1Department of Reproductive Endocrinology, King George V and Royal Prince Alfred Hospitals, Camperdown, NSW and 2Department of Obstetrics and Gynaecology, University of Sydney, NSW 2006, Australia
3To whom correspondence should be addressed. E-mail: helena@obsgyn.usyd.edu.au
Over the past 10 years there has been an upsurge of interest in the mechanisms underlying normal and disturbed menstrual bleeding. These studies have particularly focused on the mechanisms underlying the common problems of menorrhagia associated with ovulatory and anovulatory dysfunctional uterine bleeding (DUB) and of unpredictable breakthrough bleeding during hormonal contraceptive use. A wide range of abnormalities of endometrial morphology and function have been demonstrated, but it is still not clear how all the pieces of this complex jigsaw puzzle ®t together. Ovulatory DUB is predominantly associated with decreased endometrial vasoconstriction and vascular haemostatic plug formation, leading to defective control of the volume of blood which is lost during menstruation. By contrast, breakthrough bleeding is associated with a wide range of molecular disturbances which appear to result in unpredictable vessel breakdown through disturbed endometrial angiogenesis, increased vascular fragility and loss of the integrity of the endothelial, epithelial and stromal supporting structures. Anovulatory DUB is very poorly understood, but may be associated with disturbed angiogenesis, fragile vessels and defective haemostatic processes. Little is known about the actual mechanisms of the common problem of abnormal bleeding associated with speci®c genital tract pathologies such as uterine myomata.
Key words: endometrium/menorrhagia/menstrual bleeding/progestogens
Mechanism of abnormal menstruation
Abnormalities of regularity and frequency
Conclusion
References
Introduction
The spectrum of abnormal uterine bleeding affects up to one-third
of women of child bearing age and is therefore one of the
commonest complaints seen by family doctors and gynaecolo-
gists. These disturbances lead to considerable social and physical
morbidity in all societies, and may also be a re¯ection of serious
underlying pathology. Menorrhagia affects 10±30% of menstru-
ating women at any one time, and may occur at some time during
the perimenopause in up to 50% of women (Ballinger et al., 1987;
Prentice, 1999). Disturbance of the menstrual pattern is almost
universal with long-acting methods of hormonal contraception,
although the patterns usually improve with time and most women
tolerate the changes well (Odlind and Fraser, 1990). Nevertheless,
>50% of women who prematurely discontinue use of these
methods cite menstrual disturbance as the main reason for this
decision. A similar problem is found with peri- or postmenopausal
hormone replacement therapy (Nachtigall, 1990).
The need to have a regular menstruation has been culturally
seenÐerroneouslyÐas essential for women's health and episodes
of amenorrhoea are still usually perceived as pathological. Many
women have rejected modes of contraception that did not give a
regular bleed, despite the fact that in primitive human cultures
women had many fewer periods in a lifetime as a result of later
menarche, more pregnancies, longer episodes of breast feeding and
earlier menopause (Short, 1976; Thomas and Ellertson, 2000).
The nature of abnormal uterine bleeding
The perception of menstrual bleeding varies considerably
amongst individual women. Studies designed to quantify
menstrual loss have demonstrated considerable inaccuracy in
subjective assessment of volume and women's recall of menstrual
events has been shown to be unreliable (Rodriguez et al., 1976;
Snowden, 1977; Fraser et al., 1984). This unreliability relates just
as much to perception of regularity, frequency, duration and dates
of events as to perception of volume.
Of those women who present with a clinically signi®cant
complaint of menorrhagia, <50% will have the complaint
Human Reproduction Update, Vol.8, No.1 pp. 60±67, 2002
60 Ó European Society of Human Reproduction and Embryology
objectively con®rmed (Chimbira et al., 1980; Fraser et al., 1981).
However, what is normal for one woman can be abnormal for
another and change in pattern may be important. Most research
techniques require dedicated subjects to maintain daily diary
records and to collect all sanitary towels for quanti®cation. These
studies have shown mean menstrual blood loss to be ~30 ml per
cycle in most societies, with loss >60±80 ml per month being
associated with an increased tendency towards iron de®ciency and
anaemia (Hallberg et al., 1966; Cole et al., 1971). An upper limit
of 60 ml may be more appropriate clinically. However, the total
menstrual loss is very dilute blood, since half is transudate from
the endometrium (Fraser et al., 1985).
