CHAPTER 10: Endometriosis

Endometriosis is a common benign disorder defined as the presence of endometrial glands and stroma outside of the normal location. Implants of endometriosis are most often found on the pelvic peritoneum, but other frequent sites include the ovaries and uterosacral ligaments. Endometrial tissue located within the myometrium is termed adenomyosis and discussed in Chapter 9. Women with endometriosis may be asymptomatic, subfertile, or suffer varying degrees of pelvic pain. This is an estrogen-dependent disease and thus lends itself to hormone-based treatment. However, in those with disease refractory to medical management, surgery may be required.

The incidence of endometriosis is difficult to quantify, as women with the disease are often asymptomatic. Moreover, imaging modalities have low sensitivities for small implants (Wall, 2015). The primary method of diagnosis is laparoscopy, with or without biopsy for histologic diagnosis (Dunselman, 2014). Using this standard, the annual incidence of surgically diagnosed endometriosis was 1.6 cases per 1000 women aged between 15 and 49 years (Houston, 1987). In asymptomatic women, the prevalence of endometriosis ranges from 6 to 11 percent, depending on the population studied and mode of diagnosis (Buck Louis, 2011; Mahmood, 1991). However, because of its link with infertility and pelvic pain, endometriosis is notably more prevalent in subpopulations of women with these complaints. From studies, the prevalence lies between 20 to 50 percent in infertile women, and in those with pelvic pain, it ranges from 40 to 50 percent (Balasch, 1996; Eskenazi, 2001; Meuleman, 2009). In adolescents, Janssen and coworkers (2013) reported that nearly two thirds of adolescents undergoing diagnostic laparoscopy for pelvic pain had evidence of endometriosis.

Previously, white women were thought to be disproportionately affected. More recent studies have provided variable results. Some show greater rates for whites and Asians, whereas others have found no statistically significant differences in endometriosis prevalence among any racial or ethnic groups (Jacoby, 2010). Of other patient characteristics, lower body mass appears to positively correlate with endometriosis risk (Peterson, 2013; Shah, 2013).

Pathogenesis

The definitive cause of endometriosis remains unknown, but theories have been proposed. A more favored one describes retrograde menstruation through the fallopian tubes (Sampson, 1927). These refluxed endometrial fragments invade the peritoneal mesothelium and develop a blood supply for implant survival and growth. Supporting data include a report that surgical obliteration of the outflow tract in baboons induces endometriosis (D’Hooghe, 1997). In correlation, women with outflow tract obstruction also have a high incidence of endometriosis, which often resolves following obstruction relief (Sanfilippo, 1986; Williams, 2014). Importantly however, most women have retrograde menstruation (Halme, 1984). Thus, other factors, such as immunologic and angiogenic components, likely aid implant persistence.

Another hypothesis, the stem cell theory, implicates undifferentiated endometrial cells that initially reside in the endometrium’s basalis layer. These cells differentiate into epithelial, stromal, and vascular cells as the endometrium is routinely regenerated each cycle. If displaced to an ectopic location, such as by retrograde menstruation, these stem cells may give rise to endometriosis (Valentijn, 2013).

Aberrant lymphatic or vascular spread of endometrial tissue has also been implicated (Jerman, 2015). Lymphatic spread of endometriosis to pelvic sentinel lymph nodes is noted in affected women (Mechsner, 2008; Tempfer, 2011). Findings of endometriosis in unusual locations, such as the groin, also bolster this theory (Mourra, 2015). Last, cases in which no peritoneal implants are found, but solely isolated retroperitoneal lesions are noted, implicate lymphatic spread (Moore, 1988).

Another theory concerns coelomic metaplasia and suggests that the parietal peritoneum is pluripotent and can undergo metaplastic transformation to tissue histologically identical to normal endometrium. Because the ovary and the progenitor of the endometrium, the müllerian ducts, are both derived from coelomic epithelium, such metaplasia may help explain endometriosis involving the ovary. This process may also underlie cases of endometriosis in those without menstruation, such as premenarchal girls and males treated with estrogen and orchiectomy for prostate cancer (Marsh, 2005; Taguchi, 2012). Last, a theory purports that müllerian remnants left along their embryonic path undergo abnormal differentiation (Batt, 2013; Signorile, 2012).

Anatomic Sites

Endometriosis may develop anywhere within the pelvis and on other extrapelvic peritoneal surfaces. Most commonly, endometriosis is found in the dependent areas of the pelvis. As such, the anterior and posterior cul-de-sacs, other pelvic peritoneum, the ovary, and uterosacral ligaments are frequently involved. Additionally, the rectovaginal septum, ureter, and bladder and rarely, pericardium, surgical scars, and pleura may be affected. One pathologic review revealed that endometriosis has been identified on all organs except the spleen (Markham, 1989). Implants may be superficial or they may be deep infiltrating endometriosis (DIE), that is, infiltrative forms that involve vital structures such as bowel, bladder, and ureters (Koninckx, 2012; Vercellini, 2004). Some definitions of DIE also quantify invasion as >5 mm (Koninckx, 1994).

As just noted, ovarian endometriomas are frequent manifestations of endometriosis (Fig. 10-1). These smooth-walled, dark-brown ovarian cysts are filled with a chocolate-appearing fluid and may be unilocular or, when larger, multilocular. Their pathogenesis is unclear, yet three theories include invagination of ovarian cortex implants, coelomic metaplasia, and secondary involvement of functional ovarian cysts by endometrial implants located on the ovarian surface (Vignali, 2002).

Molecular Mechanisms

Endometriosis is an estrogen-dependent, chronic inflammatory disease with aberrant growth of ectopic endometrial tissue. In this discussion, eutopic endometrium is that which lines the uterine cavity, whereas ectopic endometrium describes that outside the cavity. In affected patients, ectopic endometrial implants show molecular differences from the eutopic endometrium of unaffected women. The disturbed molecular mechanisms in this disease are yet to be completely defined. However, suspected underpinnings include an environment of estrogen dominance, estrogen dependence, and progesterone resistance within implants; inflammation; escape from immune clearance; local invasion and neurovascularity development; and genetic predisposition.

Estrogen and Progesterone

Estrogen plays a causative role in endometriosis formation and is derived from multiple sources. First, most estrogen in women is produced directly by the ovaries. Second, peripheral tissues also produce estrogens through conversion of ovarian and adrenal androgens by the enzyme aromatase. Endometriotic implants express aromatase and 17β-hydroxysteroid dehydrogenase type 1, which are the enzymes responsible for conversion of androstenedione to estrone and of estrone to estradiol, respectively. Implants, however, are deficient in 17β-hydroxysteroid dehydrogenase type 2, which inactivates estrogen (Kitawaki, 1997; Zeitoun, 1998). This enzymatic combination ensures that implants create an estrogenic environment. Moreover, it provides the rationale for aromatase inhibitor use to diminish aromatase activity in refractory clinical cases. Last, the endometriotic stromal cell uniquely expresses the full complement of genes in the steroidogenic cascade, which is sufficient to convert cholesterol to estradiol itself (Bulun, 2012).

In addition to an estrogenic environment, normal progesterone effects are attenuated in endometriosis. This progesterone resistance is thought to stem from an overall low concentration of progesterone receptors within implants (Attia, 2000). Specifically, pathological overexpression of estrogen receptor β in endometriosis suppresses estrogen receptor α expression. This diminishes estradiol-mediated induction of the progesterone receptor in endometriotic cells (Xue, 2007).

