused by nerve entrapment related to previous surgery will be discussed later on under the musculoskeletal causes of pelvic pain.

• The above first ultrasound image shows a circumscribed hypoechoic anterior abdominal wall mass, underneath a caesarean section scar, splitting the rectus sheath. It was immobile and very tender on palpation, especially during menstruation. The patient had 2 previous caesarean section operations, the second being 2 years previously. It proved to be an endometrioma on histological assessment. The uterus also showed an area of adenomyosis.

• The second ultrasound image above shows a well circumscribed solid mass in the rectus sheath. The patient also had 2 previous caesarean sections, the second one being 6 years previously. This mass also proved to be endometriotic in nature.

Perineal and vulvar endometriosis

Endometriosis could also be seen in episiotomy scars though this is not commonly reported. This possibility should be kept in mind especially in patients who complain repeatedly of a painful perineum after childbirth especially if an episiotomy had been performed. McCormick et al (2007) reported on 3 patients who presented with vague perineal pain adjacent to the episiotomy scars, 3-6 months after childbirth. Their symptoms got worse cyclically during menstruation and physical examination revealed vague fullness of the area. Endoanal ultrasound scan examination revealed a local mass with mixed echogenicity which proved endometriotic on histological assessment after excision. Though uncommon, they suggested that such a diagnosis should be kept in mind when dealing with patients who present with a painful perineum after childbirth. Similar case reports have been published of endometriosis involving a bartholin gland or the resulting scar following its excision by Matseoane et al in 1987; Goemen et al in 2004 and Buda et al in 2007.

Treatment of endometriosis

Endometriosis is an inflammatory progressive disease which can cause pain by different ways, depending on the stage and severity of the disease. Biochemically active lesions can produce prostaglandins, cytokines and substance P which can cause dysmenorrhoea and bowel symptoms. They may, as well, trigger afferent nerve pathways. Deep lesions can affect nerve endings directly causing more pain [Redwine and Sharp, 1971]. On the other hand, pelvic adhesions may affect tissues blood supply, and cause mechanical tension on small nerves [Olive et al, 1997]

The exact treatment of endometriosis depends on many variables including the patient’s age, her symptoms, fertility needs as well as the location and extent of the disease itself. Accordingly, the treatment plan should be tailored to suit the patient’s needs including:

• Pain relief,
• Restoration of normal or near normal anatomy,
• Maintenance or improving the fertility potential.

It is apparent that only surgical management may have a chance of fulfilling all these three requirements. But even so, this is not the case in many patients. On the other hand, medical treatment could be used to prevent further progress of the disease and to give pain relief in mild and moderate conditions. However, eventually surgery would be necessary to deal with such conditions as large endometriomas and deep endometriosis in the rectovaginal septum. 

Medical treatment of endometriosis

Pain control is the cornerstone for the total management of patients with endometriosis. In most cases patients would have used over-the-counter medication for a long time before seeking medical advice. Non-steroidal anti-inflammatory drugs (NSAID) are usually prescribed first for that purpose in primary care clinics. They act by blocking the synthesis of prostaglandins which are important factors in causing pain. Different brands within this group are used including mefenamic acid, ibuprofen and naproxen. Some patients might respond or react differently to these drugs and a different one should be used in cases of side effects or ineffectiveness. They could be used to control dysmenorrhoea or in a time contingent plan for chronic non-cyclic pelvic pain. However, they might prove ineffective in many cases and more potent non-narcotic analgesia would be needed.

