Aetiology

The oldest proposal for the formation of endometriosis is Sampson’s theory of retrograde menstruation [5] or, more aptly, retrograde passage and implantation of endometrial tissue. Menstrual material containing viable cells is transported into the peritoneal cavity in a retrograde direction along the fallopian tubes and the refluxed endometrium then implants onto the surface of exposed tissues, principally the peritoneum. The amount of menstrual backflow seems important as higher prevalence rates occur in women with increased menstrual exposure due to (i) obstructed outflow associated with Müllerian anomalies, (ii) short menstrual cycles, (iii) increased duration of bleeding, and (iv) decreased parity [6]. In addition, endometriosis is found more commonly on the left side of the pelvis, thought to be due to the cleft created by the sigmoid colon’s peritoneal reflections. However, Sampson’s theory does not explain all endometriosis since it can occur before menarche, in women with amenorrhoea, despite menstrual suppression, at the umbilicus, and (incredibly rarely) in men. Hence the theory of coelomic metaplasia, the pluripotential of coelomic epithelium to develop not only into the more appropriate normal tissue but also, through a programming defect, into endometriotic tissue. Furthermore, despite the ubiquitous occurrence of retrograde menstruation, which gynaecologists often view at the time of laparoscopy during menses, endometriosis occurs only in a minority of women, and this is explained by the third theory, the contribution of immunologic surveillance defects (which also explains the association of endometriosis with other autoimmune diseases). The expression of factors such as cell adhesion molecules, proteolytic enzymes and cytokines affecting the adherence, implantation and proliferation of tissue within the peritoneal cavity may differ between women, as may clearance of endometrial cells from the pelvis, and altered systemic humoral immunity (altered B‐cell function and antibody production) has also been implicated. It is unclear whether such abnormalities are truly a cause or a result of the disease. Finally, embolic transport of endometrial cells through the blood and lymphatic stream may contribute to the very rare occurrences of endometriosis at sites distant from the pelvis, including lungs, brain and nasal tissues (responsible for the rare case reports, so loved by journal editors, of catamenial pneumothorax and haemoptysis, catamenial epilepsy and catamenial epistaxis).

Risk factors include age, increased peripheral body fat and greater exposure to menstruation (short cycles, long duration of menses and reduced parity), whereas smoking, exercise and oral contraceptive use (current and recent) may be protective [6]. However, there is no evidence that the natural history of the disease can be influenced by controlling these factors. There is clear genetic predisposition: endometriosis is six to nine times more common in the first‐degree relatives of affected women than in controls; and in an analysis of more than 3000 Australian twin pairs, over 50% of the variance of the latent liability to the disease was attributable to additive genetic influences [7]. Stage 3–4 disease has a stronger genetic linkage than stage 1–2 disease [8]. Endometriosis is inherited as a complex genetic trait, similar to diabetes or asthma, meaning that a number of genes interact to confer disease susceptibility, but the phenotype probably only emerges in the presence of environmental risk factors. (Other than the known risk factors outlined above, specific promoting environmental agents have proven elusive. One of the most appealing environmental exposure theories, that the environmental pollutant dioxin might be an underlying cause of the disease [9], was ultimately difficult to confirm.) Genome‐wide association studies (GWAS) [] have confirmed genetic associations and nine genome‐wide significant genetic susceptibility loci have been reported in at least one dataset. However, there is no indication of high‐risk mutations, such as the BRCA mutation in breast cancer. The logarithm of the odds (LOD) score for the genetic loci on chromosome 7 [13] and chromosome 10 [14] is between 3 and 4, compared with the LOD score of over 20 for the BRCA1 gene. The currently recognized nine endometriosis gemetic loci explain only 3.5% of the heritability of endometriosis, compared with the BRCA1 gene which is considered to account for 60–70% of cases of familial breast cancer.

Prevalence

The prevalence is estimated to be 8–10% in women in the reproductive years [6]. Reported rates of endometriosis have long been recognized to be highly variable in different populations of women (Summary box 53.1).

Peritoneal (or typical) endometriosis

Peritoneal endometriosis comprises superficial lesions scattered over the peritoneal, serosal and ovarian surfaces. The appearance has been described as ‘powder‐burn’ or ‘gunshot’ deposits.

Cystic ovarian endometriosis (endometriomas)

Endometriomas develop as cystic lesions within the ovary, classically forming ‘chocolate cysts’ due to the degradation of blood over time to a thick haemosiderin‐rich fluid. Several variants on the implantation and metaplasia theories have been proposed to account for ovarian endometriomas. Thus, it has been suggested that superficial lesions on the ovarian cortex become inverted and invaginated, and that endometriomas are derived from functional ovarian cysts or metaplasia of the coelomic epithelium covering the ovary. Endometriomas have features in common with neoplasia such as clonal proliferation, which is consistent with the endometriosis disease theory. They are statistically associated with subtypes of ovarian malignancy, such as endometrioid and clear‐cell carcinoma. However, it still remains uncertain whether such cancers arise from malignant transformation of benign endometriotic tissue.