As one consequence of this resistance, survival of refluxed endometrium in affected women may be bolstered. Namely, normal endometrium does not express aromatase and has elevated levels of 17β-hydroxysteroid dehydrogenase type 2 in response to progesterone (Satyaswaroop, 1982). As a result, progesterone antagonizes the estrogen effects in normal endometrium during the luteal phase. Endometriosis, however, manifests a relative progesterone-resistant state, which prevents this antagonism in its implants.

Progesterone resistance may also enhance implantation of refluxed endometrium. Invasion of the mesothelium can be aided by matrix metalloproteinases (MMPs). These are a group of collagenase proteins that can digest and remodel extracellular matrix and are implicated in endometrial turnover during normal menstruation. Of the various MMPs, MMP-3 expression is significantly increased in women with endometriosis compared with healthy controls, and its expression is significantly elevated during the luteal phase (Kyama, 2006). Progesterone represses MMP activity (Itoh, 2012). Thus, in affected patients, progesterone resistance within these implants may augment the MMP activity necessary for implant invasion.

Inflammation

Prostaglandin E₂ (PGE₂) is the most potent inducer of aromatase activity in endometrial stromal cells (Noble, 1997). Estradiol produced in response to the increased aromatase activity subsequently augments PGE₂ production by stimulating the cyclooxygenase type 2 (COX-2) enzyme in uterine endothelial cells (Gurates, 2003). This creates a positive feedback loop and potentiates the estrogenic effects on endometriosis proliferation. As discussed on page 239, nonsteroidal antiinflammatory drugs (NSAIDs) are used clinically to reduce prostaglandin formation and thereby decrease endometriosis-linked pain.

Immune System

With retrograde menstruation, refluxed menstrual tissue in most women is usually cleared by macrophages, natural killer (NK) cells, and lymphocytes. For this reason, immune system dysfunction is one likely mechanism for endometriosis establishment (Seli, 2003). Of these immune cells, macrophages serve as scavengers, and increased numbers are found in the peritoneal cavity of women with endometriosis (Haney, 1981; Olive, 1985b). Although this increased population might logically act to suppress endometrial proliferation, macrophages in these affected women actually stimulate endometriotic tissue (Braun, 1994).

Of other immune system players, NK cells have cytotoxic activity against foreign cells. Although NK cell numbers are unaltered in the peritoneal fluid of affected women, the NK cell cytotoxicity against endometrium is decreased (Ho, 1995; Wilson, 1994).

Cellular immunity may also be disordered in women with endometriosis, and T lymphocytes are implicated. For example, in patients with endometriosis compared with unaffected individuals, total lymphocyte numbers or helper/suppressor subpopulation ratios do not differ in peripheral blood. However, peritoneal fluid lymphocyte numbers are increased (Steele, 1984). Also, the cytotoxic activity of T lymphocytes against autologous endometrium in affected women is impaired (Gleicher, 1984).

Humoral immunity is also altered in affected women and is thought to play a role. Endometrial antibodies of the IgG class are more frequently detected in the sera of women with endometriosis (Odukoya, 1995). One study also identified IgG and IgA autoantibodies against endometrial and ovarian tissues in the sera and in cervical and vaginal secretions of affected women (Mathur, 1982). These results suggest that endometriosis may be, in part, an autoimmune disease.

Cytokines are small, soluble immune factors involved in signaling of other immune cells. Numerous cytokines, especially interleukins, are suspected in endometriosis pathogenesis. Of specific interest, increased levels of interleukin-1β (IL-1β), IL-6, and IL-8 have been identified in relevant tissues and fluids (Arici, 1998; Mori, 1991; Tseng, 1996).

Other cytokines and growth factors are associated with endometriosis establishment. For example, both monocyte chemoattractant protein-1 (MCP-1) and RANTES (regulated on activation, normal T-cell expressed and secreted) can attract monocytes. Levels of these cytokines are increased in the peritoneal fluid of those with endometriosis and positively correlate with disease severity (Arici, 1997; Khorram, 1993). In addition, vascular endothelial growth factor (VEGF) is an angiogenic growth factor, which is upregulated by estradiol in endometrial stromal cells and peritoneal fluid macrophages. Levels of this factor are increased in the peritoneal fluid of affected women (McLaren, 1996). Although the exact role of these cytokines is unclear, perturbations in their expression and activity further support an immunologic role in endometriosis development.

Genetics

No mendelian genetic inheritance pattern has been identified for endometriosis. But, the increased incidence in first-degree relatives suggests a polygenic/multifactorial pattern. For example, in population studies, 4 to 8 percent of the female siblings or mothers of affected women had endometriosis (Dalsgaard, 2013). Other research revealed that women with endometriosis and an affected first-degree relative were more likely to have severe endometriosis (61 percent) than women without an affected first-degree relative (24 percent) (Malinak, 1980). Studies also demonstrate concordance for endometriosis in monozygotic twin pairs (Saha, 2015; Treloar, 1999).

To assist with identifying candidate genes, population-based genome-wide association studies (GWASs) have been performed. These studies are founded on the principle that common diseases, such as endometriosis, are caused by genetic variants that are common themselves. With GWAS, a set of several 100,000 common single nucleotide polymorphisms (SNPs or single DNA base-pair changes) are selected to provide the maximum coverage of the genome. Their frequencies are then compared between affected and unaffected groups. From GWASs of endometriosis, several candidate genes and chromosomes have been identified for further study (Burney, 2013).

The primary method of endometriosis diagnosis is visualization of endometriotic lesions by laparoscopy, with or without biopsy for histologic confirmation. The extent of endometriosis can vary widely between individuals, and thus, one classification by the American Society for Reproductive Medicine (1997) allows disease to be quantified (Fig. 10-2). With this, endometriosis on the peritoneum, ovaries, fallopian tubes, and cul-de-sac is scored at surgery. At these sites, points are assigned for disease surface area, degree of invasion, morphology, and extent of associated adhesions. Also, endometriotic lesions are morphologically categorized as white, red, or black. In this system, endometriosis is classified as stage I (minimal), stage II (mild), stage III (moderate), and stage IV (severe).

Advantages of this system are its widespread implementation, its ease of use, and its four simple-to-comprehend stages. However, the system has limitations. It correlates poorly with infertility and pain symptoms (Guzick, 1997; Vercellini, 1996). For example, women with extensive disease (stage IV) may note few complaints, whereas those with minimal disease (stage I) may have significant pain or sub­fertility or both. This poor predictive ability stems in part from scores that are derived from subjective visual examination. Moreover, disease involving ureter, bowel, or other extrapelvic sites is not scored (Adamson, 2013). To address these shortcomings, other systems have been developed but are yet to be widely used. These include the ENZIAN staging system to better represent DIE and the Endometrial Fertility Index (Adamson, 2010; Haas, 2011).

Pain

As noted, women with endometriosis may be asymptomatic, but chronic pelvic pain (CPP) or subfertility is common (Ballard, 2008). Of endometriosis-associated CPP, dysmenorrhea, dyspareunia, and noncyclic pain are frequent types. Less often and described on page 234, affected women may also complain of dyschezia (pain with defecation), dysuria, or abdominal wall pain.

