Pathophysiology of Heavy Menstrual Bleeding

Abstract

Heavy menstrual bleeding (HMB) is a common gynecological complaint with multiple etiologies and diverse pathophysiological origins. This review discusses HMB with reference to the recently proposed PALM-COEIN classification system for abnormal uterine bleeding, initially describing the endometrial events in normal menstruation followed by discussion of the perturbations of normal endometrial shedding that can result in HMB. Our present understanding of the mechanisms of menstrual bleeding as well as many of the pathological aberrations of HMB is incomplete. Further research into the pathophysiology of HMB is urgently needed, as clear knowledge of the mechanisms of this disorder will provide new therapeutic targets to formulate more effective treatments.

In the developed world an average woman undergoes ∼400 repetitive cycles of monthly menstrual bleeding with shedding of her superficial endometrial functional layer. In as many as 20–30% of women, this bleeding is excessive and is termed heavy menstrual bleeding (HMB) [1]. Apart from the physical symptoms of anemia (fatigue, lethargy and exertional dyspnea), HMB can interfere with normal daily life and may affect the social and emotional well being of women, reducing their productivity in society. HMB commonly presents to primary and secondary healthcare providers, with over 1.5 million women suffering from this problem in England and Wales alone [1]. It is the fourth most common reason for secondary gynecological referrals and each year over 30,000 women in England and Wales undergo surgical treatment for HMB [1]. The annual cost to the National Health Service in 2000 was estimated to be over GB£65 million [2] and with the spiraling cost of healthcare provision, it is clear that the current cost of HMB to women, society and the National Health Service is huge.

In this review HMB will be discussed with reference to the recently proposed polyps; adenomyosis; leiomyoma; malignancy and hyperplasia; coagulopathy; ovulatory dysfunction; endometrial; iatrogenic and not yet classified (PALM-COEIN) classification system for abnormal uterine bleeding (AUB), approved by the International Federation of Gynecology and Obstetrics [3]. HMB can present as a chronic disorder of over 12 months duration or as an acute increase in bleeding over a short time period. The age at presentation can range from adolescence to the perimenopausal phase and HMB may occur in the context of either ovulatory or anovulatory cycles. This review will briefly describe the endometrial events in normal menstruation followed by discussion of the perturbations of normal endometrial shedding that can result in HMB.

Normal menstruation

The purpose of menstruation is unknown but it is a phenomenon confined mainly to the upper order primates and humans [4], suggesting an associated evolutionary advantage. The morphology of human endometrium, relatively greater quantity of menstrual shedding/bleeding, larger uterine size relative to adult female body size and the design of the endometrial microvasculature suggest that women have higher resource allocation to endometrial function than many other primates. The average menstrual cycle length in women is 28 days and most women have bleeding for approximately 4–5 days associated with shedding of the superficial stratum functionalis of the endometrium. The endometrium is under the regulation of ovarian steroid hormones, mainly estrogen and progesterone and their involvement in the monthly endometrial cycle is well established. These ovarian sex steroids exert their effects on the endometrium via the respective steroid hormone receptors [5]. Ovarian androgens and glucocorticoids of adrenal origin may also play a role in this process, although the actions of androgens on the endometrium are yet to be elucidated [4].

After menstrual shedding every cell type of the stratum functionalis (epithelial, stromal and vascular), under the influence of estrogen, is regenerated from the cells of the remaining stratum basalis [6]. Estrogen, via estrogen receptor (ER), increases the endometrial responsiveness to all ovarian steroid hormones by upregulating ER, progesterone receptor (PR) and androgen receptor (AR) the net result being cell proliferation with consequent rebuilding and increasing thickness of the endometrial functionalis layer [5]. After ovulation, as the production of progesterone by the corpus luteum increases, histologically the endometrium assumes secretory phase changes and undergoes differentiation. Progesterone, acting via PR, reduces endometrial ER, PR and AR expression, and generally counteracts the estrogen driven mitotic activity of endometrial cells. As well as secretory changes in the endometrial glands, the progesterone dominant effects of the endometrium include predecidualization of the endometrial stroma which prepares the endometrium for implantation of the embryo, although decidualization of endometrial stroma is seen even without embryo implantation. During the secretory phase of the menstrual cycle the spiral arterioles within the stratum functionalis grow and acquire muscle and spiral arterioles are easily identified histologically by the end of the secretory phase.

