The evaluation of the effectiveness of an intrauterine-administered progestogen (levonorgestrel) in the symptomatic treatment of endometriosis and in the staging of the disease

F.B. Lockhat, J.O. Emembolu and J.C. Konje1

Department of Obstetrics and Gynaecology, Robert Kilpatrick Clinical Sciences Building, Leicester Royal Infirmary, Leicester LE2 7LX, UK

1 To whom correspondence should be addressed. e-mail: jck4{at}le.ac.uk


    Abstract
 Top
 Abstract
 Introduction
 Subjects and methods
 Results
 Discussion
 References
 
BACKGROUND: Medical treatment of endometriosis, a condition which significantly affects the quality of life in ~10–15% of women in the reproductive years, remains problematic. Although oral progestogens are effective and cheap, their efficacy is significantly influenced by poor compliance and systemic side effects. A progestogen (levonorgestrel) administered via an intrauterine system (Lng-IUS) has been demonstrated to improve symptoms of endometriosis; however, its effect on the staging of the disease has not been evaluated. The aims of this study were therefore to investigate the effectiveness of Lng-IUS in the symptomatic relief of minimal to moderate endometriosis and in the staging of the disease. METHODS: This was a prospective non-comparative observational study in which 34 women with clinically suspected and laparoscopically confirmed symptomatic minimal to moderate endometriosis had Lng-IUS inserted for 6 months. The symptom profile and stage of the disease before, during and after 6 months of treatment and patients’ satisfaction with treatment and willingness to retain the device at the end of the study period were used to assess response to treatment. RESULTS: Of the 34 women recruited, 29 (85%) completed the study; five discontinued, for personal reasons (one), side effects of worsening of acne (one) and lower abdominal/ pelvic pain (three). Significant (P < 0.05) improvements in severity and frequency of pain and menstrual symptoms as well as staging were achieved, with 68% (23) of cases electing to continue with the device after 6 months of therapy. CONCLUSION: The levonorgestrel intrauterine system is an effective hormonal option for treating symptomatic endometriosis (minimal to moderate). It also alters the American Fertility Society staging of disease. With a continuation rate of 68% after 6 months, it has the potential for providing long-term therapy in a substantial number of sufferers, although this would require further study and verification.

Key words: endometriosis/levonorgestrel-releasing intrauterine system/mirena


    Introduction
 Top
 Abstract
 Introduction
 Subjects and methods
 Results
 Discussion
 References
 
Endometriosis affects 6–20% of women of reproductive age (Sangi-Haghpeykar and Poindexter, 1995Go; Witz, 1999Go; Overton et al., 2002Go). Despite many theories, the exact aetio-pathogenesis is unknown; however, the disease is known to be estrogen dependent (Henderson and Studd, 1991Go). Medical treatment, which is predominantly palliative, for symptoms commonly of dysmenorrhoea, dyspareunia, non-cyclical pelvic pain and/or menorrhagia, is hormonally based (Vercellini et al., 1997Go; Overton et al., 2002Go). Amongst the therapeutic options are anti-estrogens (e.g. danazol), and regimens that induce either a medical menopause (e.g. GnRH agonists) or a pseudo-pregnant state (e.g. continuous combined oral contraceptive or progestogens) (Luciano et al., 1988Go; Morghissi, 1999Go). Although these drugs are effective, systemic side effects commonly affect compliance or preclude long-term use, and the need for regular administration may further result in poor compliance, which undermines efficacy. This is more so with oral progestogens which, although cheap and efficacious, are associated with poorly tolerated side effects of irregular bleeding, weight gain/fluid retention, seborrhoea and breast tenderness. More recently, concerns have also been raised over the possible effects of long-term use of systemic progestogens on bone metabolism (Scholes et al., 2002Go).