Hormonal treatment in the form of the oral contraceptive pill, progestogens including depo-provera, gonadotrophins releasing hormone analogues and danazol have all been used with good results but have different side effects which might not be acceptable to the patients. They are more effective when dealing with small endometriotic implants. They could reduce the size of the endometriotic lesion and the related inflammation as well as pelvic vascularity [Betts and Buttram Jr VC, 1980 and Cook AS and Rock JA, 1991]. They are all capable of reducing the actual or relative oestrogen supply to the endometriotic implants but each of them has a unique specific mode of action. Progestogens have the added effect of causing decidual exhaustion of the endometrial glands. On the other hand, danazol has a local androgenic effect on the lesions. The combined oral contraceptive pill is progestogen dominant and causes atrophy of the endometrium in the long term. This would reduce the monthly menstrual blood loss with less retrograde menstruation which could be useful after excisional surgery to remove endometriosis. They also block the enzymes cyclo-oxygenase type II (Cox-2) and aromatase which are necessary for the synthesis of prostaglandins and the conversion of androgens to oestrogen respectively [Maia and Casoy, 2007]. Accordingly, they could give symptomatic relief as well as reduce the progress rate of the disease. On the other hand, continuous use of a monophasic oral contraceptive pill proved to be safe and useful for patients who get pain with cyclic withdrawal bleeding. However, GnRH analogues are the ones mostly used now as they proved to be the most effective especially with dysmenorrhoea and dyspareunia [Vercellini et al, 1993]. They are given by monthly injections though certain ones could have an ultra long effect and could be given every 3 months. Nasal spray medication has been used in the past and is still available for those who prefer to avoid injections but are not practical alternatives, as two or more daily doses are usually needed. As well, efficacy could be affected by nasal congestion and catarrh. Furthermore, a patient might forget to use it or might not find the right circumstances in work or during travel to administer it unnoticed. They work by desensitising the pituitary gland reducing gonadotrophins secretion, mainly LH, leading to anovulation which would deprive endometriosis of its oestrogen supply. To prevent osteoporosis and severe vasomotor symptoms, patients should be offered an add-on medication to allow their prolonged use for 6 months or more. This is usually done in the form of low or high doses of progestogens, low dose oestrogen or combined oestrogen and progestogen medication. In essence these drugs are given as hormone replacement therapy to counteract the temporary and false menopausal state created by the GnRH analogues. Good pain control is usually possible during the GnRH-a medication period and could last for 6 months after cessation of the treatment but recurrence is almost certain thereafter. On the negative side, patients would not be able to conceive while using these drugs and hence they are not helpful in improving the fertility potential of the inflicted patient. Furthermore, as endometriotic lesions usually get smaller and change colour after such medication, they might be missed during laparoscopy compromising the chances of full excision. This is a drawback especially during second look laparoscopy to assess the effect of the first surgery and in patients with recurrent symptoms.

In recent years the levonorgestrel intrauterine contraceptive device mirena has been shown to reduce the symptoms and size of endometriotic lesions in the rectovaginal septum [Fedele et al 2001]. Furthermore, suggestions have been put forward to use it immediately after surgery to reduce the need for further operations in the future [Vercellini et al, 2003] as mentioned previously. Levonorgestrel has been found in the peritoneal fluid in significant amounts, approximately two-thirds of the serum level, in women who showed improvement in their symptoms after six month of using mirena [Lockhat et al, 2005]. Similarly, the level of the serum marker CA-125 showed equivalent decline after long term use of the device and GnRH-a for the treatment of endometriosis [de Sa Rosa e Silva et al, 2006]. All these effects were related to the high levels of peritoneal levonorgestrel causing atrophy of the ectopic endometrial glands, decidual transformation of the stroma as well as increased apoptotic activity. It has also been reported to have anti inflammatory and immunomodulatory effects [Vercellini et al, 2005]. Furthermore, levonorgestrel has been shown to decrease and then block DNA synthesis and mitotic activity by Bergeron C in 2000. The increased endometrial apoptotic activity related to levonorgestrel was found to be secondary to reduced expression of the Bcl-2 gene which has an anti apoptotic effect [Vereide et al, 2005]

Other new treatment modalities included aromatase inhibitors which have been used for the management of pelvic pain caused by endometriosis. Favourable response with reduction in pain intensity and size of the endometriotic lesions has been reported [Shippen and West, 2004]. Using an analogue pain scale, Amsterdam et al [2005] confirmed the same findings after using 1 mg anastrazole with an oral contraceptive pill daily for 6 months. In fact the mean pain scores started dropping after only one month of therapy. More important, they found no detrimental effect in the blood counts, liver and renal function tests, cholesterol level or bone density. Other experimental therapies used included selective oestrogen receptor modulators and anti progesterone medication RU 486 (mifepristone). However, they are all still tentative and more evidence is needed to support their wide scale use in such cases.

