Index
Contributors
Search

DEFINITION

Endometriosis is a disease in which endometrial glands and stroma implant and grow in areas outside the uterus (See Fig.1). The most common place to find implants is in the peritoneal cavity, but endometriotic lesions occasionally have been found in the pleural cavity, liver, kidney, gluteal muscles, bladder, and even in men. The anatomical location and inflammatory response to these lesions are believed to account for the symptoms and signs associated with endometriosis.

Figure 1. Common locations of endometriosis within the pelvis and abdomen. (Reprinted by permission from the New England Journal of Medicine 2001. Olive DL, Pritts EA. Treatment of Endometriosis. Vol. 345:267)

SYMPTOMS AND SIGNS

The most common symptom for women who have endometriosis is that of cyclic pelvic pain, although many women with endometriosis may not have this complaint. The pain usually begins just before menses and continues throughout the duration of menstrual flow and typically includes dysmenorrhea, dyspareunia, dysuria, or dyschezia. The pain also may be perceived in musculoskeletal regions, such as the flank, low back, or thighs.

The next most common symptom is that of infertility. It is apparent that women with moderate and severe disease can have fertility problems due to the mechanical blockage of sperm-egg union, either from adhesions or disruption of normal pelvic anatomy. Interestingly, women with minimal or mild disease also have decreased fertility outcomes when compared with those without clinical evidence of endometriosis. The exact causes remain elusive to date, but several postulated mechanisms will be addressed below.

Unfortunately, due to the diffuse and often pusillanimous nature of endometriotic lesions, the physical examination is typically unrevealing. The findings are variable and of limited precision in either localization or diagnosis of endometriosis. However, the astute clinician can appreciate pain or induration in the vicinity of otherwise nonpalpable lesions, most commonly in the cul-de-sac or rectovaginal septum. The clinician also may appreciate uterine or adnexal fixation, or an adnexal mass. Tender nodules may be palpable along the uterosacral ligaments or within the cul-de-sac, especially if the examination is performed just before menses. Because much disease is found in the dependent areas of the pelvis, it is critical to perform a systematic rectovaginal examination. In that way the practitioner is assured of evaluating the cul-de-sac and uterosacral ligaments as well as the adnexa (1).

CLASSIFICATION

The scheme most widely used to classify the extent of disease was promulgated by the American Fertility Society in 1985 (2) (See Fig's. 2 and 3). This scheme designates disease extent based upon the total 3-dimensional volume of endometriosis. Importance is placed upon depth of invasion, bilaterality, and ovarian involvement, as well as density of associated adhesions and extent of cul-de-sac involvement. From this system, point scores are assigned and tallied, with scores of 1-15 representing minimal or mild disease, 16-40 moderate, and > 40 severe. It is important to note that this staging system was established to predict fertility outcomes, and does not correlate with the more common symptom of pelvic pain.

Figure 2. The American Fertility Society Revised Classification of Endometriosis. (Reprinted by permission from Fertil Steril 1985;43:351)

Figure 3. Examples and guidelines for use of the American Fertility Society Revised Classification of Endometriosis. (Reprinted by permission from Fertil Steril 1985;43:352)

ETIOLOGY

Classically, three theories exist to explain the etiology of endometriosis; 1) Sampson's theory, 2) Meyer's theory, and 3) Halban's theory. The most oft quoted theory, and to date the one supported by the most evidence, is Sampson's theory of transplantation and implantation. This theory stemmed from observations made during surgeries in the 1920's that many women shed endometrial debris through their fallopian tubes into the peritoneum during menstruation (3). Not only has viable endometrial tissue been found in fallopian tubes and peritoneal fluid of women, but in humans and other animals, endometrial tissue will grow if placed ectopically. Finally, endometriosis seems to occur most commonly in the gravitationally dependent parts of the pelvis, further supporting this theory (4). One problem with this theory is that retrograde menstruation has been shown in 76-90% of menstruating women, a prevalence much higher than that of endometriosis (5,6). This discrepancy suggests that additional factors beyond the presence of ectopic tissue are needed to establish the disease.

Meyer's theory (7) suggests that metaplasia of the coelomic epithelium is the origin of endometriosis. This theory is logical, as cells from both the peritoneum and endometrium derive from a common embryological precursor: the coelomic cell. However, this has been a difficult theory to support scientifically. If this postulate were correct, one would expect much higher rates of pleural endometriosis, since this mesothelium is the same as that which lines the abdominal cavity, than are observed. Furthermore, investigators have been unable to show that peritoneal cells can be differentiated experimentally into endometrial cell types. Lastly, metaplasia is an age-related process. Endometriosis occurs early in reproductive aged women, and its incidence peaks at 28 years, a pattern that would not be predicted by a linear likelihood of metaplasia.