As the corpus luteum demises at the end of a cycle where conception has not occurred, the stratum functionalis is shed in response to the plummeting progesterone levels. This withdrawal of progesterone with luteal regression has been postulated to be the trigger for the initiation of menstrual bleeding [4]; with the induction of stromal shrinkage and spiral arteriolar vasoconstriction with relative hypoxia in the functionalis layer followed by the instigation of acute inflammatory changes including the influx of large amount of leucocytes and immune cells [7,8], launch of the inflammatory cascade (increase in proinflammatory cytokines, prostaglandins and destructive enzymes of the extracellular matrix such as MMPs exclusively in the endometrial stratum functionalis) [7] and the activation of stem/progenitor cells that are postulated to be resident in the endometrial stratum basalis [6]. The shedding of the ‘old’ stratum functionalis usually happens over 1–2 days yet the menstrual bleeding continues during the proliferation and repair of the surface epithelium of the damaged stratum functionalis which takes several days. The endometrial hemostasis which is responsible for cessation of menstrual bleeding involves platelet aggregation, fibrin deposition and thrombus formation [9].

Studies of myometrium have traditionally been limited to the pregnant uterus and labor [10], although myometrial contractions felt as menstrual period associated ‘cramps’ are a symptom that is commonly reported by women during menstruation. This dramatic increase in the amplitude (labor like, and increased from a base line of <30 mmHg up to 50–200 mmHg) of myometrial contractility, particularly seen in the inner subendometrial zone of the myometrium during menstruation, is likely to be induced by increased endometrial production of prostaglandins such as PGF_2α and PGE₂ [11,12]. These contractions may play a role in emptying the uterine cavity of menstrual debris and may also play a haemostatic role by providing a pressure effect on the endometrial vasculature. Although the pivotal role played by myometrial contractility in preventing uterine bleeding following parturition is well established [13], very little is known of the contribution of myometrial contractility in regulation of menstrual bleeding.

After menstrual shedding, the subsequent repair process has been traditionally presumed to be due to the action of estrogen [4–5,14]. However, the very early repair process is not necessarily dependent on estrogen, as the level of estrogen is at a nadir when the re-epithelialization of the endometrial surface occurs; a process that is not impaired in the postmenopausal hypoestrogenic state [5,15]. The activity of stem progenitor cells, either solely from the resident endometrial stem progenitor cell pool or from the circulating hematopoietic precursor pool, is involved in this regeneration process [6,16]. Therefore, a series of complex, orchestrated interactions between endocrine, paracrine, immunological and hemostatic factors on the endometrium results in normal menstrual bleeding and aberrations of each of these characteristics may result in HMB.

HMB

Traditionally HMB is defined as bleeding in excess of 80 ml per menstrual cycle when measured objectively [17], although among women who present with the complaint of significant HMB, fewer than 50% have been objectively shown to have >80 ml of menstrual loss [18]. HMB is a relatively uncommon complaint in younger women, yet has a prevalence of one in three women in the perimenopausal period [19]. The pathophysiology of HMB may be discussed in the context of both ovulatory and anovulatory cycles. Ovulatory HMB is heavy, regular menstrual bleeding occurring between 21 and 32 days, whereas in anovulatory cycles heavy, often prolonged bleeding occurs 35 or more days apart [4]. Adolescents and young women presenting with HMB dating from their menarche are termed as suffering from primary HMB, whereas relatively acute onset HMB occurring in later reproductive life is referred to as secondary HMB [20].

According to the Federation of Gynecology and Obstetrics classification of AUB, nine categories are listed according to the acronym PALM-COEIN [3]. The PALM group consists of structural abnormalities that can be visualized using imaging techniques or diagnosed by histopathology; whereas nonstructural disorders that cannot be imaged or diagnosed with histopathology are included in the COEIN group.

Mechanisms of HMB

Any process that interferes with the normal endocrine, paracrine or hemostatic functions of the endometrium as well as possibly any interference with myometrial contractility may cause HMB.