At the molecular level, the underlying cause of pain is unclear, but proinflammatory cytokines and prostaglandins released by endometriotic implants may be one source (Bulun, 2009). Other investigations implicate nerve growth into endometriotic implants (Barcena de Arellano, 2011; McKinnon, 2012). Once established, continued exposure of these sensory nerves to the inflammatory environment within the implants can lead to central sensitization and CPP, as described in Chapter 11 (As-Sanie, 2013; Bajaj, 2003). The variability of implant location and these chemical influences help explain the differing pain manifestations experienced by women with endometriosis. That said, typical pain scoring tools such as the visual analogue scale and the numerical rating scale are suitable for initial assessment and for evaluation of treatment efficacy (Fig. 11-3) (Bourdel, 2015).

Of pain types, endometriosis-associated dysmenorrhea typically precedes menses by 24 to 48 hours. Compared with primary dysmenorrhea, this pain is thought to be more severe and is less responsive to NSAIDs and combination oral contraceptives (Allen, 2009; Opoku-Anane, 2012). Presence of DIE also positively correlates with dysmenorrhea severity (Lafay Pillet, 2014).

Endometriosis-associated dyspareunia is often related to rectovaginal septum, uterosacral ligament, or posterior cul-de-sac disease, although other involved sites can cause painful intercourse (Vercellini, 2007, 2012). Tension on diseased uterosacral ligaments during intercourse may trigger this pain (Fauconnier, 2002). Although some women with endometriosis describe a history of dyspareunia since coitarche, endometriosis-associated dyspareunia is suspected if pain develops after years of pain-free intercourse (Ferrero, 2005).

Noncyclic chronic pelvic pain (CPP) is another frequent symptom of endometriosis. Approximately 33 percent of women with CPP are found to have endometriosis at the time of laparoscopy (Howard, 2003). This percentage is higher in adolescents with CPP (Janssen, 2013). Some studies correlate pain severity with advanced-stage disease, whereas other studies do not (Fedele, 1992; Hsu, 2011).

The focus of chronic pain may vary. If the rectovaginal septum or uterosacral ligaments are involved with disease, pain may radiate to the rectum or lower back. Alternatively, pain radiating down the leg and causing cyclic sciatica may reflect sciatic nerve involvement (Possover, 2011). That said, pain may correlate poorly with pelvic disease location (Hsu, 2011).

Infertility

The incidence of endometriosis in women with subfertility is 20 to 30 percent (Waller, 1993). In addition, although wide variability is reported, patients with infertility appear to have a greater incidence of endometriosis than fertile controls (13 to 33 percent versus 4 to 8 percent) (D’Hooghe, 2003; Strathy, 1982). Furthermore, Matorras and colleagues (2001) noted an increased prevalence of more severe stages of endometriosis in women with infertility.

Adhesions are one intuitive explanation for endometriosis-related infertility. These may impair normal oocyte pick-up and transport by the fallopian tube. Beyond mechanical impairment, numerous subtle defects also appear to be involved. Such defects include perturbations in follicle development, ovulation, sperm function, embryo quality and development, and implantation (Macer, 2012; Stilley, 2012).

A link between infertility and milder forms of endometriosis is less well supported (D’Hooghe, 1996; Schenken, 1980). An association is suggested by the differing prevalence of endometriosis between infertile and fertile women. For example, Rodriguez-Escudero and associates (1988) reported that women with minimal endometriosis had a 12-month cumulative pregnancy rate of 47 percent, which is below that of normal fertile women. Furthermore, a prospective cohort study demonstrated that women with minimal or mild endometriosis had a fecundity similar to that of those with unexplained infertility.

In moderate to severe endometriosis (stage III to IV), tubal and ovarian architecture are often distorted. As a result, impaired fertility would be expected. Few studies report fecundity rates in women with severe endometriosis. One investigation comparing mild, moderate, and severe endometriosis revealed a monthly fecundity rate of 8.7 percent in those with mild disease, 3.2 percent with moderate disease, and no pregnancies with severe disease (Olive, 1985a). In another, women with severe endometriosis undergoing in vitro fertilization (IVF) had poorer implantation and pregnancy rates compared with those with mild disease (Harb, 2013).

Symptoms from Specific Sites

Rectosigmoid Lesions

Defecatory pain develops much less often than other types of CPP in affected patients. Complaints may be chronic or cyclic, and they can be associated with constipation, diarrhea, or cyclic hematochezia (Roman, 2013). Thus, gastrointestinal causes of CPP are also entertained during evaluation (Chap. 11). The origin of symptoms can be fixation of the rectum to adjacent anatomic structures or rectal wall inflammation.

Symptoms may also stem from DIE of the gastrointestinal tract, which complicates 5 to 12 percent of proven endometriosis cases. Bowel DIE predominantly involves rectosigmoid colon and much less so the small bowel, cecum, or appendix (Ruffo, 2014b). Lesions are usually confined to the subserosa and muscularis propria. Thus, colonoscopy offers poor diagnostic sensitivity (Milone, 2015). Rarely, more severe cases may involve the bowel wall transmurally and lead to intestinal obstruction or a clinical picture suggesting malignancy (Kaufman, 2011; Ruffo, 2014a).

For diagnosis, rectal DIE can be imaged by transvaginal sonography (TVS), and sensitivity approximates 80 percent. However, TVS techniques used to diagnose DIE have a learning curve, and these are predominantly performed at tertiary care centers (Tammaa, 2014). Magnetic resonance (MR) imaging can clarify anatomy and degree of invasion, especially preoperatively (Bazot, 2009; Wall, 2015). Laparoscopy typically provides the definitive diagnosis.

Without obstructing symptoms, women may be considered for conservative management with hormonal therapy. However, treatment is often surgical, and cases often warrant a surgeon skilled in bowel surgery. Variables such as anatomic site, DIE depth, lesion size, and number of foci influence surgery. Colorectal segment resection may be needed. Less invasive techniques that shave down the lesion without opening the rectum or that excise discrete nodules are also described (Alabiso, 2015).

Urinary Tract Lesions

Logically, endometriosis should be considered if urinary tract symptoms persist despite negative urine culture results. Symptoms, if present, are more common with bladder disease. These include dysuria, suprapubic pain, urinary frequency, urgency, and hematuria (Gabriel, 2011; Seracchioli, 2010). Costovertebral angle pain may reflect ureteral endometriosis with obstruction and hydronephrosis that can progress eventually to kidney function loss (Knabben, 2015).

In a large series by Antonelli and coworkers (2006), the prevalence of urinary tract DIE was 2.6 percent. In this series of 31 patients, 12 had bladder endometriosis, 15 had ureteral endometriosis, and four had both ureteral and bladder involvement. TVS has suitable accuracy for bladder DIE but is less sensitive for ureteral disease (Exacoustos, 2014a). In unclear cases, MR imaging can add additional anatomic information. In light of associated symptoms with urinary tract DIE, cystoscopy with biopsy can also help clarify the diagnosis.

Treatment is either medical or surgical. If elected, surgery for bladder invasion is typically partial cystectomy. Surgeries for ureteral involvement vary by disease severity and include: (1) freeing the tethered ureter by ureterolysis, (2) segmental resection and reanastomosis, or (3) ureter reimplantation into the bladder, that is, ureteroneocystotomy (Seracchioli, 2010).

Anterior Abdominal Wall

Some individuals with abdominal pain can have anterior abdominal wall endometriomas. Most of these lesions develop in the abdominal scar after uterine surgery or cesarean delivery, whereas others form unrelated to prior operations (Fig. 10-3) (Ding, 2013). Implants usually are found within the subcutaneous layer, are palpable, and may involve the adjacent fascia. Less often, the rectus abdominis muscle is infiltrated (Mostafa, 2013). Diagnostic tools are variably employed, and abdominal wall sonography, computed tomography (CT), MR imaging, and fine-needle aspiration are options. The decision to perform concurrent TVS is typically guided by whether CPP symptoms coexist.