One could conclude this section by stating that the ideal medical treatment for endometriosis which could control the progress of the disease and give good symptomatic relief without affecting fertility in young women is not yet available. However, in an animal model the angiogenesis inhibitor TNP-470, endostatin, effectively interfered with the growth and maintenance of endometriosis [Nap et al, 2004]. One year later Becker et al (2005) showed 47% reduction in the growth of endometriotic lesions using the same drug, compared to a control group without affecting fertility. This was further confirmed by Jiang et al in 2007 who showed reduced expression of vascular endothelial growth factor (VEGF) in the ectopic endometrium. Such findings would hopefully pave the way for similar research in humans to limit the use of surgical interventions with their significant potential for adhesions formation.

Surgical treatment of endometriosis

Surgical management of endometriosis could be definitive by total excision of all the lesions or conservative by local destruction with laser or electrocautery. Some of the implants could extend 20 mm into the pelvic tissues especially in the rectovaginal septum and only total excision would allow effective removal of these deep lesions. Conservative surgery with preservation of the uterus, tubes and ovaries would be indicated for the younger age groups who still have fertility aspirations. This might entail incomplete excision and a higher risk of recurrence. Accordingly, patients should be counselled regarding the possibility of residual pain and the need for further surgery in the future. This would reassure them that treatment has not failed but was rather tailored to preserve or even improve their fertility potential. 

During radical or definitive surgery all visible lesions and any suspicious peritoneal areas as well as the appendix should be removed. Such complete excision was found to give pain relief for up to 5 years but pain could recur earlier due to the formation of adhesions or any other cause without the recurrence of endometriosis. This was in fact shown by Abbott et al in 2003 as 36% of their patients needed further treatment but no further endometriosis was seen during second look surgery. Previously total abdominal hysterectomy and removing both ovaries was the standard method for treating severe endometriosis resistant to medical treatment and conservative surgery. This is no longer the case as removing the uterus does not remove the deep endometriosis implants in the pelvic sidewall, uterosacral ligaments or rectovaginal septum. This is expected, as endometriosis is an extrauterine disease in the first place. Nezhat et al [1994] reported that 34% of patients with chronic pelvic pain who had hysterectomy and bilateral salpingo-oophorectomy in the past showed endometriosis during diagnostic laparoscopy. In the same year, Redwine DB [1994] came to the conclusion that endometriosis could remain symptomatic despite removing both ovaries, even without using any exogenous oestrogen, unless all deep endometriotic lesions were excised at the same time. He reported 33% intestinal involvement in such cases. In a more recent publication, Behera et al [2006] confirmed the persistence of endometriosis in many patients with chronic pelvic pain who had a hysterectomy and bilateral oophorectomy in the past. Other than the expected adhesions, endometriosis and ovarian remnants were the most common findings. It remains to be said that the uterus could be removed when the endometriosis is associated with severe adenomyosis or multiple symptomatic fibroids at the same time.  

Management of ovarian endometriomas is a common dilemma especially in young patients. Emptying and flushing the cyst repeatedly and local destruction of the sac could be the preferred option in young women, as less ovarian tissue would be destroyed. This could be done with cysts 5 cm or less in diameter. However, with larger cysts the capsule should be removed if at all possible. This might be difficult at times as it might be stuck to the ovary by scar tissue. In such cases that part of the capsule should be excised with a sacrifice of some healthy tissue underneath. Alternatively, the cyst could be aspirated and flushed repeatedly during the initial laparoscopy and GnRH-a administered monthly for 3 months. Definitive surgery to remove the capsule could be arranged after that time when less tissue inflammation is expected and stripping the capsule might become easier. Before attempting any form of ovarian surgery the ovaries should be mobilised free of any adhesions to allow proper exposure of the pathology and to select the appropriate area to start the cystectomy.