Polyps

Polyps are defined as abnormal outgrowth of hypertrophied tissue and in the endometrial cavity can be either endometrial or myometrial in origin [21]. Myometrial ‘polyps’ are submucosal leiomyomas and will be discussed in the next section. Endometrial polyps are common outgrowths of endometrial lining consisting of a monoclonal overgrowth of endometrial stromal cells with inclusion of a non-neoplastic glandular component [22]. They may be single or multiple and one study reported endometrial polyps in 28% of women, with the highest incidence in the fifth decade [21]. The size can vary from a small rounded protruberance within the endometrial lining to a large broad based or pedunculated lesion that can fill the uterine cavity. Endometrial polyps are usually benign but careful histological examination is required to exclude the possibility of focal atypical hyperplasia or even adenocarcinoma, although these findings are more common in postmenopausal women [21,23]. Diagnosis of an endometrial polyp may be difficult in a fragmented specimen but the presence of thick walled vessels, altered fibrous or collagenous stroma and irregular glands will be helpful. Endometrial polyps respond variably to estrogen and progesterone and lack the cyclical changes seen in the adjacent endometrium. They may present with the symptom of an irregular and intermenstrual HMB pattern. Up to 2% of endometrial polyps may have a premalignant or malignant potential in premenopasal women and according to a recent systematic review [23] this risk is increased when the patient has presented with bleeding irregularities.

Endometrial stimulation by estrogen is postulated as the main driving force for endometrial polyp formation and this is supported by the observation that use of tamoxifen, which acts as an ER agonist on the endometrium, increases the risk of endometrial polyps [24,25]. In addition, genetic mutations and the overexpression of endometrial aromatase, which increases the local estrogenic signal, are also thought to be involved in the formation of endometrial polyps [24,25]. Small (<1 cm) benign endometrial polyps often regress but larger lesions are likely to persist [23,24].

Endometrial polyps are the commonest endometrial pathology associated with tamoxifen treatment, with an incidence rate of 8–36% [26]; this frequency is much higher compared with the incidence of 0–11% which is seen in untreated women [27]. Compared with women not taking tamoxifen, these polyps are likely to be larger, more fibrotic and to more frequently show mucinous metaplasia. Histology may also identify hyperplasia with an incidence of malignant transformation of 3–10.7% in tamoxifen-related endometrial polyps, a much higher rate than in healthy controls [26].

The exact mechanism whereby endometrial polyps cause an increase in menstrual loss is not fully understood. An abnormal microvasculature is a feature of endometrial polyps with the presence of thick muscular walled blood vessels being an important diagnostic feature on histopathological examination and this may provide some explanation. Incomplete shedding of the endometrial covering has been described in endometrial polyps and this may also potentially contribute to HMB.

Myometrial dysfunctions (adenomyosis & leiomyoma)

Leiomyomas

Leiomyomas, commonly referred to as ‘fibroids’ are benign neoplasms of the myometrium and are the commonest tumor in women of reproductive age. As many are asymptomatic, the prevalence is difficult to determine. However, after histopathological examination of hysterectomy samples, leiomyomas were identified in 77% of uteri [28]. Leiomyomas are monoclonal arising from a single cell. Various nonrandom chromosomal abnormalities have been reported, including deletion of portions of 7q, trisomy 12 and rearrangements of 12q15, 6p21 or 10q22 [29]. There is a familial predisposition and having a first-degree relative with fibroids increases a woman's risk at least two-fold [30,31].

The prevalence of leiomyomas reduces after the menopause and they are not detected in preadolescent girls. The incidence increases in reproductive years and peaks in the fifth decade. Obesity is a risk factor, while increased parity and cigarette smoking are protective [28]. Compared with normal myometrium, leiomyomas have a higher concentration of ER, PR and aromatase [32] and they grow in response to estrogen and progesterone.

The majority of leiomyomas do not cause clinical symptoms and even when symptoms are present it is sometimes uncertain whether they are due to the fibroids [33]. The reason why many fibroids do not cause symptoms is not known. It is estimated that 20–50% women with fibroids will have symptoms that can directly be attributed to the leiomyoma(s) [34]. Various symptoms have been reported, including AUB, pelvic pressure (also causing urinary incontinence, dysuria, urgency and other bladder symptoms), bloating, congestion, heaviness, dyspareunia and constipation. The commonest symptom is HMB. An ultrasound study reported that premenopausal women with AUB had a higher incidence of submucosal leiomyomas (21 vs 1%) and intramural leiomyomas (58 vs 13%) compared with asymptomatic women [35]. A more recent US based study reported that women with leiomyomas were more likely to report excessive bleeding (46%) than those with no leiomyoma (28%). Risk of bleeding increased with size and bleeding symptoms were similar in those with submucosal and nonsubmucosal leiomyomas [36].