The duration of normal menstruation also varies greatly, with
an average of 5 days and the heaviest loss usually on the ®rst 2
days (Matsumoto et al., 1962; Rubin and Crosignani, 1990).
Duration of ¯ow is considered abnormal when it lasts <2 days or
>7 days.
life has been well characterized (Treloar et al., 1967; Vollman,
1977). Abnormal uterine bleeding may involve any disturbance of
regularity, frequency, duration or volume of menstrual ¯ow, and
the causes may be physiological, pathological or pharmacological
(Fraser and Sungertekin, 2000). Sound clinical management
demands that the underlying cause is de®ned. An understanding
of pathophysiological mechanisms may also be valuable in
determining appropriate therapies. Two of the most important
symptoms are menorrhagia and the occurrence of irregular
breakthrough bleeding with hormonal contraception, and empha-
sis will be placed
endothelium. It circulates in plasma as a complex with factor
VIII, which it protects from proteolytic degradation. Certain
variants of Von Willebrand's Disease may be associated with
severe menorrhagia, but it is not yet clear how often less severe
variants cause moderate menorrhagia.
patient who presents with menorrhagia, a family history of a
bleeding disorder, easy bruising or an increase in bleeding
associated with trauma should alert the clinician to the possibility
of a non-gynaecological cause.
as a consequence of anovulatory dysfunctional uterine bleeding
(DUB), but it is a rare cause of this symptom. Most women with
untreated hypothyroidism and hyperthyroidism will develop
amenorrhoea.
Dysfunctional uterine bleeding
DUB is a diagnosis of exclusion and a common cause of the
symptom, `menorrhagia'. These terms are frequently and
mistakenly used synonymously and there is a great deal of
confusion as to exactly which types of abnormal bleeding should
be termed DUB. All are agreed that the terms should refer to
excessive bleeding without pelvic pathology, but in the USA, the
de®nition of `dysfunctional uterine bleeding' refers predomi-
nantly to anovulatory bleeding and the term `menorrhagia' to both
the symptom of excessive bleeding and to the diagnosis of
ovulatory dysfunctional bleeding (Cowan, 1992; Fraser and
Sungurtekin, 2000). This is quite different from most of the rest
of the world. Our favoured de®nition of DUB, which has been
endorsed by the European Society for Human Reproduction and
Embryology, is `excessive bleeding (excessively heavy, pro-
longed or frequent) of uterine origin which is not due to
demonstrable pelvic disease, complications of pregnancy or
systemic disease' (Rubin and Crosignani, 1990; Fraser and
Sungurtekin, 2000). Dysfunctional uterine bleeding accounts for
~50% of all cases of excessive menstruation (Beazley, 1972) and
is ovulatory in ~80% of cases (Cameron, 1989; Fraser, 1989).
Anovulatory DUB. Anovulatory DUB is recognizable as irregular,
prolonged and usually excessive bleeding caused by a disturbed
function of the hypothalamic±pituitary±ovarian axis. It is most
commonly seen in polycystic ovary syndrome and at the extremes
of female reproductive life, namely the perimenarchal and
perimenopausal years. At these stages of life, cycles may be
intermittently ovulatory and anovulatory, thus leading to great
irregularity of menstruation as well as great variability in blood
loss.
The exact mechanisms behind anovulatory bleeding are
uncertain (Fraser et al., 1996) but it is known that unopposed
estrogen can lead to excessive endometrial proliferation and
hyperplasia with increased and dilated draining veins and
suppression of spiral arterioles (Beilby et al., 1971). Large,
thin-walled, tortuous, super®cial endometrial vessels can often be
demonstrated on the surface of hyperplastic endometrium
(Hamou, 1985) and increased fragility is a probable contributor
to increased blood loss. Unopposed estrogen has a direct effect on
the uterine blood supply by reducing vascular tone (Fraser et al.,
1987) and possibly an indirect effect through inhibiting
vasopressin release (Akerlund et al., 1975) leading to vasodilata-
tion and increased blood ¯ow. Unopposed estrogen also
stimulates stromal VEGF expression which may contribute to
disturbed angiogenesis (Zhang et al., 1995; Smith, 1998). In
addition, endometrium exposed to prolonged, unopposed estrogen
synthesizes less PG and a higher proportion of PGE than PGF
(Smith et al., 1982).