The levonorgestrel intrauterine system (Lng-IUS) MirenaTM provides an alternative means of administering progestogens. It delivers levonorgestrel (a 19-C progestogen) into the uterine cavity at a steady rate of 20 µg/day over its 5-year lifespan. The systemic levels following such administration are less than those achieved with therapeutic oral or parenteral doses of progestogens (Nilsson et al., 1980Go; Luciano et al., 1988Go; Du et al., 1999Go; Morghissi, 1999Go; Fedele et al., 2001Go), hence side effects should theoretically be less severe. Its effects are predominantly localized to the endometrium where the high concentrations of levonorgestrel induce atrophy and pseudo-decidualization (Nilsson et al., 1978Go; Silverberg et al., 1986Go; Maruo et al., 2001Go). It is this action on the endometrium that enables the Lng-IUS to be used as a highly effective intrauterine contraceptive device and has popularized its use in the management of menorrhagia (Andersson et al., 1990Go; Irvine et al., 1998Go). More recently, its role in protecting the endometrium has been advocated in women on estrogen-only hormone replacement therapy or tamoxifen (Gardner et al., 2000Go). Its role in extrauterine pathology such as endometriosis is uncertain, as the levels of levonorgestrel reaching the peritoneal fluid to potentially affect these lesions are unknown.

In a cohort of 20 women with known endometriosis who had previously undergone conservative surgical treatment and had recurrent dysmenorrhoea, Vercellini et al. (1999Go) demonstrated that the Lng-IUS was associated with symptomatic relief in ~70% of cases after 12 months therapy. Fedele et al. (2001Go) also demonstrated a significant symptom improvement as well as radiological evidence of lesion regression in 11 women with recto-vaginal endometriosis. Neither of these studies related the changes in the symptoms to the stage or score of the disease. The aims of this pilot observational study were therefore to investigate the efficacy of the Lng-IUS in the symptomatic relief of endometriosis, and the changes in the staging of the disease induced by the Lng-IUS.


    Subjects and methods
 Top
 Abstract
 Introduction
 Subjects and methods
 Results
 Discussion
 References
 
Women with known or suspected symptomatic endometriosis and who initially met the inclusion criteria were recruited into the study prior to laparoscopy, after giving signed informed consent. The exclusion criteria included, age <18 years or >45 years, any hormonal therapy in the preceding 3 months, a desire to conceive in the ensuing 6 months, infertility as a symptom, a clinical history of pelvic inflammatory disease (PID), other contra-indications to the use of the IUS, laparoscopically diagnosed severe endometriosis (American Fertility Society, 1985Go) and refusal of consent to undergo a second-look laparoscopy 6 months after entering the trial. The local ethics committee approved the study protocol.

At the time of recruitment in the clinic, each subject marked the severity of their pelvic pain on a 10 cm visual analogue scale (VAS) and also rated the severity of pelvic pain (dysmenorrhoea and/or non-cyclical pelvic pain) on a 4-point verbal rating scale (VRS) which has well recognized advantages and drawbacks (Karoly, 1987Go). Other associated symptoms were also noted. They then completed a 1 month (28 day) daily symptom diary, for pelvic pain using a 4-point VRS, and menstrual loss using a pictorial blood loss assessment chart (PBAC) (Higham et al., 1990Go). The former was used to generate a monthly pain score (where each day was given a score 0–3 and the monthly score was the sum of the daily scores for 28 consecutive days) and the latter to generate a bleeding score (>100 suggestive of menorrhagia).

All the subjects underwent a standard two-port diagnostic laparoscopy to confirm the diagnosis of endometriosis. Where possible, peritoneal fluid was immediately aspirated from the pouch of Douglas (for a later quantification of levonorgestrel), after which the disease was then staged systematically with video documentation. A single punch biopsy of ectopic endometrium (endometriosis) was obtained where possible for future study on the effects of levonorgestrel on steroid receptor expression. The MirenaTM was then inserted following standard aseptic techniques.

Follow-up visits were at 1, 3 and 6 months after the insertion. Severity of pain was re-evaluated at these visits using the VAS and VRS. Due to irregularity of menstrual loss, these data were timed to duration of insertion of the IUS rather than day of menstrual cycle. Each subject also kept a symptom diary identical to that pre-laparoscopy; these were completed for months 1, 3 and 6 after the insertion. After 6 months, satisfaction with treatment was assessed using a 4-point VRS.

