Stephen C. L. Koh, Department of Obstetrics and Gynaecology, National University of Singapore, Yong Loo Lin School of Medicine, National University Hospital, Lower Kent Ridge Road, Singapore 119074, Singapore. Tel.: +65 67724114; fax: +65 67794753; e-mail: email@example.com
Summary. Background: Menorrhagia is known to be associated with uterine fibroids, adenomyosis, pelvic infections, endometrial polyps and clotting defects. A viable alternative therapy to hysterectomy should alleviate heavy menstrual blood flow and consequently improve the quality-of-life measures in women presenting with menorrhagia. The levonorgestrel-releasing intrauterine system (LNG-IUS) ranks higher than medical treatments in terms of efficacy, comparable improvements in quality of life and psychological well-being. Objective: The purpose of the study was to determine the effects of 6 months of LNG-IUS use on menstrual blood loss and the hemostatic, fibrinolytic/inhibitor systems in blood and the endometrium in women with menorrhagia with known pathologic causes. Patients and methods: Samples from 41 women were analyzed. Hemoglobin, hematocrit, thrombelastography, tissue-type plasminogen activator (t-PA), urokinase plasminogen activator (u-PA), u-PA receptor (u-PAR), plasminogen activator inhibitor-1/2 (PAI-1/2), D-dimer and von Willebrand factor (VWF) were determined, and t-PA, u-PA and PAI-1/2 were also determined in endometrial tissue extracts. Results: Menorrhagia was reduced in 89% of women by 3 months; by 6 months all women had no menorrhagia, and 39% of women had become amenorrhoeic. Hemoglobin and hematocrit levels showed improvement, and reached normal reference levels by 6 months. There were no systemic changes in the fibrinolytic/inhibitor systems and VWF, except for a decreased u-PAR level. However, in the endometrium, significant elevations in PAI-1/2 together with u-PAR levels were seen at 6 months. Conclusions: The slow levonorgestrel-release intrauterine device use results in high expression of fibrinolytic inhibitors (PAI-1/2) and upregulated u-PAR expression in the endometrium. Systemic hemostasis was not significantly altered. The study demonstrated that LNG-IUS is highly effective in the treatment of menorrhagia with known pathologic causes.
Menorrhagia is defined as total menstrual blood loss (MBL) of > 80 mL per menstrual cycle and is subjectively defined as a complaint of heavy cyclic bleeding over several consecutive cycles in women in their reproductive years. Menorrhagia is known to be associated with uterine fibroids, adenomyosis, pelvic infections, endometrial polyps and clotting defects. It is a problem that afflicts 9–14% of the female population . Apart from hysterectomy, no other methods of treatment have been extensively validated to date. Hysterectomy carries a significant risk of morbidity and mortality. A viable alternative should be available to alleviate heavy menstrual flow and, consequently, improve the quality of life in women with menorrhagia. Current treatments used for menorrhagia include various medical therapies and the levonorgestrel-intrauterine system (LNG-IUS). Medical management of menorrhagia includes using oral contraceptives, tranexamic acid or merenamic acid . The LNG-IUS, developed primarily as a contraceptive device, is now used in the management of menorrhagia. The LNG released into the uterine cavity suppresses endometrial proliferation and activity , resulting in reduced MBL. The LNG-IUS was associated with a substantial reduction in MBL, as well as having the benefits of retaining fertility with comparatively few side effects [4–6]. One study reported that 64.3% of women receiving LNG-IUS treatment canceled their decision to undergo hysterectomy as compared to 14.3% in the control group . Decreases in MBL of 86% after 3 months and 97% after 12 months of LNG-IUS use have been reported , and Tang and Lo  found a decrease of MBL by 95% after 6 months of use in 10 women with menorrhagia and anemia. Use of the LNG-IUS is highly effective [4,8], and has potential as a long-term effective alternative to surgery in the treatment of women with menorrhagia [7,9].
Coagulation disorders can lead to excessive uterine bleeding, leading to menorrhagia and associated anemia. Hysterectomy may be necessary for the ultimate control of bleeding. Inherited bleeding disorders have been reported in 17% of 150 women diagnosed with menorrhagia [13% von Willebrand factor (VWF), 4% factor XI deficiency] despite normal pelvic examination and sonography findings . Bleeding disorders, especially von Willebrand disease, were also diagnosed in another two studies in 10.7%  and 32%  of women. This indicates the importance of considering inherited bleeding disorders as a cause of menorrhagia to reduce unnecessary surgical intervention. Underlying hemostatic defects, and in particular qualitative platelet abnormalities, occur in the majority of women with unexplained menorrhagia . Other than indicating the severity of anemia, hemoglobin is of limited value in screening women with menorrhagia . Anemia is defined as a hemoglobin level of < 12 g dL−1 .