In women with endometrial hyperplasia, the endometrium often
breaks down unevenly, even when circulating estrogen concen-
trations are high or rising (Brown et al., 1959), with scattered red
patches at hysteroscopy corresponding to thrombotic foci of
necrotic disintegration, adjacent to the abnormally proliferated
endometrium (Schroder, 1954). An increased production of
endometrial nitric oxide (endothelium-derived relaxing factor)
in response to excessive and unopposed estrogen has been
postulated as another mechanism leading to excessive blood loss
in anovulatory menstruation (Chwalisz and Gar®eld, 2000).
Ovulatory DUB. Ovulatory DUB is characterized by regular
episodes of heavy menstrual loss, with 90% of the loss on the
®rst 3 days as in normal menstruation (Haynes et al., 1979). There
is no disturbance of the hypothalamic±pituitary±ovarian axis and
indeed the gonadotrophin and steroid hormone pro®les are no
different to those seen in a normal menstrual cycle (Haynes et al.,
1979; Eldred and Thomas, 1994).
The decline in estrogen and progesterone levels in the late
luteal phase initiates many processes which act simultaneously,
leading to disintegration then re-epithelialization of the functional
layer of the endometrium during menstruation. There is little
evidence to suggest that there are substantial abnormalities in the
sequence of processes of breakdown, regeneration or remodelling
of the endometrium leading to the increased menstrual loss in
ovulatory DUB. The main defect appears to be in the control of
processes regulating the volume of blood lost during menstrual
breakdown of the endometrium, primarily the processes of
vasoconstriction and haemostasis.
Alterations of follicular phase endometrial blood ¯ow have
been noted in women with ovulatory DUB and may re¯ect or
in¯uence disturbances of function occurring in the tissue (Fraser
et al., 1987). Endometrial glandular and stromal estrogen and
progesterone receptor levels may be increased in the late secretory
phase in women suffering from DUB (Gleeson et al., 1993) and
these may be responsible for mediating the effects of increased
estrogen. However, others have not been able to demonstrate
differences in expression of these receptors in women with
menorrhagia (Critchley et al., 1994).
Endometrial small surface vessels were studied at hysteroscopy
in women with ovulatory DUB (Hickey et al., 1996), and these
vessels appear to be similar in size and appearance to those of
normal women (Hickey et al., 1996, 1998). These vessels also
appear to be normal in their response to intrauterine stretching, in
direct contrast to the increased fragility exhibited by small
endometrial vessels exposed to low dose progestogens, such as
Norplant (Hickey et al., 2000a). However, minor abnormalities of
endometrial and myometrial veins have been described (Hourihan
et al., 1989).
63
A range of endometrial molecular systems have been studied in
preliminary fashion in women with ovulatory DUB and some
functional abnormalities have been demonstrated. Endothelins
probably act on receptors at the endometrial±myometrial inter-
face, which increase at the time of normal menstruation (Cameron
et al., 1992). Reduced levels of endothelins may lead to an
increase in the volume of blood lost, and Marsh et al. have
demonstrated less intense endothelin-like immunoreactivity in
endometrial glandular and luminal epithelium in women with
menorrhagia than in controls (Marsh et al., 1996).
This may be one of the more important mechanisms limiting
blood loss in normal menstruation. PGs also have powerful
vasoactive effects and extensive evidence points to an important
role in the prevention of excessive menstrual blood loss (Baird et
al., 1996). Endometrial PG release is greatly in¯uenced by
circulating steroid levels (Smith and Kelly, 1987). PGF2a induces
vasoconstriction and PGE2 and prostacyclin (PGI2) induce
vasodilation. PGI2 is also one of the most potent substances
known for preventing platelet aggregation and the formation of
haemostatic plugs. An increase in total PG release and
disproportionate rise in PGE2 have been demonstrated in
ovulatory DUB (Smith et al., 1981a). It has also been shown
that there is an increase in PGE2 and PGI2 receptors, predisposing
to vasodilatation, in women with menorrhagia (Adelantado et al.,
1988). A potential for increased synthesis of PGI2 by myome-
trium from endometrial precursors in ovulatory DUB has also
been clearly demonstrated (Smith et al., 1981b).