Endometriosis‐associated pain symptoms

Severe dysmenorrhoea, deep dyspareunia, chronic pelvic pain, ovulation pain, cyclical or perimenstrual symptoms – often bowel or bladder related, causing dyschezia or dysuria – with or without abnormal bleeding, and chronic fatigue have all been associated with endometriosis. Dyspareunia is the symptom that women often find most distressing and which usually has the biggest negative impact on their quality of life [30]. This is probably because, whilst other symptoms such dysmenorrhoea occur only for a limited time each cycle and are thus more manageable, dyspareunia can be present throughout the cycle and it may also have profound effects on a woman’s relationship with her partner. The predictive value of any single symptom or set of symptoms is limited as each symptom can have other gynaecological or non‐gynaecological causes. Interestingly, the symptom complex that most accurately predicts endometriosis from a woman’s medical history is menstrual dyschezia and a history of benign ovarian cysts [31]. There is little correlation between disease stage and the type, nature and severity of pain symptoms, underscoring that the current classification systems are inadequate for women with pain related to endometriosis. It has long been accepted that typical lesions can cause moderate pain (although around half of all women with such lesions are pain‐free), ovarian endometriomas tend to be associated with more severe pain (even though 10–20% of women are pain‐free), and deep endometriosis can be associated with very severe pain (although, again, women sometimes have no pain at all). The association of other endometriosis phenotypes, including subtle lesions, with pain is less clear. The suggested causes for endometriosis‐related pain include tissue damage and distorted anatomy, peritoneal inflammation, activation of nociceptors, and nerve irritation/invasion in deep endometriosis. If pain is persistent, it may become chronic and, through central sensitization, develop the hallmarks of a chronic pain syndrome [32]. Pain symptoms are usually assessed in clinical trials using a four‐point verbal rating scale for three symptoms (dysmenorrhoea, dyspareunia and pelvic pain) and two examination signs (pelvic tenderness and induration). More recently, health‐related quality of life has also been evaluated, as traditional outcome measures may not adequately assess what the patient considers important. The most useful patient‐generated, disease‐specific tool is the Endometriosis Health Profile (EHP)‐30, a 30‐item questionnaire that covers five dimensions: pain, control and powerlessness, emotional well‐being, social support and self‐image [33].

Endometriosis‐associated subfertility

Endometriosis is associated with subfertility. The infertility associated with the anatomical distortion of fallopian tubes and ovaries, ovarian damage through endometrioma formation and tubal damage consequent on endometriosis is obvious. Most women do not have such severe endometriosis, but there is even an association between minimal/mild endometriosis and subfertility, with cohorts showing that fecundity is approximately halved and time to pregnancy doubled for women with these so‐called milder forms of endometriosis [34,35]. Even women undergoing medically assisted reproduction, including in vitro fertilization (IVF), have poorer outcomes if they have endometriosis [36] or adenomyosis [37]. Numerous mechanisms have been proposed, including negative effects on ovarian reserve, ovulation dysfunction, sperm survival in the female genital tract and egg quality (and thus embryo quality). Potentially subtle effects on egg quality have variously been ascribed to abnormal folliculogenesis: the follicular fluid of women with endometriosis has been found to have increased levels of interleukin (IL)‐6 and progesterone, and decreased levels of cortisol and insulin‐like growth factor binding protein (IGFBP)‐1; and granulosa cells express increased levels of tumour necrosis factor (TNF)‐α, soluble Fas ligand and corresponding apoptotic activity. Intraperitoneal inflammation is also a feature of endometriosis, including evidence of increased phagocytosis of sperm by peritoneal macrophages, as well as inflammatory, proteolytic and angiogenic activity of the peritoneum and peritoneal fluid, involving IL‐6, other growth factors (including vascular endothelial growth factors), cytokines and haptoglobins. Other proposed mechanisms have included reduced sexual frequency due to dyspareunia, luteinized unruptured follicle syndrome, luteal insufficiency and recurrent miscarriage. Whilst traditionally the biggest negative impact of endometriosis on a woman’s fertility was thought to be mediated through egg quality, it is becoming more apparent that endometrial dysfunction, with the possibility of reduced endometrial receptivity, also seems likely. Women with endometriosis have reduced expression of integrins, the key receptivity molecules, in their endometrium at the time of the implantation window; they also overexpress osteopontin, a ligand that binds integrins.

