ESHRE Logo ESHRE Guideline for the Diagnosis and Treatment of Endometriosis

Treatment of pain
(supporting documentation)

Rationale for treatment

In most women with endometriosis, preservation of reproductive function is desirable. The least invasive and least expensive approach that is effective with the least risks in the long run should be chosen. It is important to involve the woman in the decision taken about the treatment options. Symptomatic endometriosis patients can be treated by analgesics, hormones, surgery, assisted reproduction or a combination of these. Many women with endometriosis have pain and subfertility at the same time, which complicates the choice of treatment. Unfortunately, as endometriosis is a chronic disease, elimination of the endometriotic implants by surgical or medical treatment often provides only temporary relief. Therefore, the goal should be to eliminate the endometriotic lesions and, more importantly, to treat the symptoms (pain and subfertility) and prevent recurrence. It has to be kept in mind that endometriosis is a chronic disease and the recurrence rate is high after both hormonal and surgical treatment.

General considerations on medical treatment

Because oestrogen is known to stimulate the growth of endometriosis, hormonal therapy has been designed to suppress oestrogen synthesis, thereby inducing atrophy of ectopic endometrial implants or interrupting the cycle of stimulation and bleeding. Implants of endometriosis react to gonadal steroid hormones in a manner similar but not identical to normally stimulated eutopic endometrium.


Progestins exert an antiproliferative effect by causing initial decidualisation of endometrial tissue followed by atrophy. They can be considered as a first choice for the treatment of endometriosis because they are as effective in reducing AFS scores and pain as danazol or GnRH analogues and have a lower cost and a lower incidence of side effects than danazol or GnRH analogues (Vercellini et al., 1997). Their use in endometriosis has been the subject of a Cochrane Review (Moore et al., 1997; Prentice et al., 2000). There is no evidence that any single agent or any particular dose is preferable to another. In most studies, the effect of treatment has been evaluated after 3 to 6 months of therapy. Medroxyprogesterone acetate (MPA) has been the most studied agent and is effective in relieving pain starting at a dose of 30 mg/day and increasing the dose based on the clinical response and bleeding patterns (Moghissi and Boyce, 1976; Luciano et al., 1988). Pain was reduced significantly during luteal phase treatment with 60 mg dydrogesterone and this improvement still was evident at 12-month follow-up (Overton et al., 1994). Other progestagens, such as desogestrel, are now being looked at as alternative treatments (Razzi et al., 2006).

Side effects of progestagens include nausea, weight gain, fluid retention, and breakthrough bleeding due to hypo-oestrogenemia. Breakthrough bleeding, although common, is usually corrected by short-term (7-day) administration of oestrogen. Depression and other mood disorders are a significant problem in approximately 1% of women taking these medications. Reports from studies of local progesterone treatment of endometriosis-associated dysmenorrhea with a levonorgestrel-releasing intrauterine system during 12 months resulted in a significant reduction in dysmenorrhea, pelvic pain and dyspareunia, a high degree of patient satisfaction (Vercellini et al., 1999b; Vercellini et al., 2005; Lockhat et al., 2005; Petta et al., 2005; Varma et al., 2005) and a significant reduction in volume of rectovaginal endometriotic nodules (Fedele et al., 2001). Recent data on the use of a depot preparation of Medroxyprogesterone acetate (DMPA-SC 104) demonstrates that pain reduction is as effective as that observed with GnRH analogues (Crosignani et al., 2006). Limited data also exists on the use of another depot preparations (Implanon – Etonogestrel) in the management of endometriosis (Yisa et al., 2004; Yisa et al., 2005).

Combined oral contraceptives - continuous administration

Manipulation of the endogenous hormonal milieu is the basis for the medical management of endometriosis. The treatment of endometriosis with combination of estrogens and progestagens was originally used to induce "pseudopregnancy" with resultant amenorrhea due to decidualisation of endometrial tissue (Kistner, 1959). The modern equivalent is the continuous use of the combined oral contraceptive pill. Any low-dose combination oral contraceptive pill containing 30-35 mg of ethinyl oestradiol used continuously (to achieve amenorrhea) can be effective in the management of endometriosis (Moghissi, 1999). Symptomatic relief of dysmenorrhea and pelvic pain is reported in 60-95% of patients. Following a first-year recurrence rate of 17-18%, a 5-10% annual recurrence rate has been observed. Oral contraceptives are less costly than other treatment modalities and may be helpful in the management of endometriosis with potential long-term benefits in some women (Moore et al., 1997).