Antiprostaglandin medications are effective in treating ovulatory
DUB, probably mainly by reducing endometrial synthesis of all
PGs, but also by inhibiting the binding of PGE to its receptors.
Haemostatic mechanisms are very important in limiting the
volume of blood lost at menstruation. Prevention of platelet
aggregation by increased PGI2 release may be an important
contributing factor in ovulatory DUB (Smith et al., 1981b) as may
increased endometrial tissue plasminogen activator content,
increased local ®brinolytic activity (Bonnar et al., 1983;
Casslen et al., 1986; Gleeson et al., 1993) and excessive
endometrial heparin-like activity (Paton et al., 1980). This
combination of increased PGI2, ®brinolysis and heparin-like
activity will result in very limited and de®cient haemostatic plug
formation. Mast cells degranulate at menstruation to release a
number of substances including heparin, which reduces ®brin
formation, and histamine which causes endothelial cell contrac-
tion, resulting in gaps between the vascular endothelial cells and
both transudation and red cell loss. In women with ovulatory
DUB or with menorrhagia secondary to the use of an IUCD, an
increase in the enzymatic activity of lysosomes has also been
demonstrated (Wang et al., 2000). This may contribute to
abnormalities of the process of tissue breakdown and remodelling.
Matrix metalloproteinases may be even more important in
contributing to abnormal endometrial breakdown and abnormal-
ities of menstrual bleeding (Salamonsen et al., 2000).
When other systems are also investigated in detail, it is likely
that other disturbances of molecular function will be demon-
strated. Recent research has begun to focus on the activation of
speci®c genes which may be associated with abnormal uterine
bleeding of various types. A novel gene called `endometrial
bleeding associated factor' (ebaf) has recently been demonstrated
to be transiently expressed in endometrium at normal menstrua-
tion, and much more strongly expressed when bleeding is
abnormal (Kothapalli et al., 2000). This gene is located on
chromosome 1 and codes for a member of the TGFb superfamily.
The exact function is not clear. The concept of multigene
activation patterns is now being explored as a possible
explanation for complex functional disturbances of the type
exempli®ed by ovulatory DUB.
Abnormalities of regularity and frequency
Oligomenorrhoea, amenorrhoea and polymenorrhoea
not fall directly within the remit of this review. Oligomenorrhoea
may sometimes be associated with anovulatory DUB, especially
during adolescence (Fraser et al., 1973).
Intermenstrual and postcoital bleeding, premenstrual spotting
These symptoms are usually superimposed on otherwise normal
and regular menstrual periods, and are mainly caused by surface
lesions of the genital tract (such as cervical or endometrial polyps,
infection or carcinoma) (Fraser and Petrucco, 1996). Actual
pathophysiological mechanisms leading to this abnormal bleeding
are poorly understood, but are probably due to the presence and
dilatation of abnormal surface vessels on these lesions.
Breakthrough bleeding
for women to discontinue hormonal contraceptive use, even
though most BTB is usually prominent mainly in the ®rst few
months of treatment. The oral contraceptive pill contains both
estrogen and progesterone and BTB is not a common reason for
discontinuation of use as >90% of pill users report no disturbance
of their bleeding patterns (Belsey et al., 1988). Most episodes of
BTB are caused by missed pills or the occasional failure of a
withdrawal bleed. Any other factor interfering with the circulating
concentrations of contraceptive steroids can lead to BTB and it is
especially common when simultaneous medications, such as
anticonvulsants, interact.