Several theories have been proposed to account for the HMB that is seen in association with leiomyomas. These include an increase in the uterine surface area, increased vascularity and vascular flow into the uterus, reduction in myometrial contractility particularly of the inner junctional zone, endometrial ulceration over a submucosal leiomyoma and compression of the venous plexus within the myometrium leading to congestion of myometrium and endometrium [11]. There is limited objective evidence for many of these suggested pathogenetic mechanisms. Location has been suggested to influence symptoms, with submucosal leiomyomas having a greater association with HMB, although objective evidence for this is limited; there is no consistent relationship between the size and location of fibroids and HMB [37,38].

Early studies revealed that the myometrium in leiomyomatous uteri showed an increase in the venous plexus, especially at the edge of the leiomyomas [39]. Early studies also noted an increased vascular supply to leiomyomas; fibroids are very vascular, containing thick walled muscular vessels. This dysregulation of normal vascular function in leiomyomas may be attributed to abnormalities in expression of angiogenic growth factors and their receptors [39,40]. In a recent study, blood flow and angiogenic gene expression was investigated in fibroid, perifibroid and distant myometrium. Blood flow in tissue around leiomyomas was higher than within the leiomyoma, although there was heterogeneity. There was no difference in angiogenic gene expression between perifibroid and distant myometrium but expression of nine angiogenesis related genes differed significantly between fibroid and distant myometrium and two genes were significantly different between fibroid and perifibroid myometrium. However, there was no correlation between blood flow and any clinical or molecular parameter [41].

The presence of leiomyomas may also affect the composition of the overlying endometrium. When endometrial leucocytes in endometrium near to leiomyomas was compared with distant endometrium, the number of uterine natural killer cells was reduced in endometrium overlying leiomyomas during the mid and late secretory phases, while macrophages were increased in the same area throughout the menstrual cycle [42]. Both uterine natural killer cells and macrophages are potential producers of angiogenic growth factors [43,44] which may influence vessels within the endometrium overlying leiomyomas.

Despite advances in understanding of the molecular changes in leiomyomas and associated myometrium and endometrium, it remains unclear why clinical symptoms are so varied. HMB is seen in some women with leiomyomas but it does not correlate clearly with the size or location of fibroids or with the expression of angiogenic growth factors or blood flow.

Adenomyosis

Adenomyosis is a common disorder that is characterized by the presence of endometrial glands and stroma within the myometrium, usually surrounded by hypertrophied myometrial smooth muscle [45–47]. Until recently the diagnosis depended on examination of hysterectomy specimens and the reported prevalence varied from 5 to 70%; this variation is partly related to the number of histological sections examined, with the incidence of adenomyotic foci increasing as more sections are examined in hysterectomy specimens [45]. Variations in prevalence may also be explained by inconsistent histological definitions of adenomyosis, with some specifying a minimum depth based on myometrial thickness, while others make the diagnosis based on microscopic fields beneath the endometrial–myometrial interface, with variations from one high power field to two low power fields [46,47]. Diagnosis can now also be made using imaging techniques such as MRI or transvaginal ultrasound. A recent ultrasound study of 985 symptomatic women attending a gynecology clinic noted adenomyosis in 20.9%, with good correlation between ultrasound and histological diagnosis in the 45 women who subsequently had a hysterectomy [48]. Adenomyosis is commonest in women of late reproductive age and the majority of women with this condition are parous [48].

Approximately one-third of women with adenomyosis are asymptomatic. The commonest symptom is HMB, with other symptoms and signs including dysmenorrhoea, an enlarged tender uterus, dyspareunia and metrorrhagia [45,46]. A study of hysterectomy specimens has suggested that the extent and spread of adenomyosis may relate to the clinical symptoms: the degree of penetration of adenomyosis into the myometrium did not affect symptoms but spread of adenomyosis correlated with both pelvic pain and dysmenorrhea but not with HMB or dyspareunia [49]. The symptoms reported by women with adenomyosis are nonspecific and as there is often co-existent uterine pathology such as leiomyomas, it is uncertain whether the symptoms can be solely attributed to the adenomyosis.