Combined oral contraceptives - cyclical administration

The cyclical use of combination oral contraceptives may provide prophylaxis against either the development or recurrence of endometriosis. Oestrogens in oral contraceptives potentially may stimulate the proliferation of endometriosis. However, the reduced menstrual bleeding that often occurs in women taking oral contraceptives may be beneficial to women with prolonged, frequent menstrual bleeding, which is a known risk factor for endometriosis (Cramer et al., 1986). Further research is warranted to assess the effect of low-dose oral contraceptives in preventing endometriosis and in treating associated pain. A Cochrane review of the use of the combined oral contraceptive pill identified only one study which demonstrated a reduction in non menstrual symptoms.


Gestrinone is a 19-nortestosterone derivative with androgenic, anti-progestagenic, anti-oestrogenic, and anti-gonadotropic properties. It creates a hormonal environment that results in the cellular inactivation and degeneration of endometriotic implants but not their disappearance (Brosens et al., 1987). Amenorrhea occurs in 50-100% of women and is dose-dependent.

The standard dose has been 2.5 mg twice a week, although it has been reported that 1.25 mg twice weekly is equally effective (Hornstein et al., 1990). The clinical side effects are dose-dependent and similar but less intense than those caused by danazol (Fedele et al., 1989). They include nausea, muscle cramps, and androgenic effects such as weight gain, acne, seborrhoea, oily hair/skin, and irreversible voice changes.

Gestrinone is as effective as GnRHa for the treatment of pelvic pain associated with endometriosis (Gestrinone Italian Study Group, 1996). However, gestrinone has fewer side effects. Pregnancy is contraindicated while taking gestrinone because of the risk of masculinisation of the foetus.


Danazol suppresses GnRH or gonadotropin secretion, directly inhibits steroidogenesis, increases metabolic clearance of oestradiol and progesterone, and interacts with endometrial androgen and progesterone receptors. In addition it causes immunologic attenuation of potentially adverse reproductive effects (Barbieri and Ryan, 1981; Hill et al., 1987). The multiple effects of danazol produce a high-androgen, low-oestrogen environment that does not support the growth of endometriosis, and the amenorrhea that is produced prevents new seeding of implants from the uterus into the peritoneal cavity.

Doses of 800 mg/day are frequently used in North America, whereas 600 mg/day is commonly prescribed in Europe and Australia. It appears that the absence of menstruation is a better indicator of response than drug dose. A practical strategy for the use of danazol is to start treatment with 400 mg daily (200 mg twice a day) and increase the dose, if necessary, to achieve amenorrhea and relieve symptoms (Wingfield and Healy, 1993).

The significant adverse side effects of danazol are related to its androgenic and hypo-oestrogenic properties. The most common side effects include weight gain, fluid retention, acne, oily skin, hirsutism, hot flushes, atrophic vaginitis, reduced breast size, reduced libido, fatigue, nausea, muscle cramps, and emotional instability. Deepening of the voice is another potential side effect that is non reversible. Danazol is contraindicated in patients with liver disease because it is largely metabolized in the liver and may cause hepatocellular damage. Danazol is also contrindicated in patients with hypertension, congestive heart failure, or impaired renal function because it can cause fluid retention. The use of danazol is contrindicated in pregnancy because of its androgenic effects on the foetus.

Gonadotropin-releasing hormone agonists

GnRH agonists bind to pituitary GnRH receptors and initially stimulate LH and FSH synthesis and release. However, prolonged stimulation causes down regulation of gonadotrophic activity. Consequently, ovarian steroid production is suppressed, providing a medically induced and reversible state of pseudo menopause.
Various GnRH agonists have been developed and used in treating endometriosis. These include leuprorelin buserelin, nafarelin, histrelin, goserelin, deslorelin, and tryptorelin. These drugs are inactive orally and must be administered intramuscularly, subcutaneously, or intranasally.

The side effects of GnRH agonists are caused by hypo-oestrogenism and include hot flushes, vaginal dryness, reduced libido, and reduction in bone density. Reversibility of bone loss is equivocal and therefore of concern (Barbieri, 1992; Riis et al., 1990), especially because treatment periods of longer than 6 months may be required. Where treatment is restricted to 6 months the effect on bone mineral density virtually resolves by 12 months (Makita et al., 2005). Side effects can be ameliorated by the use of "add-back". The goal of add-back is to effectively treat endometriosis and endometriosis-associated pain, while preventing vasomotor symptoms and bone loss. Add back can be achieved by progestagens only including norethisterone 1.2 mg (Riis et al., 1990), norethindrone acetate 5 mg (Hornstein et al., 1998), but bone loss is not prevented by medrogestone 10 mg/day (Sillem et al., 1999). Add-back can also be achieved by tibolone 2.5 mg/day (Taskin et al., 1997; Lindsay et al., 1996), or by an oestrogen/progestagen combination, i.e. conjugated oestrogens 0.625 mg combined with medroxyprogesterone acetate 2.5 mg (Friedman et al., 1993) or with norethindrone acetate 5 mg (Hornstein et al., 1998), oestradiol 2 mg and norethisterone acetate 1 mg (Franke et al., 2000). However, some concern remains about the long term effects of GnRH analogues on bone loss. In a recent report (Pierce et al., 2000), bone mineral density reduction occurred during long-term GnRH agonist use and was not fully recovered up to 6 years after treatment.