Progestogen-only contraceptives now come in many different
forms with a range of contraceptive effects at all levels of the
hypothalamic±pituitary±ovarian axis. Depot medroxyprogester-
one acetate (DMPA) produces highly unpredictable bleeding
patterns with a substantial incidence of amenorrhoea, oligome-
norrhoea and erratic, frequent and prolonged light bleeding and
spotting, while lower dose preparations like the subdermal
levonorgestrel implant system, Norplant, produce some unpre-
dictable, irregular light bleeding which is somewhat better
tolerated. Nevertheless, all progestogen-only systems produce
unpredictable bleeding which many women eventually ®nd
intolerable (Belsey et al., 1988). With most methods the bleeding
pattern improves with time.
continuous exposure of the endometrium to relatively constant
doses of progestogen, with simultaneous exposure to low levels of
estrogen. This results in a variety of endometrial histological
pictures which are broadly described as `suppressed secretory' but
sometimes verge on complete atrophy (Maqueo, 1980). Surface
M.Livingstone and I.S.Fraser
breakdown of the endometrium is erratic and does not occur
synchronously over the whole surface (Hickey et al., 1996).
Hysteroscopic studies in Norplant users have shown irregular
surface breakdown which correlates positively with the duration
of BTB over the previous month (Hickey et al., 1996). These
same hysteroscopic studies have also demonstrated abnormal
angiogenesis with neovascularization patterns over the endome-
trial surface. These vessels vary considerably in size and some are
much wider in diameter and thinner-walled than normal (Hickey
et al., 1998). These vessels are also much more fragile than
normal and bleed easily in response to minor stress (Hickey et al.,
2000a). There are also disturbances in surface vessel ¯ow and
capillary vasomotion as demonstrated by laser Doppler probe
(Hickey et al., 2000b). All of these abnormalities are a re¯ection
of disturbed angiogenesis.
The cause of the increased vascular fragility is not entirely
clear (Fraser and Diczfalusy, 1980; Hickey et al., 1996), but there
is important evidence that endometrial pericytes (Rogers et al.,
2000) and basement membrane composition are de®cient in many
of these women (Hickey et al., 1999). There is limited evidence
that endothelial tight junctions and adhesion molecule expression
may also be affected. In order for bleeding to occur, the surface
epithelium must also be breached, and limited evidence indicates
that progestogen exposure leads to loosening of the surface
epithelial attachments (Murphy, 2000).
Increasing evidence points to a series of disturbances of cellular
and molecular function in the endometrium of women exposed to
continuous progestogens, all of which could contribute to
disturbed angiogenesis, increased spontaneous tissue breakdown
or defective repair (Hickey and Fraser, 2000). These include
increased expression of matrix metalloproteinases (Marbaix et al.,
2000; Salamonsen et al., 2000; Vincent and Salamonsen, 2000),
endothelial cell dysfunction (Rodriguez-Manzaneque et al.,
2000), vascular endothelial growth factor (Charnock-Jones et
al., 2000), reduced epithelial cytokeratin expression
(Wonodirekso et al., 2000), increased expression of tissue factor
(Lockwood et al., 2000), multiple alterations in concentrations
and functions of migratory endometrial leukocytes (Song et al.,
1995; Clark et al., 1996) and a range of other disturbed
angiogenic factors (Smith, 2000).
It should not be forgotten that many cases of BTB in modern
clinical practice are associated with the use of hormone
replacement therapy. Somewhat surprisingly, relatively little is
known about the mechanisms of this form of BTB, although there
are some simularities to BTB associated with progestogen-only
contraception (Thomas et al., 2000).
Conclusion
problem, with varied presentations and multiple causes. Over the
past 10±15 years there has been a major escalation in the scienti®c
effort aimed at elucidating the underlying mechanisms. It is now
clear that these mechanisms are complex and involve a number of
different molecular systems. This review has attempted to pull
together the evidence from different disciplines to provide a
picture of the way in which these systems interact to cause
different symptoms.
References
Adelantado, J.M., Rees, M.C.P., Lopez Bernal, A. et al. (1988) Increased uterine prostaglandin receptors in menorrhagic women. Br. J. Obstet. Gynaecol., 95, 162±165.
Akerlund, M. Bengtsson, L.P. and Carter, A.M. (1975) A technique for monitoring endometrial or decidual blood ¯ow with an intrauterine thermistor probe. Acta Obstet. Gynecol. Scand., 54, 469±477.
Andersson, K. and Rybo, G. (1990) Levonorgestrel-releasing intrauterine device in the treatment of menorrhagia. Br. J. Obstet. Gynaecol., 97, 690± 694.
Andrade, A.T.L. and Pizarro-Orchard, E. (1987) Quantitative studies…