Abstract
Endometriosis-associated symptomatology can be safely and effectively treated with intrauterine-released progestin, which is associated with fewer adverse effects than other therapeutic options and may be used on a long-term basis. We have herein reviewed the current literature in relation to the biological and clinical rationale for the use of an intrauterine system releasing 20 μg/day of levonorgestrel for the treatment of pelvic pain symptoms associated with endometriosis. Levonorgestrel induces endometrial glandular atrophy and decidual transformation of the stroma, reduces endometrial cell proliferation and increases apoptotic activity. After the first year of use, a 70–90% reduction in menstrual blood loss is observed. The levonorgestrel-releasing intrauterine system has proven effective in relieving pelvic pain symptoms caused by peritoneal and rectovaginal endometriosis and in reducing the risk of recurrence of dysmenorrhea after conservative surgery. Thus, the intrauterine delivery of a potent progestin may constitute an innovative, effective, safe and convenient alternative for local delivery of a potent progestin in the long-term therapy of symptomatic endometriosis.
Keywords
Several forms of therapy have been used for the conservative treatment of endometriosis, including gonadotropin-releasing hormone (<GnRH) analogues, androgen derivatives, combined oral contraceptives and progestins. Drugs used in the treatment of endometriosis are not cytoreductive and, at suspension, the endometrium, both eutopic and ectopic, resumes its metabolic activity [1] and the patient frequently suffers symptomatic relapse [2]. Thus, one of the major aims of current clinical research on the management of symptomatic endometriosis is to identify safe and effective therapeutic alternatives that would allow prolonged treatment [1].
Progestins are generally well tolerated, are associated with fewer adverse effects than danazol and GnRH agonists, are inexpensive and may be used on a long-term basis [1]. In this regard, the possibility of administering progestin locally, aiming its therapeutic action specifically at the endometrium, could further reduce the metabolic impact without limiting efficacy on pain symptoms. This seems particularly advantageous in the case of endometriosis, which is generally a localized disease [2].
The possibility of prolonging progestin treatment indefinitely without the need for daily drug administration has been achieved with the use of an intrauterine device that releases 20 μg/day of levonorgestrel (LNG), a potent 19-nortestosterone derivative progestin, over a 5-year period [3]. The LNG-releasing intrauterine device or system (LNG-IUS) consists of a 32 mm T-shaped polyethylene frame with a cylinder containing 52 mg of LNG wrapped around its stem [4–6].
In this review, we focus on biological and clinical studies addressing the effect of LNG-IUS in endometriosis in order to provide a general idea of the pros and cons of this specific treatment over other conventional therapies.
We used the Medline and Embase databases to identify English-language articles on the basis of combinations of medical subject heading terms: endometriosis, pelvic pain, medical therapy, progestins and levonorgestrel-releasing intrauterine device.
Use of the levonorgestrel-releasing intrauterine system in endometriosis: biological & vascular effects
The pharmacokinetics and pharmacodynamics of LNG released in utero have been studied in detail. The LNG-IUS initially releases 20 μg of LNG/day, whereas by the end of the fifth year the daily release is approximately 14 μg/24 h. Accordingly, serum plasma concentrations of LNG in LNG-IUS users have been reported to decrease with increased time of use. Table 1 shows serum, peritoneal fluid and tissue concentrations of LNG according to different studies.
Concentrations of levonorgestrel in serum, tissue and peritoneal fluid of levonorgestrel-releasing intrauterine system users.
Similar after insertion of the levonorgestrel intrauterine devices or after oral administration.
Many-fold higher with the levonorgestrel intrauterine devices than with oral administration.
Ovarian function
Ovulation is usually suppressed when serum LNG levels are 200–300 pg/ml, with a marked intra-individual response [7]. Accordingly, anovulatory rates of 70–85% have been reported during the first few months of the device use [8], whereas thereafter ovulation is suppressed in only 15–40% of users in relation to length of use [9]. Ovarian steroidogenesis is not consistently affected and circulating levels of estradiol remain well within the range of women of reproductive age independently of menstrual pattern [10,11]. Enlarged ovarian follicles have been detected by ultrasound in 19% of users of the LNG system but most of the enlarged ovarian images disappeared after 2 weeks of follow-up [12].