Six months after the insertion of the device, a second-look laparoscopy was performed, again with video documentation, systematic re-staging, collection of peritoneal fluid and biopsies. At this point, those who requested had the device removed.

The outcome measures included severity and frequency of pain (which was assessed by using a VAS, a VRS for dysmenorrhoea and non-cyclical pelvic pain, a monthly pain score and number of pain days per month), amount and frequency of bleeding (assessed using a monthly bleed score and number of bleeding days per month) and American Fertility Society staging and score of the disease.

Data analysis
SPSS version 11.0 was used to record and statistically analyse data. Values at the time of insertion of the IUS (i.e. time zero) were compared with those at different time points after insertion using paired t-test or Wilcoxon rank analysis as appropriate. Additionally, perception by the patient of the efficacy of the device in pain control was evaluated at 1, 3 and 6 months using a VAS, as well as overall satisfaction with treatment (taking into account the undesirable side effects) as indicated on a 4-point VRS.


    Results
 Top
 Abstract
 Introduction
 Subjects and methods
 Results
 Discussion
 References
 
Thirty-seven women were recruited into the study over a 2-year period (2000–2002). All had symptoms suggestive of endometriosis, e.g. secondary dysmenorrhoea, non-cyclic pelvic pain, deep dyspareunia, menorrhagia or recurrent symptoms with a known history of endometriosis. None were being investigated for or presented with infertility. Three cases were excluded at the initial laparoscopy because: of Fitz–Hughes–Curtis syndrome (which was taken to indicate previous Chlamydia trachomatis infection) (one) and severe disease requiring surgical intervention (two). The mean age of the remaining 34 subjects in whom the device was inserted was 31 ± 7.2 years (range 18–42); 50% were known cases of endometriosis.

The device was removed in five cases prior to the end of the trial period. These cases were not included in paired outcome analyses. Table I shows the indications for and the timing of all removals within and following the study period. Of the 29 women completing the 6 months, one was excluded from the pain analyses as the potent analgesics she had received during a significant part of the trial period due to her involvement in a road traffic accident were considered to have the potential to mask pains from endometriosis.


View this table:
[in this window]
[in a new window]
 
Table I. Indication for and timing of MirenaTM removal
 
Figure 1 shows the changes in the mean VAS for pain severity during the study period. There was a statistically significantly fall from 7.7 ± 1.3 cm pre-therapy to 6.1 ± 2.4 cm at 3 months post-IUS insertion (P < 0.01) and to 4.6 ± 3.0 cm after 6 months (P < 0.01). The VRS for dysmenorrhoea was also significantly different, with the proportion of patients experiencing moderate or severe dysmenorrhoea falling from 96% (27) before therapy to 68% (19) at 3 months (P = 0.001) and to 50% (14) at 6 months (P < 0.001). Non-cyclical pelvic pain, however, was not significantly reduced over the 6-month period (P = 0.087).



View larger version (12K):
[in this window]
[in a new window]
 
Figure 1. Changes in mean VAS (cm) for severity of pain over the 6-month trial.

 
The number of days of pain experienced per month (28 days) was significantly reduced from a mean of 15.0 ± 6.9 to 10.7 ± 8.7 days, after 6 months therapy (P < 0.05). From a maximum possible score of 84 per month, the mean score for pain per month (generated from the daily VRS) dropped from 25 (±13.8) pre-insertion to 19 (±18.9) 6 months later, but this did not reach statistical significance (P = 0.076). Of the 20 (69%) cases experiencing dyspareunia at the start of the study, 13 (45%) felt this had improved after 6 months therapy, two (7%) felt it had worsened, two (7%) felt there had been no change and three (10%) were unsure as they were not sexually active by the end of the 6-month trial period.