"Draw-back therapy" has been suggested as an alternative in a recent study showing that 6 months intake of 400 microgram nafarelin/day was as effective as "draw-back regimen" consisting of 1 month intake of 400 microgram nafarelin/day followed by 5 months 200 microgram nafarelin/day, with similar oestradiol levels (30 pg/mL) but less loss of bone mineral density (Tahara et al., 2000).

Aromatase inhibitors

Theoretically aromatase inhibitors may have a role to play in the medical management of endometriosis, particularly in postmenopausal women (Attar and Bulun, 2006; Bulun et al., 2000; D'Hooghe, 2003b). Although a number of small studies have confirmed this theoretical promise (Ailawadi et al., 2004; Amsterdam et al., 2005) there is insufficient data to support their widespread use in the management of endometriosis at the present time and particularly in a group of women of predominately reproductive age. It is likely that if aromatase inhibitors do find a role in the management of endometriosis then it will be as part of a combination therapy with other ovarian suppressant drugs (Attar and Bulun, 2006).

Anti-angiogenic therapies, progesterone anatagonists and Selective Progesterone Receptor Modulators (SERMs)

There are some interesting preliminary results from the laboratory and in animal models of the possible use of angiostatic drugs and anti-vascular endothelial growth factor but no clinical trials have yet been performed in humans (Ferrero et al., 2006).
Antiprogesterones, such as mifepristone, have been suggested as potential treatments for endometriosis but limited data is available to advocate wider use (Spitz, 2006; Tang and Ho, 2006; Chabbert-Buffet et al., 2005).
There is also little data to support the use of selective progesterone receptor modulators although their potential use has been advocated (Chabbert-Buffet et al., 2005; Chwalisz et al., 2005).

Empirical treatment of pain symptoms without a definitive diagnosis

Empirical treatment for pain symptoms presumed to be due to endometriosis without a definitive diagnosis includes counselling, adequate analgesia, progestagens, the combined oral contraceptive (COC) and nutritional therapy. It is unclear whether the COC should be taken conventionally, continuously or in tricycle regimen. A GnRH agonist may be taken but this class of drug is more expensive, and associated with more side-effects and concerns about bone density.

With no overwhelming medical evidence to support particular treatments over others, it is important to recognise that the decisions involved in any treatment plan are individual, and that the woman is able to make these based on an informed choice and a good understanding of what is happening in her body.
The resentment and frustration, which many with chronic diseases develop over time, more often than not result from a mismatch between clinical management and the woman's expectations - a mismatch that leaves her unprepared for the possibility of side effects or recurrence.

Appropriate counselling before, during, and after treatment, either by the treating physician and/or a counsellor/psychologist, is therefore an imperative part of the treatment process in endometriosis. The pros and cons of each potential treatment must be comprehensively communicated prior to its commencement, as must the setting of realistic expectations for outcome and potential follow-on treatment.
It is common practice to treat women suffering from dysmenorrhoea with analgesics, in fact many women treat themselves with oral analgesics purchased over-the-counter (without a prescription). A systematic review by Zhang and coworkers (1998) found that paracetamol was not more effective than placebo in reducing pain, whilst co-proxamol (paracetamol 650 mg and dextropropoxyphen 65 mg) reduced pain compared with placebo. However, these analyses were based on two relatively small randomised controlled trials comparing paracetamol and co-proxamol with placebo respectively (Evidence Level 1a). Additionally, the dosage of paracetamol was 500 mg four times daily and this may have been somewhat suboptimal. When paracetamol 1000 mg three times daily dosage was compared with either ibuprofen or naproxen no significant differences were found (Proctor and Farquar, 2006a). A recent small RCT again demonstrated that paracetamol (acetaminophen)1000 mg four times daily was superior to placebo for the treatment of primary dysmenorrhoea (Dawood and Khan-Dawood 2007).