As mentioned before, in the presence of ovarian endometriosis other parts of the pelvis may be involved in the majority of cases. In a large series of cases, only 1.06% of the patients who had ovarian endometriosis did not show any further lesions in other parts of the pelvis [Redwine DB, 1999]. The same author suggested that such lesions should be used as markers for extensive disease. Treating only the ovarian pathology would indicate underdiagnosis and undertreatment of the concerned patients.

The first image above shows enlargement of the right ovary with an endometrioma. The second picture shows the ovarian capsule being sliced to remove the cyst. Unfortunately it started leaking (figure 3), and its cotents needed to be sucked out, before the cyst capsule could be removed (Figures 4 and 5). The sides of the ovarian defect were not approximated (Figure 6), as ovaries usually heal best without adhesions, if not resutured [Brumsted et al, 1990; Wiskind et al, 1990 and Myer et al, 1991].

Deep endometriosis

This is a different entity of endometriosis and needs to be addressed separately. It is a term used to describe lesions infiltrating 5 mm or deeper under the peritoneum [Koninckx and Martin, 1994]. It is commonly seen in the uterosacral ligaments, the rectovaginal septum and beneath the bladder peritoneum. The pathology was described by Novak in 1974 as a combination of endometrial glands and stroma within a mass of fibromuscular tissue. However, Cullen described this concept of adenomyomas in 1927 to be followed by Sampson who suggested in 1928 that rectovaginal lesions were only secondary to infiltration of that part by endometriosis from neighbouring areas. Such lesions could involve the uterosacral ligaments first before spreading laterally or medially to engage the ipsilateral ureter or rectum respectively. As well they could begin in the pouch of Douglas itself before extending in different directions to involve all neighbouring tissues. The outer serosa and superficial longitudinal muscular layers are usually the most commonly affected parts of the rectal wall.

Deep endometriosis is more likely to cause chronic pelvic pain than endometriosis in other parts of the pelvis. Compression and infiltration of the nerves in the subperitoneal space by deep endometriotic lesions was suggested to be the direct cause [Fauconnier and Chapron, 2005]. Involvement of the rectum and bladder could cause symptoms related to these organs. Such location-indicating symptoms and site-specific tenderness could give good clues to the location and distribution of the endometriotic lesions. Furthermore, a positive correlation was found between the severity of dysmenorrhoea with the presence of deep endometriosis in the uterosacral ligaments and with the depth of their invasion [Leng et al, 2007]. Such lesions as well as nodules in the rectovaginal septum were also associated with dyschesia more than lesions in other parts of the pelvis.

The rectum might fuse with the posterior vaginal wall, the back of the cervix or even the uterus itself. Such lesions are also more likely to cause pelvic pain than endometriosis in other areas of the pelvis. This is especially so during intercourse and defecation with the pain mostly felt in the back. Shooting pain up the rectum is not uncommon. Sometimes pain is felt only before or during menstruation. Rectal bleeding is a very unusual presentation of deep endometriosis. It is best diagnosed by combined vaginal and rectal examinations to feel for any indurations in the rectovaginal septum, the pouch of Douglas or uterosacral ligaments. Slight upward pushing of the cervix with the finger in the vagina would stretch the uterosacral ligaments and make them easier to feel with the finger in the rectum. It is more useful to conduct this examination during menstruation otherwise there is 50% risk of getting false negative results. The importance of such timing has been stressed by Koninckx et al in 1996.