Endometrial function
The LNG-IUS exerts profound morphological and molecular effects upon the endometrium, summarized in Table 2 . Moreover, progesterone stimulates differentiation of endometrial tissue, reduces plasminogen activator (PA) activity, down-regulates pericellular proteolysis and upregulates PA inhibitor-1 and urokinase-type PA receptor [13,14]. Local changes that occur in the endometrial microvasculature following exposure to long-acting progestogen are associated with breakthrough bleeding [15]. In general, all these histological and molecular effects of locally released LNG determine a major reduction in menstrual blood loss. Indeed, 20–30% of LNG-IUS wearers experience amenorrhea, although their serum hormonal profiles are mostly consistent with ovulation [7]. This is also the biological rationale for the use of the compound in menorrhagia [16].
Histological and molecular effects of levonorgestrel-releasing intrauterine system on the endometrium.
Effects on uterine blood circulation
It has been hypothesized that the reduction of pain induced by LNG-IUS in women with endometriosis may be mediated by a decrease in the vascular supply to the pelvis with relief from pelvic congestion. To determine the effect of the device on the impedance of blood flow in the uterine arteries, Järvela and colleagues performed transvaginal color Doppler ultrasonography in 27 fertile women before and 3 months after IUS insertion [17]. The mean ± standard deviation (SD) mid-luteal uterine artery pulsatility index rose from a pretreatment value of 2.28 ± 0.48 to 2.70 ± 0.67 after 3 months of LNG-IUS use. However, this increase was detectable only in the 16 subjects with serum LNG concentrations over 200 pg/ml. The above findings are at odds with those of Pakärinen and colleagues, who measured the impedance of uterine blood flow in ten women before and 3 months after the device insertion and did not detect variations in circulatory conditions and pulsatility index at transvaginal Doppler flow sonography [18]. Zalel and colleagues evaluated the uterine vasculature of 47 women using the LNG-IUS and compared the findings with those observed in a control group of 35 women using a standard copper IUS [19]. Doppler flow in the cervical branch of the uterine artery did not reveal variations between the groups (resistance index: 0.6 ± 0.01), whereas subendometrial flow in the spiral artery was significantly reduced in 35 subjects of the LNG-IUS group and in none of the copper IUS group. The authors concluded that the reduction in uterine bleeding experienced by LNG-IUS users may be mainly due to a local progestational effect (marked reduction in spiral artery blood flow and endometrial thickness) associated with normal systemic blood flow.
Use of intrauterine levonorgestrel in endometriosis: clinical effects
The LNG-IUS has been used in various clinical conditions of endometriosis, namely, in patients with peritoneal, superficial ovarian, rectovaginal and recurrent endometriotic lesions, and also as a postoperative measure, particularly to treat pain symptoms.
Symptomatic peritoneal & superficial ovarian lesions
Lockhat and colleagues studied women with stage I–III endometriosis of reproductive age who underwent laparoscopy for pelvic pain to investigate the effectiveness of the LNG-IUS for the relief of symptoms [20]. In 34 subjects, a LNG-IUS was inserted intraoperatively. In five women, the device was removed soon after surgery owing to side effects or worsening pain, and another subject was excluded from the pain analysis due to protocol violation. At 6-month evaluation, dysmenorrhea was significantly relieved in the remaining 28 women, as the proportion of patients experiencing moderate or severe menstrual pain fell from 96 to 50%. Dyspareunia improved after 6 months of therapy in 13 of the 20 women (45%) experiencing the symptom at the beginning of the study; in two women (7%) it worsened. Noncyclical pelvic pain was not significantly reduced. The mean number of days of pain experienced every month was reduced after 6 months of therapy. Menstrual blood loss, assessed by means of the semiquantitative pictorial blood-loss assessment chart (PBAC) score [21], varied from 204 ± 196 at pre-insertion evaluation to 90 ± 157 at 6 months postinsertion (>100 suggestive of menorrhagia). At the end of treatment, of the 29 women who completed the study, 14 patients were very satisfied, five satisfied, eight uncertain and two dissatisfied. Including dropouts as failures in an intent-to-treat analysis, slightly more than half of the recruited subjects (19/34; 56%) were satisfied or very satisfied with the treatment received. A total of six patients requested device removal at the end of the study period. In the 26 women who underwent a second laparoscopy, the mean ± SD endometriosis score varied from 12.2 ± 10.9 at baseline to 10.7 ± 11.6 after 6 months of therapy, with an improvement in the American Fertility Society classification stage in 30.8% of the cases [22]. However, an improvement in symptoms was not necessarily associated with a change in staging.