A recent systematic review evaluated the use of non-steroidal anti-inflammatory drugs (NSAIDs) for dysmenorrhoea (Marjoribanks et al., 2003). This Cochrane review included the trials evaluating the effectiveness of NSAIDs for primary dysmenorrhoea. Primary dysmenorrhoea was described as menstrual pain without organic pathology; however, exclusion of pelvic pathology was based on physical examination. Hence, it is quite likely that the trials evaluated in this review included patients with endometriosis given that some women with so called 'primary dysmenorrhoea' probably have endometriosis. It is well known that some women who have so called 'primary dysmenorrhoea' have endometriosis. Marjoribanks and coworkers (2003) concluded that NSAIDs, except niflumic acid, were more effective than placebo for pain relief and that there was insufficient evidence to suggest whether any individual NSAID was more effective than others (Evidence Level 1a). Another review concluded that selective cyclo-oxygenase-2 inhibitors rofecoxib, lumiracoxib and etoricoxib were as effective as naproxen and more effective than placebo for the treatment of primary dysmenorrhoea (Proctor and Farquar, 2006a). However, concerns have been raised about the safety of these medications and its manufacturers have recently withdrawn rofecoxib from the market in many countries (Evidence Level 1b).

Combined oral contraceptives (COCs) are also used commonly for the management of patients with dysmenorrhoea, which in some cases may be due to endometriosis. However, there is a paucity of information for the use of modern COCs for primary dysmenorrhoea. A Cochrane review by Proctor et al (2001) suggested that 1st and 2nd generation COCs with 50 mcg or more oestrogen may be more effective than placebo treatment for dysmenorrhoea, however it concluded that the RCTs included for analysis were of poor quality and heterogenous so that no recommendation could be made regarding the efficacy of modern, lower dose COCs (Evidence Level 1a). A recent RCT comparing a low dose oral contraceptive containing 20 μg ethinyl estradiol and 100 μg levonorgestrel with placebo showed better pain relief in adolescent girls with dysmenorrhoea (Davis et al., 2005). Additionally there is some evidence in general populations that combined oral contraceptives can effectively treat dysmenorrhoa (Proctor and Farquhar 2006a). The COCs have the advantage of long term safety; hence they can be used indefinitely in low risk women. In clinical practice, when they are used for menstrual pain, they may be taken tricyclically or continuously to reduce the number of periods or to avoid them altogether (Evidence Level 4). However, there is no direct comparison of these options with the conventional approach.

Some authors advocate use of empirical medical therapy, as they see laparoscopy as an unnecessary, fruitless and hazardous procedure (Lindheim 1999). Similarly, a recent consensus statement from the 'Chronic Pelvic Pain/Endometriosis Working Group' suggested use of danazol, progestagens and GnRH agonists as 'second line' treatment without laparoscopic confirmation, when endometriosis was suspected as the cause of chronic pelvic pain (Gambone et al., 2002). Their recommendation is to continue with the 'advanced' or 'second line' treatment for six months and then to initiate an appropriate 'maintenance' treatment with NSAIDs or COCs. This statement suggests long term treatment with danazol, GnRH agonists or progestagens if symptoms recur upon reversion to maintenance treatment. The document does not provide evidence for the long term efficacy and safety of the approach, nor does it give data regarding compliance and recurrence rates. In the absence of data, there is a clear need for a randomised controlled trial to compare laparoscopic surgery with 'second line' medical treatment when the 'first line' treatment options (NSAIDs and COCs) fail.

Several Cochrane reviews and one Clinical Evidence review suggest that other treatment modalities which may be helpful in primary dysmenorrhoea include thiamine, vitamin E, high frequency transcutaneous nerve stimulation, topical heat and herbal remedy toki-shakuyaku-san. They also suggest that treatment modalities with unknown benefit are vitamin B12, fish oil, magnesium, acupuncture, other herbal remedies and behavioural interventions and that spinal manipulation is unlikely to be beneficial (Proctor and Murphy 2001; Proctor et al., 2006b; Proctor et al., 2002; Proctor and Farquar 2006a).

Hormonal treatment

Suppression of ovarian function for 6 months reduces endometriosis associated pain. The hormonal drugs investigated - COCs, danazol, gestrinone, medroxyprogesterone, acetate and GnRH agonists - are equally effective but their side-effect and cost profiles differ (Davis et al., 2007; Prentice et al., 2000; Prentice et al., 1999; Selak et al., 2007).