The cause of HMB in adenomyosis is unknown. In common with leiomyoma, adenomyosis can affect normal myometrial contractility and this may contribute to the HMB [11]. Although diagnosis of superficial adenomyosis (within one low power field of the stratum basalis) is disputed, this has been reported to have an increased association with HMB compared with deeper adenomyosis [50]. Superficial adenomyosis with associated myometrial hypertrophy may lead to compression of the overlying endometrium, resulting in heavy bleeding similar to the effects of fibroids causing HMB [51]. The proportion of endometrial glands within adenomyosis has also been reported to be associated with HMB [46,50]. In contrast, a further study that related HMB to increasing depth of myometrial penetration did not find any association of HMB and superficial adenomyosis or with glandular density within foci [46,50,52]. Adenomyosis is often associated with other uterine pathology, including endometrial polyps, leiomyomas and hyperplasia and these may also be associated with HMB [51].

Although the majority of women with adenomyosis are parous, adenomyosis may be associated in some cases with infertility. In a group of women who had bowel resection for endometriosis, women who did not conceive spontaneously or after IVF were significantly more likely to have adenomyosis [53]. The explanation is uncertain and has been related to abnormal uterotubal transport in diffuse adenomyosis [54]. However, endometrial leucocyte populations are also altered in women with adenomyosis, with reports of increased endometrial macrophages and uterine natural killer (NK) cells in luteal phase endometrium of women with recurrent implantation failure and adenomyosis diagnosed by pelvic MRI [55].

The pathogenesis of adenomyosis is uncertain and various explanations have been suggested. The most popular hypothesis is that endometrium invaginates into myometrium from the stratum basalis during periods of regeneration and healing. This could occur due to trauma such as pelvic surgery or after repeated sharp curettage following failed pregnancy that disrupts the endomyometrial border [56]. Adenomyosis is influenced by steroid hormones, although reports of expression of ER, PR and AR vary. It has been reported that foci of adenomyosis express higher levels of ER than the corresponding eutopic endometrium [51], suggesting a high estrogen responsiveness in adenomyotic lesions resulting in mitosis. However, there are other reports suggesting that the expression of ER in adenomyosis is lower than in adjacent endometrium [57]. Adenomyosis tissue also express aromatase, leading to local production of estrogen that may contribute to further growth stimulation [45].

An alternative suggestion is that adenomyosis develops within the myometrium from Mullerian remnants. This is supported by studies of eutopic endometrium and adenomyomatous endometrium which show distinct differences. Compared with eutopic endometrium, the endometrium within foci of adenomyosis did not respond to hormone changes, rarely showed secretory changes, did not exhibit cyclical changes in the apoptosis regulatory protein Bcl-2 and showed altered expression of cytokines and growth factors [58]. More recently, it has been proposed that adenomyosis arises from bone marrow derived stem cells that are displaced through the vasculature or from stem cells within the stratum basalis of endometrium. Although still largely speculative, this has also been proposed as a pathogenetic mechanism in endometriosis [59,60].

It is not known why some women develop adenomyosis while others do not. There is a familial predisposition and several studies have investigated genetic abnormalities in adenomyosis. Various chromosomal and genetic abnormalities have been reported in adenomyosis but results have been variable and studies are limited compared with those for endometriosis [61]. There is an association between adenomyosis and endometriosis; in an MRI study of women with deeply infiltrating endometriosis, an irregular junctional zone was observed in 39.9% of women with endometriosis compared with 22.5% in the reference group [62]. Endometrium in adenomyosis may also display evidence of increased invasiveness; stromal cells from adenomyosis exhibit greater invasiveness compared with normal stromal cells when grown on collagen or in co-culture with myocytes from both normal uteri and those affected by adenomyosis [62,63].

Despite its prevalence in women of reproductive age, our understanding of adenomyosis remains incomplete. The pathogenesis remains unclear but the detection of stem cells in endometrium has opened up new possibilities. Many cases are asymptomatic and there are conflicting data regarding the relationship of depth and extent of adenomyosis to clinical symptoms, including HMB. Recent research points to distinct endometrial abnormalities in adenomyosis, including potentially altered local immune responses, increased capacity for invasion and altered expression of steroid hormone receptors.