In most instances, implants are surgically excised for pain relief and diagnosis. For small implants, preoperative imaging may not be needed. But with larger implants and concerns for fascial or rectus abdominis muscle involvement, CT or MR imaging can aid surgery planning (Ecker, 2014). Large fascial defects following excision may require mesh to close the defect.

Thoracic Lesions

Thoracic endometriosis defines implants inside the thoracic cavity that lead to symptoms described as menstrual or synonymously called “catemenial.” These include cyclic chest or shoulder pain, hemoptysis, or pneumothorax, which predominantly occurs on the right (Haga, 2014; Rousset-Jablonski, 2011). Chest CT is preferred imaging (Rousset, 2014). For pneumothorax, minimally invasive thorascopic surgery is usually indicated. This is often coupled with several months of postoperative gonadotropin-releasing hormone (GnRH) or of progestin therapy identical to that for pelvic endometriosis treatment (Alifano, 2010). Hemoptysis, depending on findings, may be treated hormonally or surgically.

Physical Examination

For the most part, endometriosis is a disease confined to the pelvis. Accordingly, visual cues are often lacking. Some exceptions include endometriosis within an episiotomy scar or surgical scar, most often within a Pfannenstiel incision (Koger, 1993; Zhu, 2002). Rarely, endometriosis may develop spontaneously within the perineum or perianal region (Watanabe, 2003). Occasionally, blue or red powder-burn lesions are seen on the cervix or the posterior vaginal fornix. These lesions can be tender or bleed with contact. One study found that speculum examination displayed endometriosis in 14 percent of patients diagnosed with DIE (Chapron, 2002).

During bimanual examination, pelvic organ palpation often reveals suggestive anatomic abnormalities. Uterosacral ligament nodularity and tenderness may reflect active disease or scarring along the ligament. An enlarged, cystic adnexal mass may represent an ovarian endometrioma, which can be mobile or adhered to other pelvic structures. A retroverted, fixed, tender uterus and a firm, fixed posterior cul-de-sac are among other findings. Pelvic nodularities secondary to endometriosis may be more easily detected by bimanual examination during menses (Koninckx, 1996). However, examination is generally inaccurate in assessing the extent of endometriosis, especially if the lesions are extragenital. Last, rectal examination may reveal rectovaginal septum nodularity or tenderness.

Laboratory Testing

Laboratory investigations are often undertaken to exclude other causes of pelvic pain that are listed in Table 11-1. Initially, a complete blood count (CBC), human chorionic gonadotropin assay, urinalysis and urine cultures, vaginal cultures, and cervical swabs may be collected to exclude infections or pregnancy complications. If urinary tract endometriosis is suspected, then renal function can also be assessed by creatinine levels.

Numerous serum markers have been studied as possible diagnostic tools. Cancer antigen 125 (CA125) is a glycoprotein that is found in fallopian tube epithelium, endometrium, endocervix, pleura, and peritoneum. As discussed in Chapter 35, this marker is used in ovarian cancer evaluation and surveillance. Recognized by monoclonal antibody assays, elevated CA125 levels positively correlate with endometriosis severity (Hornstein, 1995a). Unfortunately, the assay has poor sensitivity in detecting mild endometriosis and appears to be a better diagnostic test for stage III or IV endometriosis (Mol, 1998; Santulli, 2015). Although the role of this assay in clinical practice is uncertain, it may be useful in the presence of a sonographically detected ovarian cyst suggestive of an endometrioma.

As for other serum markers, May and colleagues (2010) completed a systematic review of more than 100 putative biomarkers. They were unable to identify a single biomarker or biomarker panel that they felt was clinically useful.

Diagnostic Imaging

Many women with endometriosis present with CPP, and TVS is an initial imaging tool. It is accurate in detecting endometriomas and aids exclusion of other causes of pelvic pain. However, imaging of superficial endometriosis or endometriotic adhesions is inadequate. Small endometriotic plaques or nodules may occasionally be seen, but these findings are inconsistent (Wall, 2015).

Endometriomas can be diagnosed by TVS with adequate sensitivity in most settings if they are 20 mm in diameter or greater. Sensitivity and specificity of TVS to diagnose endometriomas range from 64 to 90 percent and from 22 to 100 percent, respectively (Moore, 2002). An endometrioma classically is cystic with homogeneous, low-level internal echoes, often described as “ground glass” echogenicity. There is normal surrounding ovarian tissue (Fig. 10-4). As such, these may have an identical appearance to hemorrhagic corpus luteum cysts. Although endometriomas are most often unilocular, one to four thin septations can be found (Van Holsbeke, 2010). Less typically, these cysts can display thick septations or walls. Also less often, echogenic wall foci that lack flow when color Doppler is applied can be seen and are typically depositions of blood or blood components (Bhatt, 2006). Color Doppler TVS often demonstrates pericystic, but not intracystic, flow. Although endometriomas can be found in postmenopausal women, they are less common and more often are multilocular compared with those in reproductive-aged women.

As noted previously, TVS for DIE involving the bowel and bladder has suitable accuracy (Exacoustos, 2014a; Hudelist, 2011). That said, for the diagnosis of rectal endometriosis, TVS is highly operator dependent, and experience is often lacking (Dunselman, 2014). Thus, MR imaging may clarify anatomy for equivocal sonographic findings and offers superior resolution at soft tissue interfaces (Fig. 10-5). In some cases with DIE, MR imaging can assist preoperative planning.

CT scanning plays a limited role in the evaluation of endometriosis. This is because TVS images endometriomas well, and CT has poor sensitivity for small implants and plaques. That said, chest CT is preferred for thoracic endometriosis. CT is suitable for abdominal wall endometrioma evaluation. Also, in selected cases, CT may have a role to evaluate bowel or ureteral endometriosis (Exacoustos, 2014b).

Diagnostic Laparoscopy

Although imaging can add clinical information, laparoscopy is the primary method used for diagnosing endometriosis (American College of Obstetricians and Gynecologists, 2014b). Surgical findings vary and may include discrete endometriotic lesions, endometrioma, or adhesions. Implants are typically found on pelvic organ serosa and pelvic peritoneum. Lesions are variably colored and can be red (red, red-pink, or clear), white (white or yellow-brown), and black (black or black-blue) (Fig. 10-6). White and red lesions most commonly correlate with the characteristic histologic findings of endometriosis (Jansen, 1986). Dark lesions are pigmented by hemosiderin deposition from trapped menstrual debris. In addition to color differences, endometriotic lesions may differ morphologically. They can appear as smooth blebs on peritoneal surfaces, as holes or defects within the peritoneum, or as flat stellate lesions whose points are formed by surrounding scar tissue. Endometriotic lesions may be superficial or may deeply invade the peritoneum or pelvic organs.

Endometriomas are easily identified during laparoscopy. Laparoscopic visualization of ovarian endometriomas has a sensitivity and specificity of 97 percent and 95 percent, respectively (Vercellini, 1991). Because of this, ovarian biopsy is rarely required for diagnosis.