Manipulation of the endogenous hormonal milieu is the basis for the medical management of endometriosis. As oestrogen is known to stimulate the growth of endometriosis, hormonal therapy has been designed to suppress oestrogen synthesis, thereby inducing atrophy of ectopic endometrial implants or interrupting the cycle of stimulation and bleeding. Withdrawal of oestrogen stimulation causes cellular inactivation and degeneration of endometriotic implants but not their disappearance.

Most women with symptomatic endometriosis experience pain relief throughout treatment as shown in several prospective, randomized, placebo-controlled, double-blind studies (Davis et al., 2007; Prentice et al., 2000; Prentice et al., 1999; Selak et al., 2007). The effect lasts for a variable time after cessation of therapy. Progestagens, oral contraceptives, danazol, gestrinone and GnRH agonists are equally effective, so the choice of treatment is guided by side effects and cost.

Treatment of endometriosis-associated pain in confirmed disease

Non-steroidal anti-inflammatory drugs

There is inconclusive evidence to show whether NSAIDs (specifically naproxen) are effective in managing pain caused by endometriosis (Allen et al., 2005).
Level 1a

Non-steroidal anti-inflammatory drugs (NSAIDs) may be effective in reducing endometriosis associated pain (Kauppila et al., 1979; Kauppila and Ronnberg, 1985; Ylikorkala and Viinikka, 1983).

As endometriosis is a chronic inflammatory disease, anti-inflammatory drugs would appear attractive for treatment. Non-steroidal, anti-inflammatory drugs (NSAIDs) have become the most widely used therapeutic agents in the treatment of hyperalgesia induced by the inflammatory process. Although NSAIDs have been used extensively and have often been the first line therapy for reduction of endometriosis related pain, the analgesic effect of NSAIDs has not been studied extensively. Only one small, double-blind, placebo-controlled, four-period, cross-over clinical study has been published (Kauppila and Rönnberg, 1985). This study showed complete or substantial pain relief of endometriosis-related dysmenorrhoea in 83% of cases treated with naproxen compared with 41% in cases treated with placebo. Women who received naproxen needed significantly less supplemental analgesics compared to women on placebo. Some NSAIDs may act not only through central inhibition of the prostaglandin synthesis, but also through the activation of endogenous opioids and serotionergic mechanisms, which can explain the efficacy of NSAIDs in chronic pain conditions. In fact, NSAIDs and opiates have a synergistic effect and combined treatment may contribute to reduction or even prevention of morphine tolerance and be opiate sparing (Hanna et al., 2003).
Endometriosis-related pain is nociceptive (Bajaj et al., 2003), but persistent nociceptive input from endometriotic lesions leads to central sensitization manifested by somatic hyperalgesia and increased referred pain areas. The positive clinical experience of NSAIDs for reduction of endometriosis related pain may be explained by both a local anti-nociceptive effect and a reduced central sensitisation besides the anti-inflammatory effect.

It is important to note that NSAIDs have significant side-effects, including gastric ulceration. Prostaglandins are involved in the follicle rupture mechanism at ovulation and this is why NSAIDs should not be taken at ovulation time by women who wish to become pregnant (Duffy and Stouffer, 2002). Other analgesics may be effective as well but there is insufficient evidence to make specific recommendations.

A recent Cochrane review on the use of NSAIDs for pain in endometriosis appears to contradict the recommendation made here by stating that there is inconclusive evidence to show whether NSAIDS are effective (Allen et al., 2005). However, this review highlights the lack of high quality data in this area and the small size of the trials including the trial previously discussed (Kauppila and Ronnberg, 1985). The trend towards benefit demonstrated in this trial justifies the recommendation that NSAIDs may be effective in reducing endometriosis associated pain, especially when it is considered that as a group they have been shown to be effective for relieving the commonest symptom of endometriosis (dysmenorrhoea) in other situations (Marjoribanks et al., 2003).

Hormonal treatment

Suppression of ovarian function for 6 months reduces endometriosis associated pain. The hormonal drugs investigated - COCs, danazol, gestrinone, medroxyprogesterone, acetate and GnRH agonists - are equally effective but their side-effect and cost profiles differ (Davis et al., 2007 ; Prentice et al., 1999; Prentice et al., 2000; Selak et al., 2007).
Level 1a

There are pilot data suggesting that the aromatase inhibitor, letrozole, may be effective although it is associated with significant bone density loss (Ailawadi et al., 2004).

  A The levonorgestrel intra-uterine system (LNG IUS) reduces endomestriosis associated pain.
Evidence Level 1a

A systematic review identified two RCTs and three prospective observational studies, all involving small numbers and a heterogeneous group of patients (Varma et al., 2005). Nevertheless, the evidence suggests that the LNG IUS reduces endometriosis associated pain (Petta et al., 2005; Vercellini et al., 1999a) with symptom control maintained over 3 years (Lockhat et al., 2004; Lockhat et al., 2005).