Halban's theory is one that suggests that distant lesions are established by the hematogenous or lymphogenous spread of viable endometrial cells. Although this metastatic theory explains rare endometriotic lesions in the brain or lung, it does not explain the gravity dependent location of most foci of endometriosis.

Sampson's transplantation and implantation theory is the most widely accepted, but most researchers agree it grossly simplifies the disease process. Whilst retrograde menstruation can transplant tissue fragments into the peritoneal cavity, the cells themselves must escape apoptosis (8), adhere to the underlying peritoneum (9,10), degrade the underlying extracellular matrix (11), generate a new vascular supply (12), and evade the immune surveillance system (13). New studies have recently surfaced which address the disease on an even more fundamental level, evaluating the expression of some of the cell cycle regulatory genes. Although these data are preliminary, evidence is arising that links endometriosis with the loss of several cell cycle regulatory genes (14,15).

DEMOGRAPHY

Endometriosis is diagnosed in women aged 12-80, with the average age approximately 28. Exposure to ovarian hormones appears to be essential with the vast majority of the cases found in women aged 20-50. There is no known racial or socioeconomic predilection (16). Although with severe disease, there seems to be a familial correlation, no clear mendelian genetic trait has been identified and most researchers believe it to be multifactorial (17).

PREVALENCE

The prevalence of endometriosis ranges from 2-50% of reproductive aged women. Unfortunately, this number is widely disparate depending upon the study. In women with infertility, the prevalence ranges from 21-47%. This may be due to the fact that endometriosis plays a causative role in infertility, but it also may be due to a diagnostic selection bias, as women with infertility typically undergo laparoscopy as part of their clinical evaluation.

Between the years 1965 and 1984, endometriosis increased from 10 to19% as the primary indication for hysterectomy in the USA. Interestingly, this happened during a time in which a trend towards more conservative therapies as treatment modalities for endometriosis occurred (18). This finding suggests a true increase in the incidence of this disease. Varying theories have been proposed to explain the increase, including delay of childbearing, declining use of oral contraceptives, and exposure to environmental toxins such as dioxin as a causative agent (19). These theories, while intriguing, require further investigation.

HISTOLOGY

Although the histopathologic finding of ectopic endometrial glands and stroma is the sine qua non for establishing diagnosis of disease, only about 50-70% of presumed endometriosis specimens fulfill these criteria. Many specimens harbor fibrosis, chronic inflammation, and/or hemosiderin-laden macrophages. Most pathologists and clinicians accept these latter findings as highly suggestive of disease status.

CLINICAL APPEARANCE

Grossly, peritoneal endometriosis can take on many visual appearances. Classically, it was taught that endometriosis implants were blue-black "powder burns" or "mulberry lesions" of the peritoneum. In recent years, several stages of implant development have been appreciated, each with a corresponding appearance. Early, active lesions can appear as papular excrescences or vesicles, and can range in color from clear to pink to bright red (20). About a third of these lesions are in phase with the eutopic endometrium, and have a tendency to spontaneously grow and regress, suggesting a fluctuation of proliferation in association with hormonal production during the menstrual cycle (21). Advanced, active lesions are associated with inflammation, fibrosis and hemorrhage, and take on a more classic appearance identifiable at surgery. These lesions can express a myriad of colors, from black to brown, purple, red, or green. These are due to the presence of heme degradation products as the foci undergo hemorrhage and fibrosis. Dormant and healed lesions take on either a white or calcified appearance, and represent remnants of glands embedded in fibrous tissue (20).

A specific manifestation is the endometrioma or chocolate cyst. These ovarian cysts gained their moniker by the characteristic chocolate syrup appearance of their contents often seen at rupture. They arise after implantation of ectopic endometrial tissue and subsequent invasion into the normal ovarian cortex. The cell types present in these cysts include endometrial epithelium, both as glands and flattened cells, endometrial stroma and hemosiderin-laden macrophages. In some cases ciliated cells, similar to those of oviduct epithelium, have been observed (22).

Under scanning electron microscopy, microscopic lesions have been found in the normal appearing peritoneum of women with and without endometriosis (23). The clinical significance of these findings is a matter of controversy at present.