Pathologic Analysis

Current guidelines do not require biopsy and histologic evaluation for the diagnosis of endometriosis. However, some suggest that relying solely on laparoscopic findings in the absence of histologic confirmation often results in overdiagnosis (Buck Louis, 2011; Wykes, 2004). Specifically, the greatest discordance between laparoscopic and histologic findings is noted in scarred lesions (Walter, 2001). Histologic diagnosis requires both endometrial glands and stroma found outside the uterine cavity (Fig. 10-7). Additionally, hemosiderin deposition is frequently seen. The gross appearance of endometriotic lesions often suggests certain microscopic findings. For example, if examined microscopically, red lesions are frequently vascularized, whereas white lesions more often display fibrosis and few vessels (Nisolle, 1997).

Therapy for endometriosis depends on a woman’s specific complaints, symptom severity, location of endometriotic lesions, goals for treatment, and desire to conserve future fertility. As shown in Figure 10-8, determining whether a patient is seeking treatment for infertility or pain is essential, as therapy for these two is different.

If pain is prominent and conception is not currently desired, then medical therapy is typically selected. Treatment strives to atrophy ectopic endometrium and diminish disease-associated inflammation. Available agents include NSAIDs, sex steroid hormones, GnRH agents, and aromatase inhibitors. In general, suitable starting regimens are NSAIDS alone or combined with oral contraceptive pills or with a progestin. These agents may be initiated if endometriosis is suspected in a woman with CPP or may be started following diagnostic laparoscopy. If initial therapy fails to control pain following laparoscopy, then use of a different medication is reasonable. If initial empiric therapy is ineffective, then either diagnostic laparoscopy or medication change is suitable (American College of Obstetricians and Gynecologists, 2014b). Of note, although medical treatment improves pain, relapse rates are high with therapy discontinuation.

If infertility is the presenting symptom, then fertility-preserving treatment without ovulation suppression will be required, as outlined on page 243. In contrast, if the patient has severe, recalcitrant pain and has completed childbearing, definitive surgery may be warranted, as described on page 243.

Medical Treatment of Pain

Expectant Management

For many women, symptoms will preclude them from choosing expectant management. However, for those with mild symptoms or for asymptomatic women diagnosed incidentally, expectant management may be appropriate (Moen, 2002). Sutton and associates (1997) expectantly managed patients initially diagnosed by laparoscopy with minimal to moderate endometriosis. At second-look laparoscopy after 1 year, 29 percent of women had disease regression, 42 percent remained unchanged, and 29 percent had disease progression. Other investigators have shown similar rates of disease regression with expectant management (Thomas, 1987). These studies are confined to patients with minimal to moderate endometriosis. There are no well-designed trials examining the effect of expectant management on severe endometriosis.

Nonsteroidal Antiinflammatory Drugs

Both COX-1 and -2 enzymes promote synthesis of prostaglandins involved in the pain and inflammation associated with endometriosis. Specifically, endometriotic tissue expresses COX-2 at greater levels than eutopic endometrium (Cho, 2010). Accordingly, therapy aimed at lowering these prostaglandin levels plays a role in alleviating endometriosis-associated pain. As such, NSAIDs are often first-line therapy in women with primary dysmenorrhea or pelvic pain with suspected or known endometriosis. That said, study evidence supporting NSAIDs for this disease is scant and is extrapolated from efficacy data in primary dysmenorrhea (Kauppila, 1985; Marjoribanks, 2010).

The NSAIDs listed in Table 10-1 nonselectively inhibit both COX-1 and COX-2 enzymes. In contrast, selective COX-2 inhibitors specifically inhibit the COX-2 isoenzyme. Due to the cardiovascular risks with long-term use of selective COX-2 inhibitors, these medications are used at the lowest possible dose and for the shortest duration necessary (Jones, 2005). Thus, drugs in Table 10-1 are primarily selected.

Combination Hormonal Contraceptives

These agents are a mainstay for the treatment of endometriosis-related pain. They inhibit gonadotropin release, decrease menstrual flow, and decidualize implants. As such, abundant study evidence supports use of combination oral contraceptive (COC) pills or the contraceptive patch or ring to relieve endometriosis-related pain (Harada, 2008; Vercellini, 1993, 2010). These provide contraception and other noncontraceptive benefits, which are balanced against risks enumerated in Chapter 5.

COCs can be used conventionally in a cyclic regimen or may be used continuously, without a break for withdrawal menses. The continuous regimen decreases the frequency of painful menses and improves CPP (Guzick, 2011). For endometriosis-related pain, monophasic or multiphasic COCs are both suitable. Additionally, low-dose COCs (containing ≤20 μg ethinyl estradiol) have not proved superior to conventional-dose COCs for endometriosis treatment, but lower doses may lead to higher rates of abnormal uterine bleeding (Gallo, 2013).

Progestins

This family of hormones is often used for endometriosis therapy. Progestational agents are known to antagonize estrogenic effects on the endometrium, causing initial decidualization and subsequent endometrial atrophy. For endometriosis treatment, progestins can be administered as an oral progestin pill, depot medroxyprogesterone acetate (DMPA) (Depo-Provera), norethindrone acetate (NETA), or a levonorgestrel-releasing intrauterine system.

As supporting evidence, one randomized trial compared the effect of oral medroxyprogesterone acetate (MPA) 100 mg daily given for 6 months and placebo. At second-look laparoscopy, partial or total resolution of peritoneal implants was noted in 60 percent of progestin-treated women compared with 18 percent of the placebo group. Furthermore, pelvic pain and defecatory pain were significantly reduced (Telimaa, 1987). Side effects of high-dose MPA included acne, edema, weight gain, and irregular menstrual bleeding. In practice, MPA is prescribed in oral dosages ranging from 20 to 100 mg daily.

Alternatively, MPA may be given intramuscularly in depot form in a dosage of 150 mg every 3 months. In depot form, MPA may delay resumption of normal menses and ovulation and thus is less suitable for women contemplating imminent pregnancy. Subcutaneous formulation of MPA, marketed as Depo-SubQ Provera 104, is also effective (Schlaff, 2006).

As discussed in Chapter 5, the Depo-Provera package insert carries a “black box warning.” This describes that prolonged DMPA use may result in bone density loss, that this loss is greater with increasing duration of use, and that the loss may not be completely reversible. Labeling recommends limiting use to 2 years unless other contraceptive methods are inadequate. Thus, the risks and benefits of treatment are weighed if contemplating long-term DMPA therapy. Bone density surveillance with dual energy x-ray absorptiometry (DEXA) scanning is not recommended (American College of Obstetricians and Gynecologists, 2014a).

NETA is a 19-nortestosterone synthetic progestin that has been used to treat endometriosis. In one study, investigators administered an initial oral dosage of NETA, 5 mg daily, with increases of 2.5 mg daily until amenorrhea or a maximal dosage of 20 mg daily was reached. They found an approximately 90-percent reduction in dysmenorrhea and pelvic pain (Muneyyirci-Delale, 1998). As discussed on page in the next section, NETA is also used as adjunct therapy with GnRH agents to blunt the bone loss linked with those drugs.

Dienogest is another 19-nortestosterone synthetic progestin suitable for endometriosis. In one randomized study, it was significantly more effective than placebo for reducing endometriosis-associated pain when used orally at a dosage of 2 mg daily (Strowitzki, 2010a). Other trials show efficacy equivalent to that of GnRH agonists (Harada, 2009; Strowitzki, 2010b). Currently, this progestin as a sole agent is not available in the United States.