Fibrinolysis in the endometrium plays an important role in menstruation. Levels of plasminogen activators, especially tissue-type plasminogen activator (t-PA) and its inhibitor, plasminogen activator inhibitor-1 (PAI-1), were found to be significantly increased in the late secretory endometrium . They are influenced by the cyclic variation of the sex hormone concentration at different phases of the menstrual cycle. Increased numbers of monocytes or macrophages known to adhere to intrauterine contraceptive devices (IUCDs) are associated with increased fibrinolytic activity in the endometrium [17,18], whereas no increase was seen on progesterone-releasing devices . In dysfunctional uterine bleeding (DUB), increased fibrinolytic activity was observed in the menstrual fluid and is suggested to be a major contributing factor in the etiology of menorrhagia .
This study was designed to determine the effect of the LNG-IUS use on MBL and the hemostatic, fibrinolytic/inhibitor systems in the treatment of women with menorrhagia with known pathologic causes.
Materials and methods
Study design and treatment
The study was an open, non-randomized longitudinal study with subjects acting as their own controls.
Approval was obtained from the Hospital Ethics Committee to conduct the study. Parous women referred to the gynecologic clinic of the hospital for menorrhagia were recruited after obtaining informed consent. This only included women with known pathology and who did not require immediate surgery. MBL was quantified using a pictorial blood-loss assessment chart  for two consecutive cycles before entry into the study. Women with menorrhagia were then fitted with the LNG-IUS, and each acted as their own control for the subsequent cycles.
Pictorial blood-loss assessment chart
Each woman referred to the clinic for menorrhagia was given a pictorial chart, to chart light, moderate and heavily soiled tampons or towels. A mark was to be made in the appropriate box for every tampon or towel used. Passage of blood clots and episodes of flooding were recorded as previously described . After the menses, the pictorial chart was returned for scoring by the gynecologist (KS). Ascending scores were assigned: 1 for each lightly stained tampon; 5 if moderately soiled; and 10 if completely soaked with blood. Small and large clots scored 1 and 5 respectively. The total numerical score for each pictorial chart was then calculated. A monthly score of 100 or higher is associated with uterine blood loss of > 80 mL when measured by the alkaline hematin method. Women demonstrating menorrhagia for two menstrual cycles by this method were then included in the study. Although the pictorial method of assessing MBL is not as accurate as an objective measurement, it still provides a simple and useful alternative to enable the clinician to discriminate between those who require treatment and those who should be dissuaded .
The following criteria were necessary for inclusion in the study: documented menorrhagia with the menstrual pictorial chart, menstrual cycle length of 28 ± 7 days, parous, no hormone therapy within the previous 3 months, not taking medication affecting MBL, no contraindication to IUCD use, no adnexal pathology at clinical and ultrasonographic examination, uterine size of < 14 weeks, submucous myomas of < 3 cm, no malignant or atypical endometrial changes on histology, normal cervical cytology, and no known bleeding diathesis or hemostatic disorders.
The LNG-IUS Mirena® was originally developed as a contraceptive to be inserted into the uterine cavity after menstruation, according to the insertion instructions, to treat women with menorrhagia. The contraceptive and therapeutic effects of the LNG-IUS are based on the local effects of the 20 μg per day of LNG released in the uterine cavity, resulting in a high concentration in the endometrium. Endometrial proliferation is inhibited, and this suppression of endometrial growth with reduced MBL properties can be used in the treatment of menorrhagia. After the device is removed, the morphologic changes in the endometrium revert to normal  and menstruation returns within 30 days, including the rapid return of fertility.
Endometrial tissue sampling
The procedure was performed using the Pipelle® device in an outpatient setting for the extraction of endometrial samples. In the pretreatment phase, one preinsertion sample was obtained and subsequent samples were taken at 1, 2, 3 and 6 months. All endometrial samples were obtained in the premenstrual phase (days 24–28 of a 28-day cycle), in order to obtain a reasonable yield of endometrial tissue for evaluation. In women with irregular spotting or who became amenorrheic, the samples were obtained between days 24–28 of the menstrual cycle. The endometrial specimens were rinsed in cold saline, blotted dry on filter paper, immediately snap-frozen using liquid nitrogen, and stored at −70 °C until further processing.