Duration of GnRH agonist treatment

Treatment for 3 months with a GnRH agonist may be as effective as 6 months in terms of pain relief (Hornstein et al., 1995).
Evidence Level 1b

GnRH agonist treatment with 'add-back'

Treatment for up to 2 years with combined oestrogen and progestagen 'add-back' appears to be effective and safe in terms of pain relief and bone density protection; progestagen only 'add-back' is not protective (Sagsveen et al., 2003). However, careful consideration should be given to the use of GnRH agonists in women who may not have reached their maximum bone density.
Evidence Level 1a

Manipulation of the endogenous hormonal milieu is the basis for the medical management of endometriosis. As oestrogen is known to stimulate the growth of endometriosis, hormonal therapy has been designed to supress oestrogen synthesis, thereby inducing atrophy of ectopic endometrial implants or interrupting the cycle of stimulation and bleeding. Withdrawal of oestrogen stimulation causes cellular inactivation and degeneration of endometriotic implants but not their disappearance.

Most women with symptomatic endometriosis experience pain relief throughout treatment as shown in several prospective, randomized, placebo-controlled, double-blind studies (Davis et al., 2007; Prentice et al., 1999; Prentice et al., 2000; Selak et al., 2007). The effect lasts for a variable time after cessation of therapy. Progestagens, oral contraceptives, danazol, gestrinone and GnRH agonists are equally effective, so the choice of treatment is guided by side effects and cost.

Surgical treatment

Depending upon the severity of disease found, ideal practice is to diagnose and remove endometriosis surgically at the same time, provided that pre-operative adequate consent has been obtained (Abbott et al., 2003; Chapron et al., 2003b; Fedele et al., 2004a; Redwine and Wright, 2001).

There are no data to justify hormonal treatment prior to surgery to improve the success of surgery (Muzii et al., 1996; Audebert et al., 1998).

Ablation of endometriotic lesions plus laparoscopic uterine nerve ablation (LUNA) in minimal-moderate disease reduces endometriosis associated pain at 6 months compared to diagnostic laparoscopy; the smallest effect is seen in patients with minimal disease (Jacobson et al., 2001). However, there is no evidence that LUNA is a necessary component (Sutton et al., 2001), and LUNA by itself has no effect on dysmenorrhoea associated with endometriosis (Vercellini et al., 2003a).
Level 1b

There are no data supporting the use of uterine suspension but, in certain cases, there may be a role for pre-sacral neurectomy especially in severe dysmenorrhoea (Soysal et al., 2003).

Endometriosis associated pain can be reduced by removing the entire lesions in severe and deeply infiltrating disease. If a hysterectomy is performed, all visible endometriotic tissue should be removed at the same time (Lefebvre et al., 2002). Bilateral salpingo-oophorectomy may result in improved pain relief and a reduced chance of future surgery (Namnoum et al., 1995).

The goal of surgery is to excise or coagulate all visible endometriotic peritoneal lesions, endometriotic ovarian cysts, deep rectovaginal endometriosis and associated adhesions, and to restore normal anatomy. Laparoscopy can be used in most women, and this technique decreases cost, morbidity, and the risk of adhesions postoperatively. In the absence of deep endometriosis laparoscopy can be performed as day surgery. Laparotomy should be reserved for patients with advanced stage disease in whom laparoscopic surgery is not possible.

Peritoneal endometriosis

Endometriosis lesions can be removed during laparoscopy by excision, coagulation or vaporization by laser (carbon dioxide laser, potassium-titanyl-phosphate laser or argon laser). No controlled evidence is availale demonstrating that one laser technique is better than the other. The effectiveness of surgical ablation of peritoneal endometriosis has been convincingly shown in two RCTs where the control group underwent a laparoscopy without surgical ablations of lesions. The treated group had a significant reduction of symptoms that peristed for 12 months (Abbott et al., 2004) and 18 months (Sutton et al., 1997; Sutton et al., 1994).

The effectiveness of laparoscopic uterine nerve ablation (LUNA) in women with symptomatic endometriosis has not been proven (Vercellini et al., 2003a). The effectiveness of surgical treatment by laparotomy has not been investigated by a RCT. The many observational studies that have been published claim a high percentage of success.