Unfortunately transvaginal scan examination and barium enema might give false negative results depending on the site and extent of the disease. The role of MRI has already been discussed before. In comparison to rectal scanning it proved to be more accurate for the diagnosis of uterosacral and vaginal endometriosis where as both were equally accurate for the diagnosis of colorectal endometriosis [Bazot et al, 2007]. However, neither MRI nor rectal scanning had absolute diagnostic accuracy, but they are still very useful adjuncts to vaginal and rectal examinations. As well, there is no biochemical test which is specific or sensitive enough to help with the diagnosis of deep endometriosis; hence a high degree of clinical suspicion is needed. Occasionally, a nodule could be felt on one side or the other while the patient is under general anaesthesia before starting laparoscopy. This might not be possible in the clinic because of the immense pelvic tenderness. However, nodules in the upper part of the uterosacral ligaments and rectal wall might not be felt during vaginal or rectal examinations [Chapron et al, 2002]. As the lesion is retroperitoneal, its true extent might not be possible to assess laparoscopically and complete excision should be ascertained by vaginal and rectal examinations during the procedure.

• The first ultrasound image shown below demonstrates extensive rectovaginal endometriosis as seen during transvaginal ultrasound examination, marked as X. 

• The second image shows an endometriotic nodule medial to the right uterosacral ligament (RUSL). The bulk of the nodule was felt vaginally. To deal with such a lesion, the technique should isolate the uterosacral ligament from both the ipsilateral ureter and the rectum, before it can be excised. The advice 'Do not attack the pathology' is most important in such cases, as both the ureter and rectum may be damaged.

A normal pouch of Douglas usually extends down behind the upper one third of the vagina. Accordingly, a good part of the posterior vaginal wall would be seen between the rectum, the back of the cervix and the two uterosacral ligaments. For that reason, a swab on a stick pushed up the posterior fornix would show how much of the vagina is visible laparoscopically. This part of the vagina is also covered by an upward extension of the rectovaginal septal fascia which goes on to form part of the posterior pericervical ring and extends laterally to fuse with the uterosacral and cardinal ligaments [Richardson et al, 1993]. Rectal adhesions to the posterior vaginal wall or the cervix would cause partial or complete obliteration of the vaginal wall respectively. With partial obliteration part of the vaginal wall would be seen where as with complete obliteration the outline of the sponge on the stick in the posterior vaginal fornix would not be seen laparoscopically. This test is more accurate than simple visual assessment of the pelvis in verifying the extent of obliteration of the pouch of Douglas.  Furthermore, using a probe into the rectum could help to identify the extent of rectal wall involvement. The outer serosa would move freely on top of the probe in cases with no or only superficial involvement. Alternatively, with deeper lesions involving the inner muscular layer and mucosa itself, the serosa would not move as freely on top of the probe. 

Deep rectovaginal endometriosis is usually divided into 4 stages with no distortion or obliteration of the pouch of Douglas in the first two stages. However, a correct diagnosis and staging may become possible only during the microdissection of the nodule during surgery. The 4 stage are:

• Stage 1 with no vaginal involvement.
• Stage 2 involves the vagina.
• Stage 3 involves the vagina and rectum with cul-de-sac distortion.
• Stage 4 as for stage 3 plus cul-de-sac obliteration.

Early stage of rectovaginal endometriosis with the rectum shown tented up. Still most of the upper part of the vagina could be seen underneath the posterior part of the cervix between the two uterosacral ligaments
Severe pelvic adhesions do not necessarily indicate involvement of the rectum or vaginal wall with deep endometriosis. Accordingly, total reliance on visual assessment may give a wrong diagnosis in many cases. The neighbouring laparoscopic picture is a good example of such a case, with no rectovaginal involvement at all. This patient attended for infertility reasons, rather than pelvic pain. It was reported by Redwine and Wright in 2001 that   27% of their patients who had complete obliteration of the pelvis with endometriosis did not have rectal involvement.
For comparison purposes with the picture above,   the neighbouring image shows the rectum pulled up and attached to the posterior vaginal wall and posterior aspect of the cervix. A nodule was felt high up in the posterior fornix behind the cervix by vaginal examination. Despite infliction with deep endometriosis the pelvis looked almost normal. Furthermore, this patient had intense deep dyspareunia and dyskesia.

Treatment of deep endometriosis

Treatment of deep endometriosis could be difficult and entails thorough planning in advance. It is important to remember that deep endometriosis does not mean involvement of the rectovaginal septum in all cases. In fact such involvement occurs in only a small percentage of all deep endometriotic lesions [Chapron et al, 2001]. The majority of lesions would be in the uterosacral ligaments. Accordingly over zealous attempts to reach that area would be unnecessary in most cases.