The mean quantified menstrual loss from the PBAC chart was 204 ± 196 pre-insertion (suggestive of significant menorrhagia in most of the cases) compared with 129 ± 273 at 3 months post-insertion (P > 0.05) and 90 ± 157 at 6 months post-insertion (P < 0.001). The mean number of bleeding days per month by the end of the sixth month was not significantly different from that pre-insertion (8.7 ± 7.5 days versus 7.6 ± 3.4, P = 0.519). Tables II and III show the changes in the quantified monthly blood loss and number of bleeding days respectively, against patient satisfaction, over the study period. The blood loss in the dissatisfied and uncertain groups was more prolonged, with a smaller quantitative reduction after 6 months, when compared with those who were satisfied or very satisfied.


View this table:
[in this window]
[in a new window]
 
Table II. Quantified monthly blood loss over the 6-month study versus overall satisfaction
 

View this table:
[in this window]
[in a new window]
 
Table III. Number of days bleeding per month (28 days) before, during and after 6 months use of MirenaTM
 
Twenty-six of the 29 women who completed the study underwent a repeat laparoscopy; one refused the procedure for personal reasons, one relocated to another country and the third failed to attend. The mean scores for endometriosis in the 26 patients were 12.2 ± 10.9 and 10.7 ± 11.6 at the time of the initial laparoscopy and 6 months after the IUS insertion, respectively. The American Fertility Society stage of the disease improved in 30.8% of cases after 6 months therapy, with the difference in staging reaching statistical significance (P < 0.05) in these patients. Table IV shows the differences in staging pre- and post-therapy. An improvement in symptoms was not necessarily associated with changes in the staging. Overall, a lower score after treatment with the IUS for 6 months was demonstrated in 62% (16) of cases, although, due to the weighting of points, this did not always alter staging.


View this table:
[in this window]
[in a new window]
 
Table IV. American Fertility Society staging pre- and post-Lng-IUS therapy versus satisfaction with treatment
 
The changes in pain severity (as indicated by the VAS) versus changes in American Fertility Society scoring are shown in Table V. The patients’ perceptions of the efficacy of the IUS in pain relief as assessed on a 10 cm VAS (where 0 cm = nil effect and 10 cm = extremely good) were 5.3 ± 2.9 cm, 7.1 ± 2.5 cm and 7.4 ± 2.4 cm after 1, 3 and 6 months of therapy, respectively. There was a significant change in perception from 1 month to 3 months; this persisted at 6 months (P < 0.01). At the end of the 6 months, two (6.9%) of the 29 who completed the study (including those not undergoing repeat laparoscopy) were dissatisfied with treatment, eight (27.6%) were uncertain and 19 (65.5%) were either satisfied or very satisfied. At the second-look laparoscopy, six (20.7%) women elected to have the device removed; two for side effects, although they had reported an improvement in endometriosis-related symptoms, and the other four for persistent unchanged symptoms with or without side effects. A functional ovarian cyst (defined as a cyst of >2 cm in diameter) was identified in three (11.6%) cases at the second-look laparoscopy; none of these were symptomatic.


View this table:
[in this window]
[in a new window]
 
Table V. Change in American Fertility Society scoring in relation to VAS (pain intensity) following 6 months use of MirenaTM
 

    Discussion
 Top
 Abstract
 Introduction
 Subjects and methods
 Results
 Discussion
 References
 
Although our series is small, it is larger than those previously studied (Vercellini et al., 1999Go; Fedele et al., 2001Go) and the results confirm that the Lng-IUS is beneficial in some women with endometriosis experiencing pain and menorrhagia. In assessing response to treatment, account has to be taken of the difficulties and limitations in objectively measuring pain (Linton and Melin, 1982Go). It is recognized that the VRS and to a lesser extent the VAS still rely on patients’ subjective perception and recall, but in a longitudinal clinical study like ours, this provides a reasonably reliable assessment of symptoms response. Overall, the severity of pain was significantly reduced as indicated by the VAS, as was the frequency of pain as indicated by the number of days of pain per month. Whilst relief of dysmenorrhoea was achieved, a finding similar to that of Vercellini et al. (1999Go), non-cyclical pelvic pain was not significantly reduced, contrary to the findings of Fedele et al. (2001Go). This variation is possibly because the cohort of women in the study of Fedele et al. (2001Go) all had recto-vaginal endometriosis, which may reflect a different aetio-pathology and symptom mechanism to that of peritoneal disease, which was the predominant disease type in our series.