The levonorgestrel-releasing intrauterine system (LNG-IUS) (Mirena) delivers levonorgestrel directly to the endometrium and is effective for up to 5 years. This intrauterine device has traditionally been used for contraception, but data are accruing for endometriosis treatment. One small randomized trial that incorporated second-look laparoscopy showed improved endometriosis stage with both LNG-IUS and the comparator GnRH treatment. Other small randomized trials have shown symptom improvement when compared against expectant management, DMPA, or GnRH agonists (Petta, 2005; Tanmahasamut, 2012; Vercellini, 2003b; Wong, 2010). However, in patients with bowel endometriosis, the LNG-IUS may be ineffective for symptom control (Hinterholzer, 2007). Contraindications to LNG-IUS use are listed in Chapter 5.

Intuitively, the implant that chronically releases the progestin etonogestrel might be considered for endometriosis. Data regarding its efficacy for this indication are limited. One small randomized study comparing the implant and the LNG-IUS reported comparable efficacy (Walch, 2009).

GnRH Agonists

Endogenous pulsatile release of GnRH prompts secretory activity of the gonadotropes within the anterior pituitary. Gonadotropin release from the pituitary then leads to ovarian steroidogenesis and ovulation. However, continuous, nonpulsatile GnRH administration results in pituitary desensitization and subsequent loss of ovarian steroidogenesis. These features allow pharmacologic use of GnRH agonists for endometriosis treatment. With loss of ovarian estradiol production, the hypoestrogenic environment removes the stimulation normally provided to the endometriotic implants and creates a pseudomenopausal state during treatment. In addition to their direct effect on estrogen production, GnRH agonists also reduce COX-2 levels in patients with endometriosis, providing another mechanism of treatment (Kim, 2009).

GnRH agonists are inactive if taken orally, but intramuscular, subcutaneous, and intranasal preparations are available. Leuprolide acetate (Lupron Depot) is available in a 3.75-mg monthly dose or an 11.25-mg 3-month dose, both given IM. Less frequently used GnRH agonists include goserelin (Zoladex) administered as a 3.6-mg monthly or a 10.8-mg 3-month subcutaneous depot implant; triptorelin (Trelstar) given as a 3.75-mg monthly IM injection; and nafarelin (Synarel) used in a 200-mg twice-daily nasal spray regimen. All of these except triptorelin carry specific Food and Drug Administration (FDA) approval for endometriosis treatment.

Pain Improvement

Empirically, GnRH agonists may be used prior to laparoscopy in women with CPP and clinical suspicion of endometriosis. In a study by Ling and associates (1999), after 3 months of GnRH agonist treatment, pain scores significantly declined compared with those after placebo. Subsequent laparoscopy revealed that 93 percent of these women had surgically diagnosed endometriosis. Similarly, in suspected cases, depot leuprolide acetate may be used empirically in lieu of laparoscopy for satisfactory symptom improvement. That said, an empiric GnRH trial is not routinely offered to patients younger than 16 years because GnRH agonist effects on long-term bone mineral density (BMD) have not been adequately studied in this age group.

In those with surgically confirmed endometriosis, numerous studies demonstrate the effectiveness of GnRH agonist therapy to improve pain symptoms (Brown, 2010). The GnRH agonists provide greater relief when administered for 6 months compared with 3 months (Hornstein, 1995b). As noted in the prior sections, GnRH agonists compares favorably with other drugs used for endometriosis treatment. Although GnRH is effective, its cost and side effects often favor trials of COCs or progestins first.

Add-back Therapy

Concerns regarding the effects of prolonged hypoestrogenism preclude extended treatment with GnRH agonists. Hypoestrogenic symptoms include hot flushes, insomnia, reduced libido, vaginal dryness, and headaches. Moreover, both spine and hip BMD decrease at 3 and 6 months of GnRH agonist therapy, with only partial recovery at 12 to 15 months after treatment (Orwoll, 1994). Because of the increased osteoporosis risk, therapy is usually limited to the shortest possible duration—usually no greater than 6 months.

Estrogen may be added to GnRH agonist therapy to counteract bone loss and is termed add-back therapy. With the addition of such hormonal add-back therapy, a GnRH agonist may occasionally be used longer than 6 months (American College of Obstetricians and Gynecologists, 2014b). The goal of add-back therapy is to supply enough estrogen to minimize GnRH agonist side effects while still maintaining a hypoestrogenic state sufficient to suppress endometriosis. Barbieri (1992) explained that tissues have varied sensitivity to estrogen, and a concentration of estrogen that will partially prevent bone loss may not stimulate endometrial growth. This “estrogen threshold” has not been established but is thought to approximate 30 to 40 pg/mL of estradiol.

Several regimens are suitable and appear equally efficacious (Wu, 2014). In one study, NETA 5 mg orally given daily, with or without conjugated equine estrogen (Premarin) 0.625 mg orally daily, for 12 months provided extended pain relief beyond the duration of treatment and preserved BMD (Hornstein, 1998). Another regimen of transdermal estradiol 25 μg plus daily 5 mg oral MPA showed that the GnRH agonist remained effective in reducing endometriosis pain (Edmonds, 1996). In addition, traditional COCs may be used effectively as add-back agents.

The extent of BMD decline has been evaluated with add-back therapy. Although bone loss was noted in all patients undergoing GnRH agonist treatment, the extent of loss was lower in the add-back group (Edmonds, 1994). Quality of life is also improved with add-back therapy (Zupi, 2004).

Such therapy can be initiated either immediately with the GnRH agonist or after 3 to 6 months of agonist therapy. However, little benefit is gained by deferring add-back therapy, and patients who receive add-back concurrently with agonist therapy have reduced bone loss (Al-Azemi, 2009; Kiesel, 1996). Supplemental calcium as a 1000-mg total daily dose is recommended along with add-back regimens (American College of Obstetricians and Gynecologists, 2014b).

GnRH Antagonists

GnRH antagonists are a newer category of GnRH analogues capable of suppressing gonadotropin production. Unlike GnRH agonists, GnRH antagonists do not produce an initial release or flare of gonadotropins. Thus, suppression of gonadotropins and sex steroid hormones is immediate.

GnRH antagonists are mainly used for suppression of premature ovulation during IVF cycles. They have not been well studied for endometriosis treatment. Küpker and colleagues (2002) evaluated the effect of the antagonist cetrorelix in 15 endometriosis patients. They administered subcutaneous injections of cetrorelix at a dosage of 3 mg weekly for 8 weeks. Patients were symptom free during treatment, and second-look laparoscopy revealed disease regression in 60 percent of study participants. That said, long-term depot forms are not currently available.

Of newer agents, a nonpeptide, orally bioactive GnRH antagonist, elagolix, is undergoing evaluation. One 24-week randomized trial showed similar efficacy between elagolix and DMPA for endometriosis-associated pain (Carr, 2014).

Aromatase Inhibitors

As noted earlier, in endometriotic tissue, estrogen may be produced locally through aromatization of circulating androgens. This may clarify postmenopausal endometriosis or may explain cases in which symptoms persist despite conventional treatment. Hormonal strategies described in prior sections target ovarian estrogen production but have little effect on estrogens produced from other sources. In contrast, the aromatase inhibitors (AIs) block aromatase action and estradiol production in both the ovary and extraovarian sites. As a result, estrogen levels are dramatically suppressed, and AIs have hypoestrogenic side-effect profiles similar to those of GnRH agonists (Pavone, 2012). AIs used clinically include anastrozole (Arimidex) and letrozole (Femara).