The 3-year follow-up data of the subjects enrolled in the 6-month trial who requested continuation of therapy have been recently reported by the same authors [23]. The device was retained by 23/34 (67.6%) women at 12 months, 21 (61.8%) at 24 months and 19 (55.9%) at 36 months. Out of a total of 15 discontinuations, five (33%) were for unacceptable irregular bleeding, most of which were within the first 6 months. Pelvic pain (20.6%) and weight gain (8.8%) were the second and third most common reasons for requesting removal, respectively. There were no expulsions over the 3-year period. The mean ± SD 10 cm visual analog scale (VAS) dysmenorrhea score fell from a pre-insertion value of 7.7 ± 1.3 to 2.7 ± 1.5 at 36 months. Moderate or severe dysmenorrhea was experienced by 27 patients at pre-insertion, seven at 12 months, three at 18 months and only one at 24 months. The mean ± SD number of monthly days with pain dropped from 15.0 ± 6.9 at baseline to 6.0 ± 3.4 after 12 months of therapy. The most dramatic improvement in symptoms occurred during the first 12 months of therapy. Thereafter, there were no significant changes over the remaining 24 months.
A multicenter, randomized, controlled clinical trial with the aim of comparing the efficacy of the LNG-IUS and a depot-GnRH analogue in controlling endometriosis-related pain has been recently published by Petta and colleagues [24]. The study group consisted of 82 women with symptomatic endometriosis who were randomized to receive either LNG-IUS (n = 39) or GnRH analogue (n = 43) for 6 months. Both therapies appeared to be similarly effective for the treatment of endometriosis-associated chronic pelvic pain, since from baseline to study completion there was a six-point (standard error of the mean [SEM]: 0.3) decrease in VAS pain score in the LNG-IUS group and a six-point (SEM: 0.2) decrease in the GnRH analogue group. At the end of the study period, only five and six women, respectively, in each group failed to achieve a VAS pain score of less than three. In both groups, women with stage III and IV disease showed a more rapid improvement in the VAS score than those with stage I and II disease. LNG-IUS users had a higher bleeding score than GnRH analogue users, with 34 and 71% of patients, respectively, in each group reporting no bleeding during the first treatment month and 70 and 98% reporting no bleeding during the sixth month. No difference was observed between groups with respect to improvement in quality of life as assessed using the Psychological General Well-Being Index Questionnarie [25]. Similarly, no difference was found with respect to side effects such as abdominal distension or peripheral edema, although more symptoms of breast tenderness were recorded in LNG-IUS users.
Rectovaginal lesions
Fedele and colleagues recruited 11 symptomatic women, who previously underwent conservative surgery without excision of deep lesions, to test the effectiveness of the LNG-IUS as a therapy for rectovaginal endometriosis [26]. Variations in pain symptoms and size of plaques were evaluated. At 1-year follow-up, nine women were oligo-menorrheic and two experienced amenorrhea; dysmenorrhea, which had been moderate or severe in all cases, and nonmenstrual pelvic pain were absent. Of notable interest was the reduction of deep dyspareunia, which had been moderate or severe in eight cases prior to the device insertion, to absent or mild in all subjects throughout treatment. Dyschezia was relieved in four out of five women by the sixth month of treatment. Transrectal ultrasonography showed a slight but significant reduction of rectovaginal lesions after 6 months of therapy. Side effects of the therapy were limited. The results of this study are of particular clinical importance because they prove the efficacy of a progestinic treatment in a type of lesion that is generally considered refractory to any medical therapy. Relief of organic symptoms such as deep dyspareunia and rectal tenesmus is probably not only due to reduction in size of the fibronodular rectovaginal plaques, but also to a reduction of the intra- and peri-lesional inflammatory condition.
