The effect of a levonorgestrel-releasing intrauterine device on human endometrial oestrogen and progesterone receptors after one year of use

Zhu Pengdi, Liu Xiaoqun, Luo Hongzhi, Gu Zhao, Cheng Jie, Xu Ruhua, Lian Shizhu, Wu Shangchun and Wang Jiedong1

National Research Institute for Family Planning, Beijing, 100081, China


    Abstract
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Thirty-four women bearing a levonorgestrel-releasing intrauterine device, 20 µg/day (LNG-IUD-20), for 12–15 months were recruited. Endometrial biopsies were collected during the late proliferative phase of the cycle (on cycle days 10–12) before (control) and after the use of the IUD for 12 months, and assayed for oestrogen receptors (ER) and progesterone receptors (PR). An immunohistochemical technique with the peroxidase–antiperoxidase detection system (PAP method) was employed. D75 and JZB39 were the primary antibodies for ER and PR respectively. The immunostaining semiquantitative analysis was performed with a computerized microscope image processor, and expressed as `grey value'. Both endometrial ER and PR populations were significantly lower after insertion of the IUD (P < 0.01) than in control biopsies. The intensity of nuclear staining and the percentage of positively stained cells for ER and PR in women with LNG-IUD were each about 50% of those in control biopsies. The results suggested that LNG released locally from the IUD has a depressive action on the ER and PR, which may contribute to the contraceptive effectiveness of this type of IUD and also to the possible causes of LNG-IUD-induced irregular bleeding and amenorrhoea.

Key words: endometrium/immunohistochemistry/levonorgestrel-releasing intrauterine device/oestrogen receptor/progesterone receptor


    Introduction
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
A levonorgestrel-releasing intrauterine device (LNG-IUD) developed in Finland has been in wide clinical use for more than 15 years. Although its clinical performance, including long duration of action, low pregnancy rate, reduced menstrual blood loss and relief of symptoms of dysmenorrhoea is encouraging, some side effects such as intermenstrual spotting and amenorrhoea still cause problems for the users. Many studies have suggested that the mode of action and side effects of this IUD were, at least in part, based on the changes in endometrial function and morphology that are produced by the steroid liberated from the IUD (Nilsson et al., 1984Go; Barbosa et al., 1995Go; Gu et al., 1995Go; Luukkainen and Toivonen, 1995Go; Xiao et al., 1995Go).

The endometrium is the target of oestrogen and progesterone. The steroid hormone may not be directly responsible for the endometrial functional and morphological changes, as the hormonal actions not only coincide with the plasma hormone concentrations but also relate closely to the numbers of endometrial oestrogen receptors (ER) and progesterone receptors (PR). Some studies have demonstrated the effect of progestin-releasing intrauterine devices on steroid receptor expression in the endometrium (Janne and Ylostalo, 1980Go; Lu, 1991Go). These studies used radiochemical techniques to identify total ER and PR populations in whole endometrium, and did not permit identification of the receptors in particular cell types. In the present study we used immunohistochemistry to describe the cellular localization of ER and PR in the endometrium before and after insertion of a LNG-IUD. A semiquantitative analysis was also employed with a computerized microscope image processor, which was intended to render results more objective, the main aim of the study being to help in our understanding of the mechanism of the action and side effects of the LNG-IUD.


    Materials and methods
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Subject selection
Thirty-four healthy women, aged between 23 and 35 years, with proven fertility and regular menstrual periods were recruited for the study. The women had no known systematic or gynaecological disorder, had not received any hormonal therapy, steroid contraceptives or intrauterine device for at least three cycles, and had no history of abortion or pregnancy during the last 6 months before the start of the study. After having obtained written informed consent from each woman, a levonorgestrel-releasing intrauterine device (LNG-IUD) was inserted into each subject for a period of 12–15 months. The study was approved by the Ethical Committee of the National Research Institute for Family Planning, Beijing.

Intrauterine devices
The levonorgestrel-releasing IUD were provided by Leiras Company (Turku, Finland). Each device contains 46 mg levonorgestrel, and releases approximately 20 µg/day. The effective lifetime of the IUD is expected to be 7 years.