Ovarian endometriosis

Superficial ovarian lesions can be coagulated or vaporized. The primary indication for extirpation of an endometrioma is to ensure it is not malignant. Small ovarian endometriomata (< 3 cm diameter) can be aspirated, irrigated, and inspected for intracystic lesions. Their interior wall can be coagulated or vaporized to destroy the mucosal lining. Ovarian endometriomata > 3 cm should be removed completely (Chapron et al., 2002b). In cases where excision is technically difficult without removing a large part of the ovary, a two-step procedure (marsupialisation and rinsing followed by hormonal treatment and surgery 3 months later) should be considered (Donnez et al., 1996). Although as little as one-tenth of an ovary may be enough to preserve function and fertility, at least for a while, there is increasing concern that ovarian cystectomy with concomitant removal or destruction of normal ovarian tissue may reduce ovarian follicle reserve and reduce fertility (Loh et al., 1999). Therefore, it has been proposed to replace cystectomy by fenestration and coagulation of the inner cyst wall (Hemmings et al., 1998) but a case-control study (Saleh and Tulandi, 1999) and a randomized controlled trial (Beretta et al., 1998) have demonstrated that pain and subfertility, related to ovarian endometriomas, were improved more by cystectomy than by fenestration/coagulation. Therefore, based on the current evidence, ovarian cystectomy seems to be the method of choice (Chapron et al., 2002) with a significantly decreased risk of cyst recurrence (Vercellini et al., 2003b).


If the endometriosis-related adhesions are part of an inflammatory fibrosis, they should be removed carefully. So far there is no evidence from randomized controlled trials that routine use of pharmacological or liquid agents prevent postoperative adhesions after fertility surgery (Watson et al., 2000).

Deep rectovaginal and rectosigmoidal endometriosis

Deep endometriosis is usually multifocal and complete surgical excision must be performed in a one step surgical procedure, in order to prevent more than one surgery, provided the woman is fully informed (Chapron et al., 2003a; Chapron et al., 2003b). Because of the fact that management of deeply infiltrating endometriosis is complex, referral to a center with expertise to offer all available treatments in a multidisciplinary approach is strongly recommended. Surgical management is only for symptomatic deeply infiltrating endometriosis. Asymptomatic patients must not be operated upon. Progression of the disease and appearance of specific symptoms rarely occurred in patients with asymptomatic rectovaginal endometriosis (Fedele et al., 2004b). When surgical treatment is decided the treatment must be radical with excision of all deep lesions. In cases of intestinal endometriosis, segmental resection improves the outcome without affecting the chance of conception (Fedele et al., 2004a).

Preoperativley the patients' agreement must be obtained to perform this difficult and high risk surgical procedure. RCTs are difficult since these severe cases are all very individual and moreover, not all surgeons are familiar with all treatment options. Complete excision might comprise the resection of the uterosacral ligaments, the resection of the upper part of the posterior vaginal wall, discoid or segmental bowel resection followed by end-to-end anastomosis, partial cystectomy and ureterolysis, eventually resection, re-anastomosis and re-implantation, sometimes still preserving the uterus and ovarian tissue.

Segmental rectosigmoid resection can be performed by laparotomy, laparoscopy or by laparoscopically assisted vaginal technique (Redwine et al., 1996). Surgical excision of deep rectovaginal and rectosigmoidal endometriosis is difficult and can be associated with major complications such as bowel perforations with resulting peritonitis (Koninckx et al., 1996). Preoperative laxatives, starch-free diet and full bowel preparation are needed to allow peroperative bowel suturing, if needed. Assessment of deep invasive endometriosis by TRS or MRI might be performed. A bowel contrast enema might be performed preoperatively if deep endometriosis is suspected by clinical exam and bowel symptoms. As endometriosis sometimes involves non-gynaecological organs, i.e. the bowel, the urinary tract or pelvic bones, other surgically devoted specialists such as bowel surgeons and urologists sometimes have to be involved. Bowel surgery has to be discussed preoperatively when indicated and planned accordingly. Ureteric catheters may be required before excision of periureteral endometriosis. A multidisciplinary approach involving gynaecological surgeons, bowel surgeons and urologists may be the safest approach; these severe cases should be handled by centres with special expertise. Moreover, as the pattern of pain in endometriosis is comlicated and pain does not always respond to treatment, having access to a multidisciplinary team including pain specialists is very important. As the women sometimes have multiple problems close collaboration with other health-care professionals is mandatory.

Oophorectomy and hysterectomy

Radical procedures such as oophorectomy or total hysterectomy are indicted only in severe cases. If a hysterectomy is performed, the cervix should be extirpated as persistent pain in a remaining cervix is common due to endometriosis in the cervix or endometriosis in the utero-sacral ligaments. However, it is important to note that women younger than 30 years at the time of hysterectomy for endometriosis-associated pain are more likely than older women to have residual symptoms, to report a sense of loss, and to report more disruption from pain in different aspects of their lives. Radical resection is an effective treatment for rectovaginal endometriosis. Hysterectomy and rectal resection were associated with a better response and quality of life (Ford et al., 2004).