DIAGNOSTIC MODALITIES

Laparoscopy is considered the gold standard in diagnosis of endometriosis. Dogma states that a biopsy of the lesion is the only way to truly confirm a disease diagnosis. For those surgeons trained in advanced laparoscopy, direct visual recognition of endometriosis also allows treatment to be instituted immediately. Laparoscopy is preferred over laparotomy as it provides visualization of the entire abdomen and pelvis at magnified views, has less morbidity than laparotomy, and carries a decreased risk of adhesion formation (24, 25).

One of the major limitations of diagnostic laparoscopy for pelvic pain is the assumption that those lesions seen, (or not seen), directly correlate with the symptoms of pain. This assumption may or may not hold true; thus "conscious pain mapping" was developed in an attempt to add information to the simple process of visual inspection of the pelvis at laparoscopy. Utilizing local anesthesia with light sedation, minimal insufflation of the pelvis, and laparoscopes and instrumentation as small as 2 mm in diameter, the pelvis is examined and probed in a standardized fashion. Patients are asked to rate the pain during systematic probing of the pelvis, as well as during direct contact of any visible endometriosis lesions. In this manner, the pain can be localized to certain organs, to specific lesions or not. The ensuing operative laparoscopy can then be modified to focus on excision or repair of the abnormalities that are associated with the pain (26).

Attempts to develop radiological imaging techniques as diagnostic tools for endometriosis have been compromised by lack of specificity. Computed tomography was used initially to diagnose endometriomas in the setting of adnexal masses. Unfortunately, not only was it difficult to distinguish between benign versus malignant masses, it was often impossible to distinguish between adnexal structures and loops of bowel (27,28). Hence, this modality is of little value for this purpose. With more recent technological advances, transvaginal ultrasonography has been employed to evaluate possible endometriomas pre-operatively. Sonographic features characteristic of endometriomas have been described (29), but specificity of these findings to endometriomas versus malignancy or other benign ovarian masses is lacking.

At present, magnetic resonance imaging is the best imaging for identifying endometriomas, and can resolve implants as small as 3 mm in size (30,31). It also has been shown to differentiate benign from malignant lesions, with excellent sensitivity and specificity, and is the method of choice for preoperative evaluation of any adnexal mass (33,34).

Monoclonal antibodies raised against a high molecular weight ovarian cancer epithelial cell antigen, CA-125, have been used as a biochemical marker of endometriosis. Moderate and severe endometriosis is associated with elevated levels of CA-125 in the peripheral blood (35). Unfortunately, this marker is relatively nonspecific, being increased in ovarian cancer, liver disease, colon cancer, pelvic inflammatory disease, and even during menstruation of women without disease. Although it may help define progression of disease in individual patients, its diagnostic utility is limited.

INFERTILITY AND ENDOMETRIOSIS

Women with moderate and severe endometriosis, particularly those in which the ovaries and oviducts are involved with adhesive disease, have decreased fertility rates. It is theorized that this stems from mechanical obstruction between the ovary and oviduct, with subsequent failure of gamete transport into the tubal ampulla. Although no published trials compare surgical treatment versus no treatment, many publications cite a near zero pregnancy rate for those with advanced stage disease (36), and a plethora of uncontrolled trials show that there is a benefit of surgical restoration of normal pelvic anatomy in these patients (37). Hence, surgery can be useful in enhancing pregnancy rates for women with moderate to severe endometriosis.

When considering the effects of minimal or mild endometriosis, there is much more controversy. As indicated above, an association between early stage endometriosis and impaired reproductive function exists, but the exact relationship remains unclear (38). Theories as to pathophysiology in mild endometriosis include ovulatory dysfunction, oocyte, sperm, or embryo toxicity, abnormalities of the immune system, and defective endometrial receptivity in women with endometriosis (39). A meta-analysis of the two randomized trials in the literature that address this issue demonstrated that surgical treatment enhanced the rate of fertility after ablation of minimal or mild disease (40). Hence surgical ablation or excision confers fertility benefit in women even with minimal or mild disease.

Ovulation induction in women with endometriosis confers fertility benefit for women with endometriosis as well. The three randomized trials in the literature addressing this issue, employing either GnRH agonists with follicle stimulating hormone and luteinizing hormone, clomiphene citrate/intrauterine insemination, or follicle stimulating hormone/intrauterine insemination, all showed increased pregnancy rates relative to the no treatment arms (40). Thus, assisted reproductive technology (ART) has a place in the treatment of the women with endometriosis-associated infertility.