Sample collection
An endometrial biopsy was taken from the anterior or posterior wall of the mid uterus on cycle days 10–12, just before insertion of the IUD, and served as the control representing the proliferative phase of the normal menstrual cycle. One year later, immediately after removal of the IUD on cycle days 10–12, a second biopsy was taken. In women with amenorrhoea who bore the IUD, the device was removed on any day after 12 months of use, and a biopsy taken at the time of removal. No biopsy could be obtained in 16 cases because of endometrial atrophy.

Specimen preparation
A small portion of the endometrial specimen was fixed in Bouin's solution and paraffin sections were stained with haematoxylin and eosin (H&E) for morphology and dating. The major portion of the specimen was rinsed immediately in cold normal saline, frozen in liquid nitrogen, and stored at –70°C until analysed.

Immunohistochemistry
Monoclonal antibodies D75 to human ER and JZB39 to human PR were the generous gifts of Dr G.L.Greene, University of Chicago. A peroxidase–antiperoxidase method was employed (Wang et al., 1992Go), the frozen serial sections of 5 µm thickness being fixed in picric acid–paraformaldehyde for 15 min, then immersed in 0.5% hydrogen peroxide solution for 15 min to block endogenous peroxidase activity. After washing the sections in phosphate-buffered saline (PBS), primary antibodies (rat monoclonal antibody to human ER, rat monoclonal antibody to human PR respectively), secondary (bridge) antibody (rabbit anti-rat IgG), and rat peroxidase–antiperoxidase complex were applied to sections sequentially. Between each step the section was rinsed in PBS. The formation of an immunoprecipitate in the section was demonstrated with 3,3-diaminobenzidine (Serva, Heidelberg, Germany). One of the serial sections in each case was treated with normal rat IgG instead of the primary antibody to serve as a negative control. A section from receptor-positive tissue, processed using the same procedure, served as a positive control in each assay. No sections were counterstained, as this procedure is thought to reduce the sensitivity of the image analysis system in terms of recording the integrated optical density of the nuclei.

Analysis of endometrial immunoreactivity
In order to assess the effect of LNG-IUD (in terms of ER and PR), two types of scoring system were used.

Subjective score: the immunostaining intensity of all tissue sections was scored on a five-point scale, where 0 = no staining, 1 = mild, 2 = moderate, 3 = high and 4 = intensive staining. The recorded immunostaining score was based on the intensity of staining expressed by the majority of cells in each section. In brief, 3–10 areas were chosen randomly in each section according to the size of the sample. In each area (at x200 magnification), the percentage of stained cells of each scale was estimated by eye. The score of each area was obtained as {sigma}(IÏP), where I is the intensity of staining from 0–4 and P is the corresponding percentage. The mean value of all areas gives the score of the section. All samples were evaluated by the same observer to rule out interobserver variability.

Objective score: the immunostaining semiquantitative analysis was performed with a computerized microscope image processor. The method was described previously by Zhu et al. (1995). Sections were analysed in blind fashion (Olympus BH light microscope, x200 magnification). In each case, about 30 positive and negative nuclei at the level of stroma and glandular epithelium respectively, were evaluated and expressed as a `grey value'. The only selection criterion was the absence of superimposition of the nuclei to obtain the real optical density. The background was measured at the same time for each cell in the area immediately outside the cell, in order to eliminate batch or area differences in staining intensity. The score of each cell was expressed as background minus nuclear values.

Data management and statistical analyses
The grey values of the endometrial cells (stromal and glandular cells) in each subject were measured with minicomputer, VAXII, SAS-language Program. Statistical analysis was performed, results were evaluated with a t-test, and significant differences between the pre- and post-insertion data were calculated.


    Results
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Histologically, the cellular components of the endometrium could be clearly subdivided into glands and surrounding stroma cells comprising the vasculature. Samples taken before IUD insertion showed the morphology of the proliferative phase (Figure 1aGo) according to the criteria of Noyes et al. (1950) and Johannisson et al. (1987). Samples from the IUD users displayed a pattern of uniform suppression (Figure 2aGo), the glands being scarce in number and very small. The stroma showed pseudodecidual features.