Results of surgical treatment

The outcome of surgery in patients with endmetriosis and pain is influenced by many psychological factors related to personality, marital and psychosexual issues. It is difficult to evaluate the objective effect of different surgical approaches scientifically as not only the extirpation/destruction of the pathological tissue can impact on the results but also surgery per se, the doctor-patient relationship, complications etc. Diagnostic laparoscopy without complete removal of endometriosis has been found to alleviate pain in 50% of patients (Sutton et al., 1994). There is a significant placebo response to all kind of therapy. Similar results have been reported using oral placebo (Overton et al., 1994). Moreover, the long standing effect of surgery on pain is difficult to evaluate as the follow up time is not long enough, usually just a few months. The risk of recurrences is significantly correlated to the age of the patients. The younger the patients are at the moment of the diagnosis the higher the risk of recurrence. Higher recurrence rates in younger patients seems to justify a more radical treatment in this group (Fedele et al., 2004a).

The major shortcomings of surgical treatment in endometriosis related pain is the lack of prospective, RCTs with a follow up time long enough to draw clear clinical conclusions. In a prospective, controlled, randomized, double-blind study, surgical therapy has been shown to be superior to expectant management six months after treatment of mild and moderate endometriosis (Sutton et al., 1994). Treatment was least effective in women with minimal disease. One year later, symptom relief was still present in 90% of those who initially responded (Sutton et al., 1997).

In a randomised blinded crossover study it was confirmed that laparoscopic excision of endometriosis is more effective than placebo in reducing pain and improving quality of life (Abbott et al., 2004). Surgery resulted in pain relief in 80% of patients with severe disease who did not respond to medical therapy (Sutton and Hill, 1990). All these studies, however, have the drawback of no or very short follow up time.

Pre-operative treatment

Although hormonal therapy prior to surgery improves rAFS scores, there is insufficient evidence of any effect on outcome measures such as pain relief (Yap et al., 2004).
Level 1a

Post-operative treatment

Compared to surgery alone or surgery plus placebo, post-operative hormonal treatment does not produce a significant reduction in pain recurrence at 12 or 24 months, and has no effect on disease recurrence (Yap et al., 2004).
Evidence Level 1a

Postoperative hormonal treatment (GnRHa) does not produce a significant reduction in pain recurrence, but has a tendency to delay recurrence. Postoperative GnRHa treatment resulted in reduced pain scores, and in a delay of pain recurrence with more than 12 months, if the agonists were given for 6 months (Hornstein et al., 1997; Vercellini et al., 1999b) but not if they were administered only for 3 months (Parazzini et al., 1994). Similarly, postoperative hormonal treatment with danazol 100 mg/day (low dose) during 12 months after surgery for moderate to severe endometriosis resulted in a significantly lower pain score in the treated group when compared to a placebo group. In contrast, high dose danazol (600 mg/day) for 3 months was not superior to expectant management with respect to pain recurrence in  an identical patient population (Bianchi et al., 1999). In a RCT postoperative administration of low-dose cyclic oral contraceptives did not significantly affect the long-term recurrence rate of endometriosis after surgical treatment. A delay in recurrence was evident at life-table analysis (Muzii et al., 2000). In a small RCT, the LNG IUS, inserted after laparoscopic surgery for endometriosis associated pain, significantly reduced the risk of recurrent moderate-severe dysmenorrhoea at 1 year follow-up (Vercellini et al., 2003c).
As endometriosis is a chronic oestrogen-dependent disease, further hormonal treatment may be needed, but the data is insufficient and further studied are needed.

Hormone replacement therapy

Hormone replacement therapy (HRT) is recommended after bilateral oophorectomy in young women given the overall health benefits and small risk of recurrent disease while taking HRT (Matorras et al., 2002). The ideal regimen is unclear: adding a progestagen after hysterectomy is unnecessary but should protect against the unopposed action of oestrogen on any residual disease. However, the theoretical benefit of avoiding disease reactivation and malignant transformation should be balanced against the increase in breast cancer risk reported to be associated with combined oestrogen and progestagen HRT and tibolone (Beral and Million Women Study Collaborators, 2003).

Level 4





This guideline, which is reviewed annually, was last updated on 30 June 2007

The guideline is hosted free of charge by | Contact webmaster