In addition to hypoestrogenic side effects, a second concern is ovarian cyst formation. As explanation and shown in Figure 20-3, by blocking the conversion of androgens to estrogens in ovarian granulosa cells, AIs reduce the negative feedback at the pituitary–hypothalamus level. This leads to increased GnRH secretion. Resulting elevations in luteinizing hormone (LH) and follicle-stimulating hormone promote increased ovarian follicular development. Therefore, combining AI with a progestins or COCs helps blunt this side effect (Shippen, 2004).

Small studies that combined aromatase inhibitors with NETA or with COCs support this approach for pain relief (Amsterdam, 2005; Ferrero, 2009). However, due to side effects and limited data, such AI combinations are usually prescribed to women after other options for medical or surgical treatment have been exhausted (Dunselman, 2014).

Selective Progesterone-Receptor Modulators

Progestins produce agonist effects upon binding to progesterone receptors. In contrast, progesterone antagonists and selective progesterone-receptor modulators (SPRMs) are agents that vary in their progesterone-receptor binding. Progesterone antagonists universally bind to and inactivate these receptors. SPRMS, depending on their individual pharmacologic profile, may activate or inactive progesterone receptors variably within different tissue types (Chap. 15). Currently, these are not used in the United States to treat endometriosis.

Of progesterone antagonists, mifepristone (RU-486; Mifeprex) is FDA-approved solely for early pregnancy termination. Studied in women with endometriosis, mifepristone reduced pelvic pain and extent of endometriosis (Kettel, 1996). However, as a side effect, its antiprogestational effects expose the endometrium to chronic unopposed estrogen. Resulting endometrial changes range from simple endometrial hyperplasia to a new category described as progesterone-receptor-modulator–associated endometrial changes (PAEC) (Mutter, 2008). The clinical significance of PAEC is still unclear.

Of SPRMs, ulipristal acetate is available in the United States for emergency contraception as Ella and in Europe for presurgical treatment of leiomyomas as Esmya. Again, long-term endometrial safety for both eutopic and ectopic endometria is unclear with this SPRM, and this limits its chronic use at this time. Most other SPRMs are still experimental.

Androgens

These drugs are now used as second-line agents for endometriosis due to their androgenic side effects. Of these, danazol is a synthetic 17α-ethinyl testosterone derivative. Its predominant action suppresses the midcycle LH surge to promote chronic anovulation (Floyd, 1980). Danazol occupies receptor sites on sex hormone-binding globulin (SHBG) and thereby increases serum free testosterone levels. It also binds directly to androgen and progesterone receptors. As a result, danazol creates a hypoestrogenic, hyperandrogenic state that induces endometrial atrophy in endometriotic implants (Fedele, 1990). Regarding efficacy, danazol given orally at dosages of 200 mg three times daily proved superior to placebo to diminish endometriotic implants and pelvic pain symptoms after 6 months of therapy (Telimaa, 1987).

The recommended dosage of danazol is 600 to 800 mg orally daily. Unfortunately, significant androgenic side effects develop and include acne, hot flushes, hirsutism, adverse serum lipid profiles, voice deepening (possibly irreversible), elevation of liver enzyme levels, and mood changes. Moreover, due to possible teratogenicity, this medication should be taken in conjunction with effective contraception. Because of its adverse side-effect profile, danazol is prescribed less frequently, and if administered, its duration is limited.

Gestrinone (ethylnorgestrienone; R2323) is an antiprogestational agent prescribed in Europe for endometriosis. It has antiprogestational, antiestrogenic, and androgenic effects. Gestrinone equals the effectiveness of danazol and of GnRH agonists for relief of endometriosis-related pain (Prentice, 2000). Furthermore, during 6 months of treatment, gestrinone was not associated with the bone density loss commonly seen with GnRH agonist use and was more effective in persistently decreasing moderate to severe pelvic pain (Gestrinone Italian Study Group, 1996). Disadvantageously, gestrinone appears to lower high-density lipoprotein (HDL) levels. Gestrinone is administered orally, 2.5 to 10 mg weekly, given daily or three times weekly.

Surgical Treatment of Endometriosis-Related Pain

Lesion Removal and Adhesiolysis

Because laparoscopy is the primary method for endometriosis diagnosis, surgical treatment at the time of diagnosis is an attractive option. Numerous studies have examined removal of endometriotic lesions, through either excision or ablation. In one randomized trial, diagnostic laparoscopy alone was compared with laparoscopic endometriotic lesion ablation plus uterine nerve ablation. In the ablation group, 63 percent of women attained significant symptom relief compared with 23 percent in the expectant management group (Jones, 2001).

The optimal method to address endometriotic implants for maximal symptom relief is controversial. First, laser ablation does not appear to be more effective than conventional electrosurgical ablation of endometriosis (Blackwell, 1991). Second, ablation and excision both appear to perform suitably. In one randomized trial, ablation was compared with excision of lesions in women with stage I or II endometriosis. At 6 months, similar reductions in pain scores were found (Wright, 2005). Another study showed no significant difference between ablation and excision at 12 months (Healey, 2010). However, at 5 years, the need for further hormonal or analgesic treatment was greater in the ablation group (Healey, 2014). For deeply infiltrative endometriosis, some authors have advocated radical surgical excision, although well-designed trials are lacking (Chapron, 2004).

Unfortunately, recurrence is common following surgical excision. Jones and associates (2001) demonstrated pain recurrence in 74 percent of patients at a mean time of 73 months postoperatively. The median time for recurrence was 20 months. After surgery for pain-related endometriosis, postoperative medical treatment may be elected to extend pain relief or treat residual pain. For this, the most rigorous evidence supports COCs or the LNG-IUS (Somigliana, 2014).

Adhesiolysis is postulated to effectively treat pain symptoms in women with endometriosis by restoring normal anatomy. However, most studies are poorly designed and retrospective. As a result, a definitive link between adhesions and pelvic pain is unclear (Hammoud, 2004). For example, one randomized trial demonstrated no overall pain relief from adhesiolysis compared with expectant management (Peters, 1992). However, within this study, one woman with severe, dense vascularized bowel adhesions experienced pain relief following adhesiolysis.

Adhesion prevention during endometriosis surgery emphasizes sound surgical techniques described in Chapter 40. Of adhesion-prevention agents available in the United States, small studies show lower adhesions reformation rates with use of the cellulose barrier Interceed in endometriosis cases (Mais, 1995a; Sekiba, 1992). But, as noted by the American Society for Reproductive Medicine (2013), although peritoneal instillates and barriers may reduce postoperative adhesions, this has not translated clinically into improved pain, fertility, or bowel obstruction rates.

Endometrioma Resection

Endometriomas are typically treated surgically to exclude malignancy or treat associated pain. To determine the best technique, total ovarian cystectomy compared against aspiration coupled with cyst wall ablation has been studied. Findings note that cystectomy lowers endometrioma recurrence rates and pain symptoms and improves subsequent spontaneous pregnancy rates (Dan, 2013; Hart, 2008). During surgery, ideally normal ovarian tissue is preserved. Toward this goal, electrosurgical coagulation of bleeding sites should be limited. As alternatives, some have described use of dilute vasopressin or suture (Pergialiotis, 2015; Qiong-Zhen, 2014). Other technical steps are described in Chapter 44. Despite cystectomy, endometriomas may recur. Liu and coworkers (2007) found an approximately 15 percent recurrence rate at 2 years following initial surgery.