Malignancy & hyperplasia

Estrogen, via its cognate receptors ERα (ESR1) and ERβ (ESR2), signals normal endometrial epithelial and stromal cell proliferation and hence when present in excess estrogen can promote endometrial tumorigenesis [5]. Conversely progesterone is the natural tumor suppressor for endometrium, counteracting almost all the pro-proliferative actions of estrogen, arresting the cell cycle, inhibiting inflammation [5]; and promoting differentiation and apoptosis of the glandular epithelium. Therefore any condition that alters the balance between estrogen and progesterone, either factors causing excessive, prolonged estrogenic stimulation or preventing the counteracting effects of progesterone will promote excessive endometrial growth with potential development of endometrial hyperplasia (EH) and carcinogenesis [4,5].

EH is histologically defined as the abnormal overgrowth of endometrial glands in relation to the endometrial stroma [64]. Several classification schemes have been proposed and reproducibility is variable [64]. The WHO (2003) classification includes six categories: benign cycling endometrium, simple hyperplasia without cytological atypia, complex hyperplasia without cytological atypia, simple hyperplasia with cytological atypia, complex hyperplasia with cytological atypia and carcinoma. This classification led to confusion and diagnostic difficulty and interobserver agreement was poor, leading to other classifications being proposed, although interobserver agreement remained relatively low [64]. More recently the WHO has simplified the classification of EH to hyperplasia without atypia (including simple and complex hyperplasia) and atypical hyperplasia/endometrioid intraepithelial neoplasia [65]. These categories are based on the recognition that hyperplasia without cytological atypia is not associated with genetic changes, whereas hyperplasia with cytological atypia is a known precursor of endometrial cancer (EC) and exhibits many of the mutations typical of invasive endometrioid adenocarcinoma [64]. Unopposed estrogen stimulation, usually associated with anovulation or occasional ovulation in premenopausal women, is a common cause of EH which may be seen in 20% of women with polycystic ovarian syndrome (PCOS) with oligomenorrhea, and when present with cytological atypia may carry approximately a 30% risk of developing into or co-existing with EC [66]. Further risk factors include obesity, nulliparity, infertility, unopposed estrogen therapy, selective ER modulators, diabetes and Lynch syndrome [64]. The mechanisms by which EH induces HMB are not fully understood. The plausible explanations include lack of the usual reduction in progesterone to initiate shedding of the thick and excessive endometrial tissues; ongoing proliferative activity in the endometrium even at the time of shedding and possibly generous blood flow that is established to support the enhanced endometrial growth; all of which may cause prolonged, heavy bleeding as well as incomplete cessation of bleeding.

Endometrial carcinoma is the commonest gynecological malignancy in the western world. Two types of EC have been described. Type 1 tumors are usually endometrioid in type and arise on a background of EH associated with excessive unopposed estrogen secretion, often with cytological atypia. These tumors are seen in association with obesity and PCOS, as well as other causes of excessive estrogen. By contrast, type 2 tumors arise on a background of atrophic endometrium and are not estrogen dependent tumors. These tumors include uterine serous carcinoma and clear cell carcinoma and typically have an adverse prognosis. Despite the different associations with hyperplastic and atrophic endometrium, mixed endometrial and serous/clear cell carcinomas are not uncommon [67] and recent molecular studies have suggested that the classification of endometrial carcinomas is more complex [68].

Most women presenting with EC are postmenopausal [69]. EC is rarely seen in premenopausal women, particularly in women below the age of 40 years, and hence EC is a rare cause of HMB. The carcinogenesis process is common in highly regenerative tissues and yet the large number of repetitive regeneration cycles that the endometrium endures does not appear to increase the potential for endometrial carcinogenesis. However, EC is a hormonally driven disease and the continuous, unopposed effect of the mitotic action of estrogen is thought to be the primary oncogenic promoter in at least 80% of endometrial cancers [5]. Therefore, conditions that dysregulate the sequential exposure to progesterone of the estrogen primed endometrium, such as obesity, PCOS, diabetes, estrogen secreting ovarian tumors and tamoxifen use, among others, will increase the risk of EC [69,70]. The usual pattern of the bleeding associated with EC is irregular and continuous rather than regular HMB and is thought to be secondary to disruption of endometrial vessels due to invasion by malignant cells and abnormal neovasculogenesis [37].