Recurrent endometriosis
Vercellini and colleagues studied 20 parous patients not wishing to conceive, who had been operated conservatively for endometriosis in the previous 12 months [27]. These women, who did not want to undergo further surgery, were evaluated for recurrent moderate-to-severe menstrual pain. Variations of dysmenorrhea severity, as well as patients' satisfaction, were assessed after 1 year of therapy with the LNG-IUS. Two women withdrew from the study (one requested IUS removal because of weight gain and abdominal bloating; in the other the medicated IUS was expelled 3 months after insertion) and one was lost to follow-up. The menstrual pattern in the remaining 17 women after 12 months of treatment was characterized by amenorrhea, hypomenorrhea or spotting in 71%, and normal flow in 29%. Baseline PBAC score was 111 and the 12-month follow-up mean score was 27 [21]. During the study period the mean ± SD 100 mm visual analog and 0–3 verbal rating scale scores dropped from 76 ± 12 to 34 ± 23 points and from 2.5 ± 0.5 to 1.2 ± 0.5 points, respectively. Only five (29%) women were considered still symptomatic at final evaluation. Side effects were reported by nine patients. A total of 75% of women were very satisfied or satisfied with treatment, 10% were uncertain and only 15% were dissatisfied at 12-month follow-up.
Postoperative therapy
Vercellini and colleagues designed a pilot study to verify if the frequency and severity of dysmenorrhea is reduced at 1-year follow-up in women in whom a LNG-IUS is inserted immediately after laparoscopic surgery for endometriosis compared with women treated with laparoscopic surgery only [28]. Parous women aged 40 years or younger, not wanting children, undergoing first-line operative laparoscopy for mild-to-severe endometriosis and reporting disabling dysmenorrhea of over 6-month duration were considered. Pain symptoms were graded with a 0- to 3-point multidimensional categorical rating scale and a 100 mm VAS. After complete excision or coagulation of all endometriotic lesions, 20 subjects were randomized to LNG-IUS insertion and 20 to postoperative expectant management. Displacement of the device was observed in one woman 5 months after insertion. One subject in each group was lost to follow-up. At the 12-month evaluation, amenorrhea was reported by five (28%) of the remaining 18 women in the LNG-IUS arm, hypomenorrhea or spotting by nine (50%) and normal flows by four (22%). Median (interquartile range) dysmenorrhea visual analog and multidimensional categorical rating scale scores fell by 50 mm (35–65 mm) and one point in the postoperative LNG-IUS group and by 30 mm (25–40 mm) and one point in the surgery-only group (p = 0.012 and 0.021, respectively). According to an intention-to-treat analysis, postoperative moderate or severe dysmenorrhea recurrence was less frequent in the former group (2/20 subjects; 10%) than in the latter (9/20; 45%; p = 0.03; relative risk: 0.22; 95% confidence interval: 0.05–0.90). The absolute risk reduction of dysmenorrhea recurrence in subjects undergoing LNG-IUS insertion compared with those allocated to expectant management was 35%. Accordingly, if a LNG-IUS was inserted postoperatively in three patients, one would avoid moderate or severe dysmenorrhea at 1 year after surgery.
Dyspareunia and nonmenstrual pain scores were also reduced to a great extent with the postoperative use of LNG-IUS. Side effects were reported by eight of the 20 women allocated to LNG-IUS insertion, but were deemed tolerable, and removal of the IUS was not necessary except in the case of a displaced device. At 12 months, 30% of the patients in the surgery plus LNG-IUS group were very satisfied with the treatment received, 45% were satisfied, 10% uncertain, 5% dissatisfied and 10% very dissatisfied compared with 15, 35, 15, 25 and 10% in the laparoscopic surgery-only group. Overall, 75% of subjects in the surgery plus LNG-IUS group were satisfied or very satisfied after 1 year of treatment compared with 50% in the surgery-only group, respectively. Based on these data, insertion of the medicated device after conservative surgery for endometriosis may constitute an innovative, effective, safe and convenient adjuvant treatment for the reduction of risk of dysmenorrhea recurrence.
Discussion
Levonorgestrel is a potent progestin with androgenic and antiestrogenic activity on the endometrium [29]. The induction of endometrial glandular atrophy and extensive decidual transformation of the stroma, the downregulation of endometrial cell proliferation, and the increase in apoptotic activities all constitute a convincing theoretical basis for the use of locally released LNG in the treatment of symptomatic endometriosis [5,30–34]. The hypoamenorrhea that many women experience during LNG-IUS use may result from a combination of endometrial atrophy and ovarian suppression. Up to 85% of subjects using the device have anovulatory cycles during the first 3 months of use, but the proportion falls to less than 40% by 12 months [17]. Accordingly, the LNG-releasing device probably acts mainly via a local rather than a systemic mechanism. Studies on uterine vascular resistance and impedance of flow in the presence of the device have failed to consistently support an alteration of uterine perfusion.