Endometrial tissue extracts
Briefly, about 100 mg of frozen tissue was first rinsed in cold normal saline and blotted dry on filter paper. In an ultracentrifuge tube of 10 mL capacity, 2 mL of Tris–buffered saline (pH 8.5) [extraction buffer as indicated for tissue PAI-1, urokinase plasminogen activator (u-PA) and u-PA receptor (u-PAR) kits; American Diagnostic Inc., Greenwich, IL, USA] was added to the washed frozen tissue. The tissue was then homogenized for 2–3 min in an Ultraturrex Homogenizer, and the tube was left standing in a beaker of crushed ice. Further homogenization was performed with the sonic dismembrator for another 5 min. The suspension was centrifuged at 30 000 times g in an ultracentrifuge for 30 min, and the supernatant was aliquoted and stored at −70 °C until assayed. Lowry's method for protein determination on the supernatant using bovine albumin standard was used, and the results were used for calculation of concentration per milligram of protein. The choice to use Tris–buffered saline (pH 8.5) for tissue homogenization was in agreement with the indications in the kits for the enzyme-linked immunmosorbent assays (ELISAs) for PAI-1, u-PA and u-PAR, supplied by American Diagnostics Inc., USA. In our preliminary studies of endometrial t-PA, u-PA and PAI-1 at different phases of the menstrual cycle, normal saline alone was used for tissue homogenization . However, in t-PA activity determinations, 0.1 m HCl was used for activation in both plasma and tissue extracts and then neutralized with 0.1 m NaOH; the acid-treated aliquots were then used for assays .
Blood sampling were performed before 11.00 h in the morning but prior to endometrial sampling, as indicated above. Nine parts of blood from a clean venepuncture were mixed with one part of 0.129 mol L−1 trisodium citrate (Merck, Sharp & Dohme, New York, NY, USA) containing 0.21 ml L−1 HEPES (Sigma-Aldrich, St Louis, MO, USA) in cold plastic tubes. The blood was immediately spun at 2000 × g for 15 min in a refrigerated centrifuge. For plasma t-PA activity, a 0.5-mL aliquot was acid-treated as described elsewhere . Acid-treated and untreated plasma aliquots were stored at −70 °C until assayed.
The following assays were performed on these blood samples – hematological indices: routine hemoglobin and hematocrit levels; citrated whole blood thrombelastography (cTEG) (Haemoscope, Niles, IL, USA) for hemostasis status ; t-PA activity  and antigen (American Diagnostic Inc.); u-PA activity , u-PA antigen and soluble u-PAR (Amer Diagn. Inc.); D-dimer (Stago, Ansiéres, France); PAI-1 activity  and PAI-1 antigen (Zymutest, Andrésy, France); PAI-2 antigen (Amer Diagn Inc.); and VWF, with an in-house chromogenic substrate assay using an international standard from the National Institute of Biological Standards and Control, Hemel Hempstead, UK.
Assays performed for endometrial tissue extracts were: t-PA activity and antigen; u-PA activity and antigen; soluble u-PAR antigen; and PAI-1 and PAI-2 antigen levels.
The Statistical Package for Social Sciences (SPSS12; SPSS Inc., Chicago, IL, USA) was used to perform statistical analysis. The analysis of variance (anova) and including post hoc Duncan's test for significance during the periods of LNG-IUS use was determined. Independent t-test and Mann–Whitney test for differences between groups was also performed. A P-value of < 0.5 was considered to be statistically significant.
In total, 52 women who fulfilled the above inclusion criteria were recruited into the study. Eleven women who did not complete the menstrual chart assessment score monthly or had only one blood sample taken during the period of the study were not included in the analysis. Data from 41 women who completed 6 months of pictorial chart for MBL and had available two or more samples of either blood or endometrial tissues during the period of the study were analyzed. One woman with adenomyosis had no preinsertion sample. The median of their menstrual cycle was 28 days, and the median age was 42.0 years (range: 30–49 years). The total number of blood and endometrial samples available from the 41 women in the 6-month study and known pathologic causes are shown in Table 1. Not all women had complete blood samples taken at the intervals mentioned, mainly because of their not turning up at specific times. Furthermore, endometrial aspiration in women who had become amenorrheic did not yield sufficient tissue material for analysis.