View larger version (643K):
[in this window]
[in a new window]
 
Figure 1. Light micrographs of endometrial sections for the immunoreactivity of oestrogen receptors (ER) and progesterone receptors (PR) in IUD users. Before insertion, at cycle day 12. (a) A haematoxylin and eosin-stained section shows a normal late proliferative phase endometrial pattern. Epithelial cells and the stromal cells in the upper functionalis exhibit strong immunoreactivity for ER (b) and PR (c) with the specific nuclear staining using D75 and B39 as primary antibodies (PAP method). No specific cytoplasmic staining was observed. No immunoreactivity for PR or ER was detected in the vascular endothelial cell nuclei and smooth muscle cell nuclei. (d) Background for PR (ER not shown) in the control section. Scale bars = 100 µm.

 





View larger version (601K):
[in this window]
[in a new window]
 
Figure 2. Light micrographs of endometrial sections stained for the immunoreactivity of ER and PR in IUD users. Post-insertion; cycle day 12. (a) A haematoxylin and eosin-stained section showing typical uniform suppression of endometrium. A substantial reduction in both the intensity of nuclear staining and the proportion of cells stained for ER (b) and PR (c) in comparison with the control. Specific cytoplasmic staining was not observed. The endometrial endothelial cells and vascular myometrial cells remained unstained for ER and PR, as in the control. (d) Background for PR (ER not shown) was shown in the control section by using normal non-immunized rat immunoglobulin in place of monoclonal anti-receptor antibody, B39.

 
Immunoreactivity of endometrial ER and PR in control samples
As shown in Figure 1b and cGo, both ER and PR were localized in the nuclei of epithelial and stromal cells. Luminal and glandular epithelial cells showed strong staining in nearly all of their nuclei, and stromal cells were strongly stained in most nuclei. The glandular cells were slightly more heavily stained, the percentage of stained cells being greater than that in stromal cells. The intensity and percentage of cell nuclei staining for PR were greater than those for ER (Figure 1cGo; Table IGo). Many of the nuclei contained a small focal area, interpreted as a nucleolus, which was unstained. In contrast to the densely stained nuclei, the cytoplasm and intercellular spaces were unstained. Endometrial vascular smooth muscle cell nuclei and endothelial cell nuclei were consistently unstained. Results obtained from the two scoring systems were parallel (Table IGo). The percentages of stained cells for ER and PR were not shown.


View this table:
[in this window]
[in a new window]
 
Table I. The immunoreactivity score of endometrial cells for oestrogen and progesterone receptors (arbitrary units) before and after insertion of LNG-IUD-20. Values are mean ± SD
 
Immunoreactivity of endometrial ER and PR in LNG-IUD users
By using a series of adjacent sections, it was shown that ER and PR immunostaining remained in the nuclei of the endometrial cells, as before. There was a substantial reduction (P <0.01) in both the intensity of nuclear staining and the proportion of cells stained for ER and PR (Figure 2bGo–d) in comparison with the control (Figure 1bGo–d). The immunostaining in most cases was almost eliminated, with only occasional isolated cells showing a variable amount of staining. However, these occasional isolated cells were quite strongly stained and contrasted sharply with the minimal staining in adjacent cells. The nuclear staining intensity and percentage of stained cells in the gland and stroma were similar (P >0.05), but the stromal cells were slightly more intensely stained and the percentage of stained cells was higher than that of the glandular cells only for ER. During this post-insertion period, the intensity and percentage of cell nuclear staining for PR remained slightly greater than that for ER as in the pre-insertion period, though no significant difference was found (Table IGo). The endometrial vascular cells remained unstained, as in the controls.


    Discussion
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
A number of studies have defined the location and distribution of ER and PR in the endometrium (Lessy et al., 1988Go; Snijders et al., 1992Go; Bergqvist et al., 1994). The expression of ER and PR in the endometrium changes cyclically, although some differences exist among the reports. In general, ER and PR are located in the nuclei and the positive rate and degree of expression peak during the mid cycle and then decrease. Progestins downregulate this expression (Natrajan et al., 1982Go). In this study, using an immunohistochemical method, we showed that the staining for endometrial ER and PR in LNG-IUD users was rather similar to the pattern of the mid to late secretory phase of the normal menstrual cycle reported previously (Lessy et al., 1988Go; Snijders et al., 1992Go; Bergqvist et al., 1993). There was a substantial reduction in both the intensity of the nuclear staining and the proportion of the cells stained for ER and PR in comparison with controls. The released LNG caused the downregulation of both ER and PR, which is in agreement with other studies as the populations of both endometrial oestrogen and progesterone cytoplasmic receptors were reduced significantly after insertion of a LNG-IUD (Janne and Ylostalo, 1980Go; Lu, 1991Go).