Examination under anaesthesia is usually more informative than laparoscopy in localising the exact site and size of the lesion, as it is retroperitoneal. However, laparoscopy with its magnification and bright light gives easy access to the rectovaginal area in comparison to open surgery. Furthermore, the exact extent of the disease might not be revealed completely till microdissection of the involved area has been started. Excellent exposure could be possible by proper positioning of the patient as described before. Using a Zumi or a similar intrauterine manipulator would also allow adequate movements of the uterus in all directions. Excision of a nodule from the anterior rectal wall, posterior vaginal wall and cervix as well as the involved uterosacral ligaments is the only reliable way to remove the whole lesion. Fulguration and desiccation would not be adequate as the lesions usually extend for 10-20 mm into the retroperitoneal tissues. Using the tripolar forceps (bipolar and cutting) would reduce tissue damage, blood loss and operative time as one instrument is used to coagulate and cut at the same time. 

Involvement of the rectal wall is usually localised to the covering serosa and outer longitudinal muscle layer. In such cases treatment would entail partial thickness resection which involves peeling such lesions off the rectal wall. It is rare for deep endometriosis to penetrate the whole rectal wall necessitating full thickness disc resection of the anterior rectal wall and laparoscopic resuturing. However, with large or multifocal lesions segmental resection of the whole area with end-to-end reanastomosis would be necessary. This is the domain of the Colorectal Surgeons and might be the best option in such cases to prevent the risk of incomplete resection if either of the other two techniques was used [Daraï et al, 2007].

The following protocol is useful to follow during surgery:

• The bowel should be prepared with picolax the day before admission and the patient should take only fluid diet on that day.

• Prophylactic intravenous antibiotics should be given on the table after taking note of any relevant drug allergies.

• Vaginal and rectal examinations should be done under general anaesthesia for the exact location of the nodules, bearing in mind that high rectal lesions may not be reachable and can only be reached during the micro-dissection of the involved area.

• Assess the exact extent of the rectal and vaginal walls involvement by using a rectal probe and a swab on a stick inserted into the posterior vaginal fornix as described before.

• Do not attack the pathology and always work from a normal area into the abnormal one.

• Adhesiolysis, mobilisation of the ovaries and ovarian cystectomy should be performed if needed to improve the operative field.

• Transparietal fixation of one or both ovaries to the anterior abdominal wall should be done to allow good exposure of the operative field and to free all ports for operative purposes. Care should be taken not to injure the inferior epigastric vessels during this procedure as well. The sigmoid colon could be similarly fixed to the anterior abdominal wall if it obstructed access to the operative field.

• The ureters should be identified as high as possible near the pelvic brim and ureterolysis should be done taking care not to devitalise them. This might need mobilisation of the sigmoid colon to expose the left ureter as it enters into the pelvis.

• Dissect the paracolic spaces between the rectum and the uterosacral ligaments till healthy fatty or loose areolar tissue is reached. Make use of the rectal probe to move it away and reduce the risk of rectal damage.
• Excise the uterosacral ligaments down to their cervical attachment taking care not to damage the neighbouring venous blood channels.

• Dissect the rectum off the back of the cervix and upper vagina making use of the rectal probe to move the rectum from side to side to help in finding the best line of cleavage.

• The endometriotic nodule should be excised off the posterior vaginal wall making use of the swab on the stick in the posterior fornix.

• The nodule and scar tissue should be peeled off the rectal wall making use of the rectal probe to find the exact line of cleavage.

• Check haemostasis under water after filling the pelvis with Ringer’s lactate.

• Check rectal wall integrity under water by blowing air into the rectum.

The technique can be changed as necessary to suit different cases and situations. This is especially so as it may be easier to dissect the nodule off the rectum first in one case and off the vagina at other times. In certain cases when the lesions are mainly in the vaginal wall, a vaginal approach will be necessary to dissect the rectum free from the vagina before complete excision of the vaginal lesions. This can be done as a laparoscopically assisted procedure if necessary.