Coagulopathy

Disruption of the endometrial vasculature at menstrual shedding is the reason for initiation of menstrual blood loss. Spurting of blood from gaping exposed vessels has been observed at hysteroscopic examination of the menstrual endometrium. The cessation of this bleeding requires platelet aggregation and clot formation; damage to endothelial cells causes secretion of von Willebrand factor which initiates both events. The deposition of thrombin adds strength to the clot and seals the vascular lumen preventing continued blood loss, followed by fibrinolysis of the clot and sealing of the open end of the vessel by PAI-1, urokinase and tissue plasminogen activators (uPAs and tPAs) [9]; this process is likely to be required for the end of menstrual blood loss. Conversion of plasminogen to plasmin by PAs initiates fibrinolysis in the clot and this process is inhibited by PAI-1. The equilibrium between activation and inhibition of the fibrinolytic system plays a major role in the maintenance of the intravascular clot. Increased endothelial cell tPA activity may be associated with the dissolution of the clot found within 24 h of menstruation [9].

Leucocytes are a prominent component of endometrial stroma throughout the menstrual cycle but numbers increase dramatically in the mid and late secretory phase so that in premenstrual endometrium >30% of endometrial stromal cells are leucocytes [71]. Infiltration of stromal leucocytes is a feature of late secretory phase and premenstrual endometrium. These leucocytes produce a variety of cytokines, chemokines and proteases that may play a role in initial disruption of the integrity of the stratum functionalis tissue including blood vessels. The vasoactive prostaglandins that are produced in the menstrual endometrium also induce vasculogenesis and epithelial repair via VEGF [4,37].

The overall prevalence of a laboratory diagnosis of von Willebrand disease in women presenting with HMB has been reported to be as high as 13% [72]. Abnormal platelet aggregation has also been reported to be common and these two conditions should be ruled out before embarking on investigations for other haemostatic disorders in women presenting with primary HMB [9]. Other conditions such as disorders of deficiency in clotting factors (inherited or acquired hemophilia A and B, chronic liver disease, vitamin K deficiency and factor VIII inhibitor); disorders of fibrinolysis and thrombocytopenic disorders can in theory cause HMB but are extremely rare [9]. Platelet function disorders (PFDs), a heterogeneous group of inherited, qualitative platelet defects have emerged as an important cause of HMB, particularly in adolescents [27]. Women with PFDs and HMB have been reported to be a clinically distinctive subgroup of women with HMB, with significantly high incidences of blood group O and the δ-storage pool deficiency with a PFD diagnosed well after menarche. High false negative standard platelet function study results are reported therefore additional diagnostic strategies, such as electron microscopy to detect significantly reduced platelet δ-granule numbers should be considered [27].

A decrease in clot strength and integrity due to the breakdown of fibrin results in an increased blood loss during menstruation. Plasminogen activators are fibrinolytic in that they induce lysis or dissolution of blood clots and endometrial and menstrual effluent levels of plasminogen activators are increased in women with excessive menstrual blood loss [73].

Ovulatory dysfunction (hormonal causes of HMB)

Normal menstrual bleeding depends on the sequential exposure of the estrogen-primed endometrium to estrogen and progesterone, followed by withdrawal of progesterone at the end of the menstrual cycle. Any endocrine irregularity that prevents these sequential events may potentially result in HMB.

Polycystic ovarian syndrome is a common gynecological condition presenting with anovulatory cycles, obesity, features of excessive androgens and HMB. Androgens are converted into estrones in peripheral tissue, resulting in prolonged periods of excessive and unopposed estrogen action on endometrium during the anovulatory cycle. Anovulatory PCOS endometrium is thick prior to the start of bleeding and also lacks the drop in progesterone which is the trigger for the normal but highly orchestrated cellular events associated with menstrual shedding [74].

A further increasingly common condition that interferes with normal hormonal equilibrium in the endometrium is obesity. The conversion of androstenedione secreted by the adrenal gland into estrone by aromatase in adipose tissue provides an important source of additional estrogen for the endometrium [75]. This results in excessive estrogen driven endometrial growth and often produces HMB associated with shedding of the thick stratum functionalis.

Endometrial causes of HMB

According to hysteroscopic observations, menstrual shedding of the endometrium does not occur as an orderly process; rather there is patchy loss of the superficial functionalis layer [76]. There are focal islands of epithelial denudation, stromal breakdown and loss of vascular integrity, which cause isolated areas of bleeding, while simultaneous tissue regeneration is initiated in other areas. With this seemingly chaotic process of contrasting changes that occur in different areas of the endometrium during menstruation, it is not surprising that aberrations are widespread. However, consistent HMB with every menses is only reported in up to one in five women [1], suggesting that in healthy women a tight regulation of this process exists, whereas in persistent HMB, abnormalities in the endometrial progenitor zone in the stratum basalis may produce an abnormal stratum functionalis and abnormal regeneration process.