This modality of treatment confers several advantages over other conventional systemic forms of therapy (avoidance of the need for repeated administration, delivery of a steady amount of levonorgestrel, effective contraception and fewer systemic side effects). Although it may be expensive at the outset, the cumulative long-term costs are significantly less than those of other forms of medical therapy, such as GnRH analogues and danazol. Side effects were often transient and generally well tolerated; the side effects profile was similar to that previously described for the device (Backman et al., 2000Go). The PBAC scores reflect that most of the women had significant menorrhagia (defined by a score of >100) prior to therapy, which was effectively reduced after 6 months treatment. However, those experiencing prolonged and/or heavy bleeding persisting after 6 months of use were less likely to be satisfied and therefore potentially more likely to discontinue therapy, even if their symptoms improved, as is the case in those using the device as a contraceptive (Backman et al., 2000Go). In our series, three of the patients discontinuing treatment (either during or after the trial period) did so for side effects (which included exacerbation of acne, weight gain and headaches), even though there was marked improvement in symptoms of pain and menorrhagia.

We found an improvement in the staging of the disease after 6 months of therapy (suggesting that whatever mechanism is involved, the lesions themselves are affected), an observation which has not been documented before. The actual significance of this in relation to symptoms is uncertain as it is widely accepted that clinical symptoms do not necessarily correlate with the extent or distribution of lesions. Staging is based on an arbitrary scoring system that was devised for prognostication of infertility rather than treatment and symptom response, and hence has a greater weighting/emphasis on the presence of adhesions rather than peritoneal endometriotic lesions. It was largely the peritoneal endometriotic lesions which appeared to respond to therapy whilst, as with all medical regimens, the presence of adhesions was not altered. Consequently, scores and hence staging may not have been drastically altered following treatment, despite a marked visible reduction in the extent of haemorrhagic and neovascular peritoneal lesions (thought to represent active disease) and a marked improvement in symptoms.

The exact mechanism by which this device is effective is uncertain, but we believe that this may be achieved via both local and systemic routes. The hypomenorrhoea which many of these patients experienced is likely to be a combination of atrophy of the endometrium (local activity) and ovarian suppression (systemic activity). It is recognized that in the first 3 months on the device, up to 85% of women have anovulatory cycles and by 12 months this figure falls to <35% (Jarvela et al., 1998Go). If anovulation were the only mechanism responsible for symptom relief, then we would expect a gradual recurrence and worsening of symptoms in those retaining the device after the 6 months, as ovarian activity resumes and endogenous hormone levels increase. However, both Fedele et al. (2001Go) and Vercellini et al. (1999Go) noted a persistent symptom improvement at 12 months. In our series (data not presented), >80% of those who elected to continue with the Lng-IUS had no significant exacerbation of symptoms after 12 months therapy, suggesting that another mechanism contributes to the efficacy of the IUS in symptom relief. This second mechanism is likely to be local. Although it may be suggested that the Lng-IUS may alter uterine perfusion and decrease pelvic congestion that may contribute to symptom relief, studies on uterine vascular resistance and impedance to flow in the presence of the Lng-IUS have failed to support this. It is suggested that decreased pelvic congestion may be mediated by systemic levels of levonorgestrel (>200 pg/ml) via receptors on uterine vasculature and, if this were the case, the effect may only last for 3 months after insertion as the systemic levels decline thereafter. Additionally, this vascular impedance is only evident in the mid-luteal phase, and no differences have been noted at menstruation (Jarvela, 1998Go).