Table 1. Blood and endometrial samples available from the 41 women from five sampling intervals at pre-insertion, 1, 2, 3 and 6 months (n = number of women) in the study with known pathologies
DUB, dysfunctional uterine bleeding.
MBL following LNG-IUS use
MBL of > 80 mL per cycle was assessed using the pictorial blood-loss assessment chart. Menorrhagia was reduced in 63% (26/41) by the first month; at 3 months, menorrhagia was reduced in 88% (36/41) of women; and by 6 months, all women in the study had no menorrhagia; 39% (16/41) of women had in fact become amenorrhoeic (Fig. 1).
Hematologic indices and systemic hemostasis status in women with menorrhagia following 6 months of LNG-IUS use
Hemoglobin and hematocrit
At preinsertion, the median hemoglobin level of 117 g L−1 showed the presence of anemia, but by 6 months of use, the hemoglobin and hematocrit levels showed a significant elevation, reaching normal reference levels (Table 2). This was associated with a significant reduction in MBL in the treatment of menorrhagia by 6 months of LNG-IUS use.
Table 2. Hemoglobin, hematocrit, coagulation status and systemic fibrinolytic and inhibitor levels following six months of LNG-IUS use in the treatment of menorrhagia. Median and (inter-quartile range)
t-PA, tissue-type plasminogen activator; u-PA, urokinase plasminogen activator; u-PAR, u-PA receptor; PAI-1, plasminogen activator inhibitor-1; VWF, von Willebrand factor; LNG-IUS, levonorgestrel intra-uterine system; cTEG, computerized thrombelastograph; r, reaction time; k, coagulation time; mA, maximum amplitude. †post hoc Duncan test significance at 0.05. Not all women had complete blood sampling at intervals mentioned due mainly for not turning up at specified times.
Hemoglobin g L−1
Hematocrit L L−1
r time (min)
k time (min)
activity IU mL−1
antigen ng mL−1
activity IU mL−1
antigen ng mL−1
antigen ng mL−1
activity IU mL−1
antigen ng mL−1
Systemic hemostasis and fibrinolytic/inhibitor systems
There were no significant differences in the hemostastic parameters studied, except for a decreased soluble u-PAR level (P < 0.001) during the 6-month period of LNG-IUS use. The fibrinolytic/inhibitor systems were not affected, and similarly, cTEG, which indicates hemostasis status, was not influenced by LNG-IUS use. VWF, an endothelial cell component involved in inflammation and platelet aggregation, was also not affected. This suggests that systemic hemostasis was not influenced by LNG-IUS use in the cohort of women with menorrhagia, except for a depressed soluble u-PAR state. u-PAR is a receptor for u-PA fibrinolytic activation in tissue degradation and repair, and its decreased activity is further supported by non-elevated levels of D-dimer, a product of fibrin lysis during LNG-IUS use (Table 2).
The fibrinolytic/inhibitor systems in the endometrium following LNG-IUS use
No significant changes in t-PA or u-PA levels were seen in the 6 months of LNG-IUS use. In contrast, significantly elevated levels of soluble u-PAR, PAI-1 and PAI-2 were seen during use. Although premenstrual t-PA antigen median levels showed increased trends during use, they did not reach statistical significance. The greatly increased PAI-1 and PAI-2 levels would have inhibited fibrinolytic activity in the endometrium or degradation by u-PA activation, thus correcting MBL in menorrhagic women. The elevated u-PAR and PAI-1/2 levels were evident from the first month of treatment, and maximum levels were reached by 6 months (Table 3).
Table 3. Endometrial plasminogen activators/inhibitors and u-PA receptor following 6 months of LNG-IUS use in the treatment of menorrhagia. Median (inter-quartile range)
(per mg protein) n
t-PA, tissue-type plasminogen activator; u-PA, urokinase plasminogen activator; u-PAR, u-PA receptor; PAI-1, plasminogen activator inhibitor-1; LNG-IUS, levonorgestrel intra-uterine system. †post hoc Duncan test significance at 0.05. PAI-2 levels not detected at pre (n = 12), month 1 = (n = 5), month 2 = (n = 3) and month 3 = (n = 3). Not all women had endometrial aspiration done mainly due to not turning up at specified times and that amenorrhoeic women at endometrial aspiration did not reveal any tissue material for analysis.