Diagnostic delay

The average delay of around a decade between symptom onset and a definitive diagnosis is well recognized. The key to avoiding diagnostic delay is improved education and awareness. This education and a willingness to consider endometriosis as a diagnosis extends not just to girls and young women and their families, but to health professionals in primary care as well as gynaecologists. In the past, even the gynaecological community has been guilty of paternalistic attitudes regarding menstrual health and endometriosis. While a definitive laparoscopic diagnosis carries great importance for many women, many others would prefer to avoid a laparoscopic procedure if possible. Although a diagnosis is important, it is not a key end‐point and can be viewed as an interim stage for a woman to regain wellness. What is crucial is that a possible diagnosis of endometriosis is considered at an early stage and that the woman (or adolescent) is offered appropriate management with that possibility in mind.

History and clinical examination

Making a diagnosis on the basis of symptoms alone is difficult as the presentation is so variable and other conditions such as irritable bowel syndrome, pelvic inflammation and pelvic congestion syndrome mimic endometriosis. Eliciting pelvic tenderness, a fixed retroverted uterus, tender uterosacral ligaments or enlarged ovaries on examination is suggestive of endometriosis, although examination findings can be normal. The diagnosis is likely if nodules are palpable in the uterosacral ligaments and pouch of Douglas, and is confirmed if lesions (which can be biopsied) are visualized on vaginal speculum examination.

Laparoscopy

Laparoscopic visualization of endometriotic lesions has long been held as the gold standard for diagnostic purposes, and this remains the case. However, Wykes et al. [38] showed in a systematic review including four studies with a total of 433 participants that, when compared with histological evaluation of visualized lesions, laparoscopic visualization alone has limited accuracy (94% sensitivity and 79% specificity). Histological confirmation of lesions seen at laparoscopy is therefore ideal, and mandatory if deep endometriosis or an endometrioma is present. The entire pelvis should be inspected systematically, and the findings documented in detail, preferably with the aid of standardized laparoscopic photographs or a video recording [39]. Whilst it has long been considered best practice to surgically remove endometriosis at the time of diagnostic laparoscopy (provided that adequate consent has been obtained and the surgeon’s expertise is sufficient to deal with the extent of endometriosis diagnosed), Vercellini et al. [40] have made a yet stronger call to avoid ‘low value care’ (meaning interventions with uncertain benefits and/or defined harms, or whose effectiveness is comparable with less expensive alternatives). Hence they have argued that the concept of ‘diagnostic laparoscopy’ should disappear, with laparoscopy being reserved for those women likely to benefit from laparoscopic removal of their endometriosis.

Low‐invasive tests

A low‐invasive diagnostic test for endometriosis, through imaging or biomarkers (in urine, blood, endometrium or other body fluids or tissues) or some combination of these, has long been sought. This is partly related to the recognition that not all women who might have endometriosis will have a laparoscopy. In low‐resource settings, access to laparoscopy may not be feasible. Even in developed countries, the availability of gynaecologists in relation to the number of women with endometriosis means that not all women with endometriosis can have a laparoscopic procedure, while some women elect to avoid laparoscopy. The concept of avoiding low‐value care is inextricably interlinked to the aspiration of an accurate and reliable method of diagnosing endometriosis non‐surgically.

Until recently, there was a dogma that laparoscopy was the only acceptable accurate method of diagnosing endometriosis and that imaging and biomarker tests were insufficiently accurate. In the case of many low‐invasive tests this remains true, but we now have a comprehensive set of systematic reviews of low‐invasive diagnostic tests for endometriosis [41–45]. Summary box 53.2 summarizes the evidence regarding the potential value of the low‐invasive tests. A low‐invasive diagnostic test might be considered suitable as a replacement test for laparoscopy if it equates to the accuracy attained by laparoscopic visualization itself (≥94% sensitivity and ≥79% specificity). Other tests might have utility as triage tests: those with very high sensitivity which, if negative, rule out endometriosis (so‐called SnOUT triage tests); or those with high specificity that, if positive, can raise the suspicion of, or rule in, endometriosis (so‐called SpIN triage tests).

Imaging tests have shown promise in diagnosing certain subtypes of endometriosis, such as endometriomas and deep endometriosis, and in mapping deep endometriosis to various sites (Summary box 53.2), although do not prove so accurate in making the actual diagnosis of endometriosis. Regarding biomarkers (measured in urine, blood and endometrial tissue), the accuracy of any one test has not been found to be sufficiently accurate to be a reasonable replacement, or even triage, diagnostic test. This is also true for the most widely used biomarker to date, serum CA125. The detection of endometrial nerve fibre antigens (protein gene product, PGP‐9.5) showed promise as a replacement test in most studies but could not be reliably detected in all settings.