Symptom improvement may therefore be due to a combination of menstrual interruption, disruption of follicular activity and a direct effect on the endometriotic lesions. Local levels of levonorgestrel in the peritoneal fluid may facilitate this direct mechanism. Unfortunately, these peritoneal fluid levonorgestrel levels are unknown, as they have never been measured. We are currently measuring these levels in addition to studying the effects of levonorgestrel on ovarian steroid receptor expression on endometriotic lesions. If the concentration of the levonorgestrel in the peritoneal fluid is high, then it may alter the steroid receptor expression on lesions (endometriotic receptor-mediated response) in a manner similar to endometrium (Silverberg et al., 1986Go). Alternatively, the progestogen may affect the peritoneal fluid macrophage activity, thus altering the production of various cytokines and factors that are responsible for the maintenance of lesions and symptoms (cytokine-mediated macrophage response); this may also be receptor mediated (McLaren et al., 1996Go). Finally, it may be possible that both systemic and local levonorgestrel downregulate or alter the local gene expressions associated with proliferation of lesions and progression of the disease.

Sixty-eight percent (23 out of 34) of the cases recruited into our study chose to continue with the device after the 6-month trial period. Whilst the exact mechanism of action of the Lng-IUS remains uncertain, in our trial the device appears to provide relief in a reasonable proportion of patients with symptoms of pain (mainly dysmenorrhoea) and menorrhagia associated with endometriosis (minimal to moderate). It is a convenient alternative to systemic progestogens, and if the side effects, particularly of bleeding dysfunction, which are most common within the first 3 months, can be tolerated, then it could be retained for as long as 5 years. However, further studies are now required not only to verify if this device is equally as effective as other medical options, in the form of larger, well designed, controlled trials, but also to verify whether the device retains its therapeutic effects for its full 5-year lifespan and whether it could be beneficial to women with severe/extensive disease presenting with similar symptoms to those we investigated.


    Acknowledgements
 
We are very grateful to Drs Sheila Thorpe and Peter Longthorne of Schering Healthcare for their support. We would like to thank all the women who participated in this study and to acknowledge the support we received from the local branch of the National Endometriosis Society. We wish to thank Schering Healthcare for funding the study.


    References
 Top
 Abstract
 Introduction
 Subjects and methods
 Results
 Discussion
 References
 
American Fertility Society (1985) Revised American Fertility Society classification of endometriosis. Fertil Steril 43,351–352.[Medline]

Andersson JK and Rybo G (1990) Levonorgestrel-releasing intrauterine device and norethisterone in the treatment of menorrhagia. Br J Obstet Gynaecol 97,690–694.[ISI][Medline]

Backman T, Huhtala S, Blom T, Luoto R, Rauramo I and Koskenvuo IM (2000) Length of use and symptoms associated with premature removal of the levonorgestrel intrauterine system: a nation-wide study of 17360 users. Br J Obstet Gynaecol 107,335–339.[ISI]

Du M, Shao O and Zhou X (1999) Serum levels of levonorgestrel during long-term use of Norplant [in Chinese]. J Zhonghua Fu Chan Ke Za Zhi 34,363–365.

Fedele L, Bianchi S, Zancomoto G, Portuese A and Ricciardia R (2001) Use of levonorgestrel-releasing intrauterine system in the treatment of rectovaginal endometriosis. Fertil Steril 75,485–488.[CrossRef][ISI][Medline]

Gardner FGE, Konje JC, KR, Brown LJR, Khanna S, Al-Azzawi F, Bell SC and Taylor DJ (2000) Endometrial protection from tamoxifen induced stimulated changes by a levonorgestrel-releasing intrauterine system: a randomised controlled trial. Lancet 356,1711–1717.[CrossRef][ISI][Medline]

Henderson AF and Studd JWW (1991) The role of definitive surgery and hormone replacement therapy in the treatment of endometriosis. In Thomas E and Rock J (eds). Modern approaches to endometriosis. Kulwer Academic Publishers, Dordrecht, The Netherlands. pp. 275–290.