However, the LNG which is released locally from the IUD does not mimic the effect of physiological progesterone on the endometrium. The endometrium of LNG-IUD users displayed a pattern of uniform depression. In the IUD users, the endometrial steroid receptor population was uniformly decreased independent of the cell types, while in the normal cycle it was cell-specific, there being a major difference between epithelial and stromal cells. Interestingly, Critchley et al. (1993) reported that the immunoreactive PR population remained high in the endometrium of Norplant users, whereas the population of ER was relatively low. These findings are significantly different from those of the present study, and may suggest that the effect of LNG released locally on the endometrium is different from that of circulating LNG. Moreover, Lau et al. (1996) showed an increase in immunoreactive PR population, but a reduction in PR mRNA levels in Norplant endometrium, while in the normal menstrual cycle the protein is consistent with the mRNA. It would be interesting to investigate the endometrial expression of isoforms of ER and PR in both LNG-IUD users and Norplant users.

In this study, the endometrial blood vessels were not stained in either controls or after insertion of the LNG-IUD. However, conflicting data have been obtained regarding this issue. Lessy et al. (1988) demonstrated no immunostaining at all, whereas others reported the presence of both ER and PR in vascular smooth muscle cells of endometrial blood vessels (Perrot-Applanat et al., 1988Go, 1994Go). A earlier study in this laboratory reported immunostaining for PR in the endothelial cells from human decidua (Wang et al., 1992Go), while Perrot-Applanat et al. (1988, 1994) took tissue at the end of the luteal phase or during pregnancy. It is possible that endometrial blood vessels express the receptors when endometrium becomes decidualized. The endometrium of LNG-IUD users also showed a pseudodecidual pattern, but this was uniformly and heavily depressed, including the vessels which were small with a thin wall that was only single-layered in most cases. In addition, combining results of the present study with those reported previously (Lessy et al., 1988Go; Snijders et al., 1992Go; Critchley et al., 1993Go; Bergqvist et al., 1994; Lau et al., 1996Go), it is likely that locally released LNG affects the endometrium in a different manner from other circulating progestins. However, Rogers et al. (1996) demonstrated highly varied ER and PR immunoreactivity in the vascular smooth muscle cells of endometrial blood vessels, ranging from 0% to 85%, with a tendency towards greater immunoreactivity in the late secretory phase. We could not find an adequate explanation for the difference between our results and those of Rogers et al. (1996), who identified receptors as being present in blood vessels throughout the menstrual cycle, albeit with high variability. However, the fact that the antibodies used in these studies were different might provide an explanation.

Many studies suggested that the changes in endometrial function and morphology are caused mainly by the local effect of progestin on the endometrium (Nilsson et al., 1984Go; Luukkainen and Toivonen, 1995Go). Previous studies of ovarian function in LNG-IUD users have suggested that the majority of women studied had ovulatory cycles (Nilsson et al., 1984Go; Barbosa et al., 1995Go; Xiao et al., 1995Go). The intrauterine release of levonorgestrel results in high local tissue concentration in the endometrium (Nilsson et al., 1982Go). Oral administration of a dose 10-fold higher than the daily dose administered with LNG-IUD resulted in endometrial concentrations of levonorgestrel which were similar to those in adjacent tissues, and the effects on the endometrium were correspondingly much weaker than with locally released LNG (Nilsson et al., 1982Go). Many LNG-IUD users—though none of the LNG intracervical device users—developed amenorrhoea, which confirms the local effect of LNG-IUD (Kurumaki et al., 1984Go).