Regulation of menstrual shedding and cessation of menstrual bleeding is not fully understood. The previously proposed theory that re-epithelialization of the luminal epithelial surface is the primary factor in cessation of bleeding [73] is not supported by the histological observations that complete re-epithelialization is frequently seen on days 1 and 2 of the menses, despite continued menstrual bleeding. The stratum basalis of the endometrium maintains tissue integrity during menstrual breakdown of the stratum functionalis [5]. The differences between the cells in the two functionally and possibly phenotypically different layers, are only beginning to be understood [6].

Angiogenesis is development of new microvessels from existing blood vessels. Human endometrium is one of the very few adult organs where regular physiological angiogenesis occurs; in the basalis layer during menstruation and in the stratum functionalis and subepithelial capillary plexus during the proliferative and early secretory phases. Mechanisms of regulation of endometrial vascular growth remain to be fully determined. VEGF produced by intravascular neutrophils has been suggested to play a role in the development of these vessels by intussusception and elongation in the endometrium [77]. Abnormal or incomplete angiogenesis, resulting in abnormal blood vessels with fragile vessel walls may cause HMB and several studies have suggested that abnormal endometrial vessel number, structure or function may play a role in HMB. Increased blood flow has been suggested [78] and reduced vascular smooth muscle proliferation and lower expression of myosin heavy chain [79,80] may reflect altered endometrial vascular development in HMB. A recent study has reported that, although vessel number did not differ between control and HMB, there was altered expression of vascular smooth muscle differentiation markers in HMB, again suggesting altered endometrial vascular development which could impact on function [81].

Iatrogenic

Agents such as tamoxifen that have an estrogenic effect can induce endometrial growth and may be associated with HMB. However, women on progestogenic hormonal therapy such as the contraceptive progestogen only pill, implant or Depot may also experience paradoxical HMB, which is thought to be secondary to the abnormal vasculogenesis associated with progestogen action [37]. Although progestogens are a common treatment for HMB, bleeding irregularities are a common reason for discontinuing progestogens, with 2% of women on Depot injections of medroxyprogesterone actetae suffering excessive bleeding [37]. By contrast, preoperative treatment with the PR modulator, asoprisnil (with antiprogestogenic properties), is associated with development of aggregates of thin walled vessels as well as and thick walled muscular vessels in endometrium, suggesting that progesterone may have an inhibitory effect on endometrial vasculogenesis, and therefore should reduce HMB [82].

Administration of a foreign object into the uterine cavity, such as inert copper contraceptive devices, is a well-established cause of increased menstrual loss and is possibly due to interference with normal endometrial development and myometrial contractility.

Not yet classified

The PALM-COEIN classification has allowed this category to include some rare uterine abnormalities of HMB and any potential future entities which can either cause or contribute to HMB. This includes some poorly defined entities such as chronic endometritis, arteriovenous malformations and myometrial hypertrophy [3]. This category also includes rare conditions such as the glycolipid storage disorder Gaucher disease, in which HMB is a common symptom [83].

Conclusion

HMB is a common gynecological complaint with multiple etiologies and diverse pathophysiological origins. Our present understanding of the mechanisms of menstrual bleeding as well as many of the pathological aberrations of HMB is incomplete. Further research into the pathophysiology of HMB is urgently needed, as clear knowledge of the mechanisms of the disorder will provide new therapeutic targets to formulate more effective treatments.

Future perspective

There is an increasing interest to find novel, non-surgical and fertility sparing management options for HMB. The future perspective therefore includes the development of therapies directed towards new avenues such as endometrial vasculature, myometrium and endometrial stem cells. Improving the basic scientific knowledge in these areas of pathophysiology of HMB in particular is warranted for timely discoveries to be made for the benefit of millions of women suffering with this distressing condition.

Executive summary

• Heavy menstrual bleeding (HMB) is a common gynecological complaint with multiple etiologies and diverse pathophysiological origins.

• Our current understanding of the mechanisms of menstrual bleeding as well as many of the pathological aberrations of HMB is incomplete.

• Further research into the pathophysiology of HMB is urgently needed, as clear knowledge of the mechanisms of the disorder will provide new therapeutic targets to formulate more effective treatments.