Higham J, O’Brien PMS and Shaw RW (1990) Assessment of menstrual loss using a pictorial chart. Br J Obstet Gynaecol 97,734–739.[ISI][Medline]

Irvine GA, Campbell-Brown MB, Lumsden MA, Heikkila A, Walker JJ and Cameron IT (1998) Randomised comparative trial of levonorgestrel intrauterine system and norethisterone for the treatment of idiopathic menorrhagia. Br J Obstet Gynaecol 105,592–598.[ISI][Medline]

Karoly P (1987) Multimethod Assessment of Chronic Pain. Pergamon Press, Oxford, UK. pp. 43–53.

Linton SJ and Melin L (1982) The accuracy of remembering chronic pain. Pain 13,281–285.[CrossRef][ISI][Medline]

Jarvela I, Tekay A and Jouppila P (1998) The effect of a levonorgestrel intrauterine system on uterine artery blood flow, hormone concentrations and ovarian cyst formation in fertile women. Hum Reprod 13,3379–3383.[Abstract]

Luciano AA, Turksoy RN and Carleo J (1988) Evaluation of oral medroxyprogesterone acetate in the treatment of endometriosis. Obstet Gynecol 72,323–327.[Abstract]

Maruo T, Laoag-Fernandez JB, Pakarinen P, Murakoshi H, Spitz IM and Johansson E (2001) Use of a levonorgestrel intrauterine system on proliferation and apoptosis in the endometrium. Hum Reprod 16,2103–2108.[Abstract/Free Full Text]

McLaren J, Prentice A, Charnock-Jones DA, Millican SA, Muller KH, Sharkey AM and Smith SK (1996) Vascular endothelial growth factor is produced by peritoneal fluid macrophages in endometriosis and is regulated by ovarian steroids. J Clin Invest 98,482–489.[Abstract/Free Full Text]

Morghissi K (1999) Medical treatment of endometriosis. Clin Obstet Gynaecol 42,620–632.[CrossRef]

Nilsson CG, Luukkainen T and Arko H (1978) Endometrial morphology of women using a D-norgestrel releasing intrauterine device. Fertil Steril 29,397–401.[ISI][Medline]

Nilsson CG, Lahteeenmaki P, Robertson DN and Luukkainen P (1980) Plasma concentrations of levonorgestrel as a function of the release rate of levonorgestrel from medicated intrauterine devices. Acta Endocrinol 93,380–384.[ISI][Medline]

Overton C and Davis M (2002) In McMillan and Shaw J (eds), An Atlas of Endometriosis. 2nd edn, Parthenon Publishing, UK. pp. 15–17.

Sangi-Haghpeykar H and Poindexter A (1995) Epidemiology of endometriosis among parous women. Obstet Gynecol 85,983–992.[Abstract/Free Full Text]

Scholes D, LaCroix AZ, Ichikawa LE, Barlow WE and Ott SM (2002) Injectable hormone contraception and bone density: results from a prospective study. Epidemiology 13,581–587.[CrossRef][ISI][Medline]

Silverberg SG, Haukkamaa M, Arko H, Nilsson CG and Luukkainen T (1986) Endometrial morphometry of women during long term use of levonorgestrel intrauterine system. Int Gynaecol Pathol 5,235–241.

Vercellini P, De Giorgi O, Aimi G, Panazza S, Uglietti A and Crosignani PG (1997) Menstrual characteristics in women with and without endometriosis. Obstet Gynecol 90,264–268.[Abstract/Free Full Text]

Vercellini P, Aimi G, Panazza S, De Giorgi O, Pesole A and Crosignani PG (1999) A levonorgestrel releasing intrauterine system for the treatment of dysmenorrhoea associated with endometriosis: a pilot study. Fertil Steril 72,505–552.[CrossRef][ISI][Medline]

Witz C (1999) Current concepts in the pathogenesis of endometriosis. Clin Obstet Gynaecol 42,566–585.[CrossRef]

Submitted on January 27, 2003; resubmitted on June 2, 2003; accepted on September 16, 2003.