The 6 laparoscopic photographs shown below demonstrate few of the steps during excision of a rectovaginal endometriotic nodule. The patient presented with severe progressive dysmenorrhoea, bad deep dyspareunia and dyskesia. She had been symptomatic for the previous 8 years and did not respond to medial treatment. Laparoscopy showed no ovarian, tubal or uterine adhesions.

• The first picture above shows endometriosis of the cul de sac. It was felt vaginally as a thickened area across the pouch of Douglas but mainly on the left side. Laparoscopy did not reflect the extent of the disease. However, pushing up the posterior vaginal fornix with a swab on a stick did not demonstrate any part of the vagina anterior to the rectum, which was stuck to the back of the cervix. This in turn could not be seen either. 

• The second picture shows opening the right pelvic sidewall at a high level in a healthy area to mobilise the right ureter, as a first stage during surgery. The right ureter can be seen behind the peritoneum on the lateral side of the incision. The release incision is made medial to the ureter to allow the parietal peritoneum to pull the ureter laterally as seen. Anatomically, the ureter and the parietal peritoneum make the first layer of the pelvic side wall. The second layer is made of the anterior branch of the internal iliac vessels, lymph nodes and deep fascia. The deep layer is made of the obturator muscle, obturator vessels and verve and external iliac vessels. Staying within the right layer usually allows proper dissection of the involved structures without damaging structures within the other layers.

• The third picture shows the rectum pushed up with a rectal probe. A cluster of endometriosis is clearly seen as  dark spots on its anterior wall.

• The fourth picture shows the free rectum and free vagina with the dissected nodule pulled sideways before being excised at its lower end near the vagina with the tripolar forceps.

• The fifth picture demonstrates the free vagina pushed up with a swab on a stick in the posterior fornix

• The last picture shows a healthy rectum and pelvic sidewalls at the end of the procedure. 

Unlike superficial endometriosis, deep rectovaginal lesions are usually less progressive and may stay asymptomatic for a long period of time. A study published by Fedele et al in 2004 showed only 9.7% of asymptomatic patients so affected became symptomatic or had an increase in the size of their lesions within 6 years of follow up. Accordingly, a coincidental diagnosis of deep rectovaginal endometriosis in an asymptomatic woman does not indicate surgical treatment bearing in mind the difficult surgery necessary to deal with these lesions. The risk of complications after surgical treatment of rectovaginal septum endometriosis has recently been published by Slack et al in 2007. Out of 128 patients who had such surgery, 32% needed intraoperative closure of a rectal defect, but only 3 of them were done for rectosigmoid resection. Furthermore, 3 patients developed rectovaginal fistulae and a further one had ureterovaginal fistula. They also reported ureteric damage in two women. This reflects the nature of the pathology itself and the difficult surgery necessary to deal with it. Patients should be counselled accordingly to prevent recriminations.

Prediction of pelvic pain in women with endometriosis

Before proceeding to another topic, it is important to re-examine the correlation between endometriosis in general and chronic pelvic pain, and to find what can be used as good predictors of pelvic pain. This is especially so since almost 40% of patients with endometriosis may not have any symptoms. This can also help in explaining the already known fact that pelvic pain is not related to the severity of endometriosis, as classified by the American Society for Reproductive Medicine. Using logistic regression analysis Porpora et al [1999] found deep endometriosis, adnexal adhesions and ovarian endometriomas were independent predictors of pelvic pain. Both chronic pelvic pain and dyspareunia can be predicted by the presence of deep endometriosis, and ovarian endometriomas associated with adnexal adhesions. It was not the size of the endometriomas, but their association with adhesions that caused pelvic pain.

Laparoscopic surgical nerve ablation

Surgical nerve ablation has been used for a long time in the treatment of chronic pelvic pain. The three techniques used are laparoscopic uterosacral nerve ablation [LUNA] and excision [LUNE] and laparoscopic presacral neurectomy [LPSN]. All these techniques are useful to different extent in treating midline lower abdominal pain mainly dysmenorrhoea. However, lateral pelvic pain is not affected as the corresponding nerves travel within the infundibuloplevic ligaments.