Conversely, suppressive medical therapy, (chronic GnRH agonists, oral contraceptives, danazol, and progestogens), has not been shown to be of clinical benefit to fertility when compared to no treatment at all (41). In fact, not only do these drugs confer no benefit by virtue of their anovulatory effects, but they also may delay pregnancy. A recent meta-analysis of the randomized trials showed a decrease in fertility when factoring in the time delay of medical treatment (42); thus it appears that current medical therapy (with the exception of ART) is not beneficial in the treatment of endometriosis-associated infertility. New evidence indicates that the eutopic endometrium of women who suffer from endometriosis is abnormal in its expression of implantation-associated proteins, e.g., complement protein C3 (43), IL-6 (44), HOX A 10 and 11 (45), and b3 integrins (46). These exciting observations may lead to beneficial medical treatments for infertility in the future.

Another issue of debate is that of the effect of endometriosis upon in vitro fertilization success. The evidence available from observational studies is at odds, with studies showing both decreases in fertility rates and no differences (47). Most of the earlier data showed a negative correlation between in vitro fertilization outcomes and severity of disease, but with the replacement of laparoscopic follicle aspiration by transvaginal aspiration techniques, there have been some studies showing no difference in pregnancy rates (48,49,50). This controversy also holds true even if women with endometriomas are included in the studies (47). However, recent data indicate that implantation rates are decreased in IVF patients with endometriosis (51). It is unclear, at present, whether treatment of endometriosis prior to attempted in vitro fertilization will enhance success rates. Further evidence is needed before a recommendation can be made.

PAIN AND ENDOMETRIOSIS

Pain is the most common finding in women with endometriosis, although the pathophysiologic mechanisms are not well understood. The heterogeneity of the disease process makes it difficult to ascertain the exact etiology of the pain. There are theories that different types of lesions cause pain through differing pathways. Early papular lesions contain higher levels of prostaglandins than older lesions. These prostaglandins may activate afferent neuronal pathways. Lesions deeply penetrating the peritoneum also have increased propensity to cause pain. Adhesions and fibrosis may cause pain by compromising the blood supply to certain neuronal plexi, or by placing small nerves on tension (52). Direct irritation of pelvic nerves by infiltration of lesions has also been observed (53). Clinicians need good evidence if they expect to treat the endometriosis patient in an acute setting, before the complications of chronic pelvic pain set in. There have been two trials comparing medical treatment versus placebo, and multiple randomized trials comparing either GnRH agonists, danazol, or oral contraceptives to address the question of the efficacy of medical treatment for endometriosis-associated pain (40). The two randomized trials with a placebo group showed significant decrease in pain when this was compared with GnRH agonists, medroxyprogesterone (MPA), or danazol. The effects of the medications were apparent up to six months after cessation of therapy (54,55). Of 18 trials comparing GnRH analogue with danazol, no differences in pain were observed, but both treatments led to improvement over premedication pain scores. In the single trial comparing a GnRH agonist and an oral contraceptive, overall pain scores were similarly decreased, but subgroup analyses revealed oral contraceptives to be less effective in suppressing dyspareunia and dysmenorrhea (56). It is important to understand that the oral contraceptives were given in a routine fashion with one week of hormone-free tablets, leading to monthly withdrawal bleeding. No studies have yet addressed continuous oral contraceptive treatment and pain. From the available evidence it is clear that medical treatment is effective for endometriosis-associated pelvic pain.

It is important to recognize that medical treatment is only effective if patient compliance remains high. Despite its efficacy, the side effects of GnRH-agonists include hot flashes, bone loss, and theoretical impairments in brain and cardiac function. "Add-back" therapy has been studied in multiple trials, and has been shown not only to diminish the side effects of the medication, but also to have no adverse effect on either the pain relief conferred from the medication, nor on the extent of disease seen at second-look laparoscopy (40). Practitioners today should feel comfortable with initiating add-back therapy at the onset of medical therapy if GnRH agonist is being used.