The natural hormones estrogen and progesterone influence endometrial development and maturation during the menstrual cycle. Estrogen stimulates endometrial growth, whereas progesterone causes secretory changes that peak at the luteal phase of the menstrual cycle, culminating in the collapse of the endometrium and menstruation, if pregnancy has not occurred. Systemic hemostatic factors were not influenced by natural hormonal changes occurring during the various phases of the menstrual cycle of healthy Asian subjects . This suggested that hormonal influence is probably localized at the level of the endometrium . However, variable systemic hemostatic changes during the menstrual cycle were seen in Caucasians [27,28]. Plasminogen activators are synthesized by macrophages and leukocytes, and their numbers are known to increase in the endometrial environment during the premenstrual phase [29,30]. Furthermore, the levels of endometrial t-PA and its inhibitor, PAI-1, were also found to be highest during this period . Excessive MBL resulting in menorrhagia is seen in conditions such as uterine fibroids, adenomyosis, pelvic infections, endometrial polyps and clotting defects.
Use of the LNG-IUS is known to suppress endometrial proliferation and activity . Blood and endometrial samples in this study were obtained during the premenstrual phase, to observe the above hemostatic changes, if any, and their association with menorrhagia before and after treatment. Moreover, the sampling was performed premenstrually, with the aim of obtaining a sufficient amount of endometrial tissue for analysis. The effects of LNG-IUS use in suppression of endometrial proliferation during treatment, resulting in insufficient or unsuccessful extraction of tissues in patients who became amenorrhoeic, were not intended.
LNG-IUS is highly effective in the treatment of menorrhagia [4,8]. The exclusion criteria for hemostatic diathesis in the study showed that fibroids (51%), adenomyosis (39%) and DUB (10%) comprise the major cause of menorrhagia in the women under study. All the women had no problems with menorrhagia by 6 months of LNG-IUS use. The hemoglobin and hematocrit levels showed significant improvement during the 6 months of treatment. Hemostatic status as indicated by cTEG and plasminogen activator/plasminogen activator inhibitor levels was not significantly altered throughout the duration of LNG-IUS use, except for a systemic reduction of the soluble u-PAR level. The apparent reduction in systemic soluble u-PAR during LNG-IUS use is difficult to explain, except that the lower u-PAR level observed will result in reduced availability to form the u-PA complex for fibrinolytic activation. VWF levels showed no significant alteration that indicated non-endothelial cell damage or enhanced release or synthesis during the period of treatment.
In the endometrium, premenstrual plasminogen activators (t-PA and u-PA) were not altered during the treatment period; t-PA antigen median levels at 3 and 6 months showed a trend upwards, but this did not reach statistical significance. In contrast, the PAI-1/2 and u-PAR levels showed significant increases during LNG-IUS use, with the highest levels seen at 6 months. In vitro studies have demonstrated that upregulation of PAI-1  and u-PAR by progesterone in human endometrial stromal cells results in increased elimination of u-PA, and thus downregulation of pericellular proteolysis . The increased u-PAR levels seen in this study suggest increased availability to promote u-PA activation and degradation in the endometrial environment. This was unlikely, as the u-PA level was not elevated during LNG treatment. Moreover, the increased levels of PAI-1/2 would have inhibited any u-PA activation, and hence the correction of menorrhagia. The increased PAI-2 level seen could not have been influenced by progesterone treatment, and could possibly be due to the presence of the intrauterine device, as increased numbers of adherent macrophages have been reported with intrauterine device use , and because PAI-2 is also expressed by these macrophages .
Large numbers of macrophages or monocytes adhere to IUCDs, and higher cell counts were seen in menorrhagia and intermenstrual bleeding associated with significant fibrinolytic activity surrounding the device [17,18,33]. However, no increase was seen in association with the use of progesterone-releasing device . Activated monocytes release u-PA, which binds to its receptor (u-PAR) on the cell surface, and this complex is regulated by its specific inhibitor (PAI-2), which is also expressed by monocytes .
The effect of LNG in the endometriun in our study would suggest enhanced PAI expression in the presence of increased u-PA receptor levels. The slow LNG-release intrauterine device has antifibrinolytic properties in the endometrial environment, as shown by high level of expression of fibrinolytic inhibitors seen in this study. Furthermore, systemic hemostasis was not affected during LNG-IUS use. The use of LNG-IUS is highly effective in the treatment of women with menorrhagia with known pathologic causes.
We thank Y. F. Fong for assistance in recruitment and follow-up of some of the patients in the study. Expert technical assistance from S. E. Chua, W. K. Yuen and B. L. Ng is gratefully acknowledged.
Disclosure of Conflict of Interests
The study was supported by a grant from the National Medical Research Council (NMRC/0346/99).