Several studies have indicated that endometrial morphology in women bearing a LNG-releasing IUD shows profound and uniform suppression which is independent of the stage of the menstrual cycle (Zhu et al., 1989Go; Gu et al., 1995Go). Locally released levonorgestrel causes atrophy of the endometrium, such that the glands are scarce and the stroma shows pseudodecidual features. This effect on the endometrium is seen as early as 1 month after insertion, and is maintained as long as levonorgestrel is released. The mechanism is likely to be a complex process (Hsueh et al., 1976Go), and many factors may be involved. For example, the decidualized cells produce insulin-like growth factor binding protein-1, which can block insulin-like growth factor I that is thought to be a mediator of oestrogen-induced mitotic effects (Pekonen et al., 1992Go). From the present data, we might suggest that the continuous high concentration of locally released LNG inhibits the expression of ER, which in turn makes the endometrium insensitive to circulating oestradiol and provokes an anti-proliferative effect during the use of LNG-IUD. Thus, endometrial proliferative activity is totally arrested, and this is thought to be the main contraceptive action of the IUD.

Taken together, the results of the present study indicate that there is a statistically significant decrease of ER and PR in the endometrium of LNG-IUD users, when compared with controls, and confirms that progesterone downregulates its receptor. However, the decrease of ER and PR populations in LNG-IUD users was somewhat uniform, and not cell-specific—a situation which is different from that in the normal menstrual cycle (Lessy et al., 1988Go; Snijders et al., 1992Go; Bergqvist et al., 1994). In addition, the immunoreactivity of ER and PR in the LNG-IUD endometrium is significantly different from that in the Norplant endometrium (Critchley et al., 1993Go; Lau et al., 1996Go), though it is unclear why such differences exist. However, the data acquired in the present study show that 16 out of 34 subjects failed to produce a tissue sample—a proportion similar to that seen in Norplant users (Hadisputra et al., 1996). As, therefore, data are unavailable from these patients, the samples that were collected may not be truly representative, and so care must be taken when interpreting our results.


    Acknowledgments
 
The authors are highly appreciative of the gifts of LNG-IUD and anti-receptor monoclonal antibodies from Prof. Luukkainen of Leiras Company, Turku, Finland, and Dr G.L.Greene of Ban May Institute, University of Chicago, USA. We are also very grateful to Dr Elisabeth Johannisson, LACQ, Geneva for scientific advice and revision of the manuscript. The present study was supported by WHO Special Programme of Research, Development and Research Training in Human Reproduction, Geneva, Switzerland.


    Notes
 
1 To whom correspondence should be addressed Back


    References
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Barbosa, I., Olsson. S.E., Odlind, V. et al. (1995) Ovarian function after seven years' use of a levonorgestrel IUD. Adv. Contracept., 11, 85–95.[ISI][Medline]

Bergqvist, A. and Ferno, M. (1993) Oestrogen and progesterone receptors in endometriotic tissue and endometrium: comparison of different cycle phases and ages. Hum. Reprod., 8, 2211–2217.[Abstract]

Critchley, H.O.D., Bailey, D.A., Au, C.L. et al. (1993) Immunohistochemical sex steroid receptor distribution in endometrium from long-term subdermal levonorgestrel users and during the normal menstrual cycle. Hum. Reprod., 8, 1632–1639.[Abstract]

Gu, Z., Zhu, P.D., Luo, H.Z. et al. (1995) A morphometric study on the endometrial activity of women before and after one year with LNG-IUD in situ. Contraception, 52, 57–61.[ISI][Medline]

Hadisaputra, W., Affandi, B., Witjaksono, J. and Rogers, P.A.W. (1996) Endometrial biopsy collection from women receiving Norplant. Hum. Reprod., 11 (suppl. 2), 31–34.[Medline]

Hsueh, A.J., Peck, E.J., Jr and Clark, J.H. (1976) Progesterone antagonism of the estrogen receptor and estrogen-induced uterine growth. Nature, 254, 337–338.[ISI]

Janne, O. and Ylostalo, P. (1980) Endometrial estrogen and progestin receptors in women bearing a progesterone-releasing intrauterine device. Contraception, 22, 19–23.[ISI][Medline]

Johannisson, E., Landgren, B.M., Rohr, H.P. and Diczfalusy, E. (1987) Endometrial morphology and peripheral hormone levels in women with regular menstrual cycles. Fertil. Steril., 48, 401–408.[ISI][Medline]

Kurumaki, H., Toivonen, J., Lahteenmaki, P. and Luukkainen, T. (1984) Pituitary and ovarian function and clinical performance during the use of a levonorgestrel-releasing intracervical contraceptive device. Contraception, 29, 31–43.[ISI][Medline]

Lau, T.M., Witjaksono, J. and Rogers, P.A.W. (1996) Progesterone receptor in Norplant endometrium. Hum. Reprod., 11 (suppl. 2), 90–94.[Medline]

Lessy, B.A., Killam, A.P., Metzger, D.A. et al. (1988) Immunohistochemical analysis of human uterine estrogen and progesterone receptors throughout the menstrual cycle. J. Clin. Endocrinol. Metab., 67, 334–340.[Abstract]

Lu, S.M. (1991) Effect of levonorgestrel intrauterine device on human endometrial estrogen and progesterone receptors. Chin. J. Obstet. Gynecol., 26, 293–294.