Good surgical trials are less common in the literature than medical trials, but there is one randomized trial comparing ablation of endometriosis, lysis of adhesions and ablation of the uterosacral nerves versus diagnostic laparoscopy with no additional intervention. Pain was decreased by 22.6% in women undergoing diagnostic laparoscopy, but there was a 62.5% decrease in pain in those women undergoing extensive surgery. Ninety percent of the latter continued to experience less pain one-year post-operatively (57). The combination of lesion ablation, adhesiolysis and uterosacral nerve ablation also appears to be an effective treatment for women with endometriosis-associated pain. Four randomized trials compared the efficacy of lesion ablation, adhesiolysis and either presacral nerve resection or uterosacral nerve ablation versus lesion ablation and adhesiolysis alone (40). Although no added benefit was noted with addition of uterosacral nerve ablation, there was more midline pain relief when the presacral nerve was resected. Lateral pain, as would be expected, was unaffected by addition of this procedure. It appears that in women with endometriosis-associated pain, lesion ablation and adhesiolysis are effective in reducing pain, with the addition of presacral neurectomy for those with midline pain. It should be emphasized that the latter procedure carries risks of bleeding from the neurovascular plexus and requires excellent operative skills.

There have been no trials comparing medical and surgical treatment to subdue endometriosis-associated pain. The studies employing each treatment modality, though, have similar success rates in all of the studies. With this in mind, it is reasonable to conclude that the treatments are equally effective. Recommendations for either drug or surgical management should be based upon patient preference and the ability of the operating team to perform these technically difficult procedures.

A logical extension of this evidence is to assess whether combined surgical and medical treatment is of value in women with endometriosis-associated pain. There are three randomized trials in the literature that evaluate the effects of postoperative medical treatment versus placebo (40). The first trial showed a significant benefit if MPA or danazol was used for 6 months after lesion ablation and adhesiolysis were performed compared to placebo. The second two trials compared GnRH agonist with placebo for either 3 or 6 months following surgery. Although the women in the placebo arms showed no differences in pain scores over 3-6 months in each study, there was a reduced need for a second operative procedure within the following 2 years in women who received GnRH agonist for 6 months. It is important to understand however, that diagnostic laparoscopy alone has an important effect on pain, as high as 50% in some studies. Beneficial effects of these diagnostic laparoscopy usually last anywhere from 3-6 months postoperatively. To further complicate this issue, a recent randomized trial compared postoperative medical treatment with either GnRH agonist or progestogen for 6 months followed by second look laparoscopy. Not only was the post-operative addition of GnRH agonist more effective than progestogen at relieving dyspareunia, dysmenorrhea and general pelvic pain, but at second look laparoscopy the AFS endometriosis scores were significantly lower in this group (58). Although the results are not conclusive at present, these small studies suggest an enhancement of surgical pain relief with the addition of medical therapy, particularly the use of GnRH analogues.

CONCLUSION

Endometriosis is an enigmatic disease, the pathophysiology of which we are just beginning to understand. Symptomatic women with endometriosis, suffering from infertility or pain, are often difficult patients to treat because we have few treatment options to offer. The therapies themselves are imperfect, with none that permanently eradicate the disease.

In deciding how to treat women with pelvic pain or infertility, we must consider the best available evidence to form our decisions (See Fig. 4, below). In women with pelvic pain and suspected endometriosis, first line treatment would be non-steroidal anti-inflammatory medications or oral contraceptives. Although the evidence is lacking, it makes theoretical sense to use continuous oral contraceptive therapy. If these conservative approaches fail, two alternative treatments would be empiric GnRH agonist therapy with estrogen and progestin add-back therapy, or operative laparoscopy. The laparoscopy should certainly include lysis of adhesions and excision of the endometriosis, with or without a presacral neurectomy depending upon the localization of pain and experience of the surgeon. Surgery for pain should be followed by medical treatment with GnRH agonist and add-back therapy. It is important to note that at no time are narcotics advocated for the treatment of endometriosis-associated pelvic pain except immediately pre-operatively or post-operatively. If the above measures fail, patients should be enrolled into a multi-disciplinary chronic pain treatment group. This should include physicians from many subspecialties, including psychiatry, anesthesiology, gynecology, and often gastroenterology and urology.

Figure 4. Algorithm for treatment of endometriosis-associated pain (panel A), and infertility (panel B). Where multiple pathways are shown, the path is guided by medical judgement and patient preference. Reprinted by permission from the New England Journal of Medicine 2001. Olive DL, Pritts EA. Treatment of Endometriosis. Vol. 345:267)

In treating endometriosis-associated infertility, the first line of defense should be operative laparoscopy, with removal of endometriosis and restoration of normal pelvic anatomy. At this juncture, either assisted reproduction could be employed, or watchful waiting. Some practitioners dispense with the operative laparoscopy and recommend ART directly. This is also a reasonable option, and no adequate data exist to strongly favor either recommendation. Ultimately, an unbiased discussion of the available evidence with the patient will best guide the practitioner to establish a treatment plan.