Importantly, women who undergo endometrioma excision may subsequently have a reduced ovarian reserve, that is, the capacity to provide ova capable of fertilization (Somigliana, 2012). Additionally, surgery increases risks for adhesion formation. Both effects may diminish future fertility. Accordingly, in a woman who is asymptomatic, has a small endometrioma that displays classic findings, has a known endometriosis diagnosis, and has normal or stable CA125 levels, surveillance is an option (American College of Obstetricians and Gynecologists, 2013, 2014b). This approach may benefit asymptomatic women with recurrent endometriomas, as repeat surgery can again diminish reserve (Ferrero, 2015). Following initial diagnosis of an endometrioma, repeat TVS is recommended 6 to 12 weeks later to exclude a hemorrhagic cyst. Endometriomas may then be sonographically surveilled in asymptomatic women yearly or sooner, at the clinician’s discretion (Levine, 2010). The main disadvantage to observation is an inability to exclude ovarian malignancy, and thus patient counseling is essential.

Presacral Neurectomy

For some women, transection of presacral nerves lying within the presacral space may provide relief of chronic pelvic pain (Fig. 38-23). Results from a randomized trial revealed significantly greater pain relief at 12 months postoperatively in women treated with presacral neurectomy (PSN) and endometriotic excision compared with that from endometriotic excision alone (86 percent versus 57 percent) (Zullo, 2003). However, all of these women had midline pain. One metaanalysis demonstrated a significant decrease in pelvic pain after PSN compared with that following more conservative procedures, but only in those with midline pain (Proctor, 2005). Neurectomy may be performed laparoscopically, but it is technically challenging. Due to involved nerve disruption, postoperative constipation and voiding dysfunction are common (Huber, 2015). For these reasons, PSN is used in a limited manner and not recommended routinely for management of endometriosis-related pain.

Laparoscopic Uterosacral Nerve Ablation

There is no evidence that laparoscopic uterosacral nerve ablation (LUNA) is effective in treating endometriosis-related pain (Vercellini, 2003a). In a randomized trial of 487 women with chronic pelvic pain lasting longer than 6 months, with or without minimal endometriosis, LUNA did not improve pain, dysmenorrhea, dyspareunia, or quality-of-life scores compared with laparoscopy without pelvic denervation (Daniels, 2009).

Abdominal versus Laparoscopic Approach

All of the surgical procedures listed above can be completed through open or laparoscopic approaches. First, for benign ovarian masses such as endometriomas, strong evidence supports laparoscopy (Mais, 1995b; Yuen, 1997). Laparoscopic treatment of endometrioma carries an associated 5-percent risk for conversion to laparotomy. However, because of its efficacy and low rates of postoperative morbidity, laparoscopy is a preferred route when feasible (Canis, 2003).

For excision of endometriotic implants, studies also demonstrate effectiveness and low morbidity rates with laparoscopy. Moreover, adhesiolysis is preferred via laparoscopy when safe, and laparoscopy leads to less de novo adhesion formation than laparotomy (Gutt, 2004). Laparoscopic presacral neurectomy appears to be as effective as laparotomy (Nezhat, 1992; Redwine, 1991).

Hysterectomy

This procedure is the definitive and most effective therapy for women with endometriosis who do not wish to retain fertility. It is appropriate for women with intractable pain, adnexal masses, or multiple previous conservative therapies or surgeries (American College of Obstetricians and Gynecologists, 2014b). Hysterectomy for patients with endometriosis may suitably be completed laparoscopically, abdominally, or vaginally. However, adhesions and distorted anatomy secondary to endometriosis often makes a laparoscopic or vaginal approach difficult. In addition, the need to remove ovaries may make a vaginal approach less feasible. Accordingly, the choice of procedure will depend on equipment availability, operator experience, and extent of disease.

Oophorectomy

Prior to hysterectomy for endometriosis, oophorectomy is discussed. A general discussion of risks and benefits is found in Chapter 43. Specific to endometriosis, the benefits of pain relief and reoperation risks are measured against complications of hypoestrogenism. In one study, of those with hysterectomy and bilateral salpingo-oophorectomy (BSO), 10 percent had recurrent chronic pelvic pain and 4 percent required reoperation. Compared with these women, those choosing ovarian conservation had a sixfold greater risk of recurrent pain and an eightfold greater risk of requiring additional surgery (Namnoum, 1995). In a second study, among all those choosing hysterectomy, ovarian conservation doubled the reoperation rate compared with those undergoing BSO (Shakiba, 2008). Moreover, in a subanalysis of those older than 40, ovary conservation lead to a sevenfold greater reoperation rate than BSO. However, in those younger than 40, reoperation rates did not differ whether ovaries were retained or removed. The American College of Obstetricians and Gynecologists (2014b) notes that ovarian conservation can be considered in patients undergoing hysterectomy if ovaries appear normal.

In epidemiologic studies, women with prior endometriosis have slightly increased ovarian cancer rates and higher proportions of clear cell and endometrioid subtypes (Kim, 2014; Pearce, 2012; Somigliana, 2006). That said, consensus guidelines do not recommend management changes in relation to this cancer risk (Dunselman, 2014).

Postoperative Hormone Replacement

Women with surgical menopause are usually younger and would likely benefit from estrogen replacement. Options are discussed in Chapter 22. Although evidence is lacking, some suggest that treatment in these women continue until the time of expected natural menopause.

Unopposed estrogen is appropriate for hypoestrogenic women without a uterus, but disease recurrence has been reported with this therapy in women with severe endometriosis first treated with hysterectomy and BSO (Taylor, 1999). Symptoms required repeat surgery and did not recur with adjuvant combined estrogen and progestin regimens. Additionally, cases of endometrial carcinoma have been reported in women with endometriosis who were treated with unopposed estrogen after hysterectomy and BSO (Reimnitz, 1988; Soliman, 2006). This is rare and may arise from incompletely resected pelvic endometriosis. Therefore, adding a progestin to estrogen replacement therapy can be considered in women with severe endometriosis treated surgically (Moen, 2010). Again, the risks of malignancy are balanced against the adverse lipid changes and breast cancer risks associated with adding progesterone to hormone replacement therapy.

The optimal timing for hormone replacement initiation following hysterectomy with BSO is support by limited data. One small study showed no significant differences in postoperative recurrent pain rates whether hormones were initiated immediately after surgery or were delayed (Hickman, 1998).

Treatment of Endometriosis-Related Infertility

For an asymptomatic woman with infertility, laparoscopy solely to exclude endometriosis is unwarranted (American Society for Reproductive Medicine, 2012). For those with endometriosis-related pain undergoing medical therapy, treatment does not raise fecundity (Hughes, 2007).

For women with infertility and minimal to mild endometriosis, surgical ablation has been suggested to be beneficial, although the effect appears small (Marcoux, 1997). However, other researchers did not report a fertility benefit to surgical ablation for mild to moderate endometriosis (Parazzini, 1999). A metaanalysis of these two studies did demonstrate an advantage for laparoscopic surgery compared with diagnostic laparoscopy (Jacobson, 2010).

Moderate to severe endometriosis may be treated with surgery to restore normal anatomy and tubal function. However, well-designed trials examining the role of surgery for subfertility in women with severe endometriosis are limited (Crosignani, 1996). In infertile women with stage III/IV endometriosis, clinicians can consider operative laparoscopy, instead of expectant management, to increase spontaneous pregnancy rates (Dunselman, 2014). However, after initial unsuccessful surgery for infertility, IVF is preferable to reoperation (Pagidas, 1996).

Alternatively, patients with endometriosis and infertility are candidates for fertility treatments such as controlled ovarian hyperstimulation, intrauterine insemination, and IVF (Chap. 20). Logically, age and disease stage factor into treatment decisions.