Luukkainen, T. and Toivonen, J. (1995) Levonorgestrel-releasing IUD as a method of contraception with therapeutic properties. Contraception, 52, 269–276.[ISI][Medline]

Natrajan, P.K., Muldoon, T.G., Greenblatt, R.B. and Mahesh, V.B. (1982) The effect of progestins on estrogen and progesterone receptors in the human endometrium. J. Reprod. Med., 27, 277–230.

Nilsson, C.G., Hankkamaa, M., Vierola, H. and Luukkainen, T. (1982) Tissue concentrations of levonorgestrel in women using a levonorgestrel-releasing IUD. Clin. Endocrinol. (Oxford), 17, 529–536.[ISI][Medline]

Nilsson, C.G., Lahteenmaki, P. and Luukkainen, T. (1984) Ovarian function in amenorrhoeic and menstruating users of levonorgestrel-releasing intrauterine devices. Fertil. Steril., 41, 52–55.[ISI][Medline]

Noyes, R.W., Hertig, A.T. and Rock, J. (1950) Dating the endometrial biopsy. Fertil. Steril., 1, 3–25.[ISI][Medline]

Pekonen, F., Nyman, T., Lahteenmaki, P. et al. (1992) Intrauterine progestin induces continuous insulin-like growth factor-binding protein production in the human endometrium. J. Clin. Endocrinol. Metab., 75, 660–664.[Abstract]

Perrot-Applanat, M., Groyer-Picard, M.T., Garcia, E. et al. (1988) Immunocytochemical demonstration of oestrogen and progesterone receptors in muscle cells of uterine arteries in rabbits and humans. Endocrinology, 123, 1511–1519.[Abstract]

Perrot-Applanat, M., Deng, M., Fernandez, H. et al. (1994) Immuno- histochemical localisation of estradiol and progesterone receptors in human uterus throughout pregnancy: expression in endometrial blood vessels. J. Clin. Endocrinol. Metab., 78, 216–224.[Abstract]

Rogers, P.A.W., Lederman, F., Kooy, J. et al. (1996) Endometrial vascular smooth muscle oestrogen and progesterone receptor distribution in women with and without menorrhagia. Hum. Reprod., 11, 2003–2008.[Abstract]

Snijders, M.P.M.L., de Goeij, A.F.P.M., Debets-Te Baerts, M.J.C. et al. (1992) Immunocytochemical analysis of oestrogen and progesterone receptors in the human uterus throughout the menstrual cycle and after menopause. J. Reprod. Fertil., 94, 363–371.[Abstract]

Wang, J.D., Fu, Y., Zhu, P. et al. (1992) Immunohistochemical localization of progesterone receptor in human decidua of early pregnancy. Hum. Reprod., 7, 123–127.[Abstract]

Xiao, B.L., Zeng, T., Wu, S.C. et al. (1995) Effect of levonorgestrel-releasing intrauterine device on hormonal profile and menstrual pattern after long term use. Contraception, 51, 319–365.[ISI][Medline]

Zhu, P., Luo, H., Xu, R. et al. (1989) The effect of intrauterine device, stainless steel ring, the copper T220, and releasing levonorgestrel, on the bleeding profile and the morphological structure of the human endometrium: a comparative study of three IUDs. A morphometric study of 96 cases. Contraception, 40, 425–438.[ISI][Medline]

Zhu, P.D., Liu, X.Q., Wang, J.D. et al. (1995) Analysis of endometrial Factor VIII activity in normal proliferative phase by digital image processing system. Contraception, 52, 23–27.[ISI][Medline]

Submitted on April 27, 1998; accepted on December 23, 1998.