A. Laparoscopic uterosacral nerve ablation

LUNA partly denervates the uterus and has up to 80% success rate in the short term, but pain usually recurs within one year or so. It is a simple procedure to perform. The uterosacral ligaments should be transected near their attachment to the posterior part of the cervix. However, it is not the appropriate procedure to perform when the ligament is heavily involved with bowel adhesions or endometriosis. In these cases, proper dissection of these adhesions and excision of the ligament will be the more effective technique [LUNE]. This will have a double benefit of reducing the endometriotic load, as well as denervating the uterus. Combining LUNA with surgical treatment of endometriosis in the same patient gave long-term success rate in 62% of the cases, compared to non-surgical controls in a controlled study conducted by Sutton et al in 1994.

To prevent ureteric injury during LUNA the uterus should be acutely anteverted and pushed up the pelvis, to one side with the uterine manipulator. This will draw out the contra lateral ligament, making it easier to transect. Furthermore, a releasing incision parallel and lateral to the uterosacral ligament should be made, taking care not to injure the venous blood channels lateral to the ligament. This will allow the ureter to move out of the operative field within its peritoneal cover. The tripolar forceps is best suited for this job as it can be used for dissection, pulling the ligament medially before being cauterised and transected using one piece of equipment. LUNE is slightly different as it entails excision of the ligament rather than just transecting it. Furthermore, it provides a specimen for histological assessment. This is important even in women with pelvic pain and normal findings during laparoscopy. It has already been reported that 73% of removed specimens from women with such a combination showed histological evidence of endometriosis, endosalpingosis or inflammation as mentioned before. This emphasises the recommendation of the American College of Obstetrician and Gynaecologists Practice Bulletin which suggested using GnRH-a for treating patients with chronic pelvic pain and negative findings during laparoscopy, with no other detectable nongynaecological abnormality. It is also reassuring that second look laparoscopy did not show significant adhesions formation following this procedure when it was performed as part of a more extensive surgery to remove all deep pelvic endometriosis. [Chapron et al, 2001].

B. Laparoscopic presacral neurectomy [LPSN]

LPSN is more effective than LUNA in relieving central lower abdominal pain and dysmenorrhoea for longer periods of time, when the procedure is properly performed [Tjaden et al, 1990]. However, Candiani et al in 1992 reported no significant difference in frequency and severity of dysmenorrhoea, pelvic pain or dyspareunia after long term follow-up in two groups of patients who had conservative surgery for endometriosis with or without presacral neurectomy. This emphasises the need for complete excision of all classical and non classical endometriotic lesions to improve the therapeutic results and not to rely only on such nerve ablation techniques.

LPSN is a more elaborate procedure than LUNA and care should be taken to avoid vascular injury and retroperitoneal haematoma formation. Furthermore, the ureters, sigmoid mesentery and common iliac veins are all at risk during this procedure. To improve operative field exposure, the patient should be put in a steep Trendelenburg position with a left lateral tilt to move the intestines and sigmoid colon upwards and laterally. The peritoneum below the bifurcation of the common iliac vessels should be pulled up with dissecting forceps to separate it from the underlying vessels and incised transversely. This will expose the nerve plexus centrally or slightly to the left side. However, no further dissection should be attempted to expose the left ureter, as that is be totally unnecessary. Following gentle dissection of the retroperitoneal tissues, the presacral nerve plexus should be transected in consecutive layers using the tripolar forceps. The peritoneal incision does not need to be reapproximated at the end of the procedure, and the excised tissue should be sent for histological examination to confirm its origin.

Some urinary and bowel dysfunction may follow excision of the presacral nerves. Minor changes may also occur in the bladder sensation due to interruption of its sympathetic nerve supply. However, as the parasympathetic nerve fibres in the hypogastric plexus are not affected by this procedure, bowel symptoms are not common though constipation has been reported.