The LNG-IUS has proven effective in relieving pelvic pain symptoms caused by peritoneal and rectovaginal endometriosis and in reducing the risk of recurrence of dysmenorrhea after conservative surgery [14,23,24,26–28].
Furthermore, the use of a LNG-IUS in women with endometriosis seems to confer several advantages over other conventional systemic therapies and may increase patients' compliance during long-term treatment. It is important to inform the patients that the local progestational effect takes approximately 3–6 months to manifest, in order to encourage their compliance [35].
The expected 70–90% reduction in menstrual blood loss could be particularly advantageous in women with endometriosis, who generally have heavier than normal flows [36]. However, women should be informed that during the first 3–4 months of use, major menstrual disorders are expected, including spotting, prolonged or continuous bleeding and even menorrhagia. After the first year of use, few women report intermenstrual bleeding and approximately 20–30% are amenorrheic. This is relevant as dysmenorrhea is the most frequent symptom in patients with endometriosis.
Intrauterine administration of LNG, with a possible direct distribution to pelvic tissues, would imply a local concentration greater than its plasma levels. This could translate into a superior effectiveness with limited adverse effects, also due to absence of the hepatic first-pass effect following oral administration of the drug. Based on the dosage of drug administered, the metabolic consequences of the LNG-IUS should be less pronounced than those of other contraceptive methods. Indeed, the release rate of 20 μg/24 h is inferior to that provided by the progestin-only pill (30 μg/day) and subcutaneous LNG implants (30–60 μg/day). However, a general effect secondary to uterine absorption of LNG cannot be excluded, as most reported side effects are typical of progestins. In fact, Lockhat and colleagues suggested that the progestin released by the IUS is rapidly absorbed by the subendometrial vascular network [23].
Other advantages include avoidance of the need for repeated administration, effective contraception and the fact that the device does not provoke hypoestrogenism. Although it may be expensive at the outset, the cumulative final costs might be less than those of other medications.
The expulsion rate of the device is over 5% and the risk of pelvic infection approximately 1.5% [6]. Accordingly, the recommended patient profile is parous women with no history of pelvic inflammatory disease, who are in stable and mutually monogamous relationships. Nulliparity is not a contraindication, but the use of IUSs in smaller uteri may be associated with increased uterine cramping. This could be particularly worrisome in patients with endometriosis-associated severe dysmenorrhea. However, encouraging data on the use of LNG-IUS among nulliparous women have been reported by Suhonen and colleagues, who, in a study comparing the safety and acceptability of LNG-IUS versus oral contraceptives for family contraception, have observed a similar performance between the two treatments in terms of continuation rates, adverse events, menstrual characteristics and effects on wellbeing [37]. Women with unproven fertility need to be informed that there is generally a return to fertility after LNG-IUS removal (within 1–2 months) [6]. This may be especially important for patients with endometriosis, in whom doubts may ensue, in case of future infertility, regarding the possible role of the LNG-IUS.
Finally, the efficacy of the LNG-IUS has been studied in subjects with peritoneal, superficial ovarian and rectovaginal endometriotic lesions, but limited information is available on the risk of endometrioma formation during long periods of therapy. It has been demonstrated that development of endometriotic ovarian cysts is associated with ovulation [38]. However, the LNG-IUS generally does not inhibit ovulation except for the first few months after insertion. In theory, this may constitute a specific drawback of the device in comparison with other forms of progestin treatment [1].
In conclusion, the LNG-IUS may constitute an innovative, effective, safe and convenient alternative for local delivery of a potent progestin in the long-term therapy of symptomatic endometriosis. Moreover, this form of treatment has been studied generally in selected populations, namely, parous women not wanting to conceive, with dysmenorrhea (i.e., functional pain) as the main symptom, and unwilling to undergo second-line surgery. The medicated device may not demonstrate the same antalgic activity or could constitute an inopportune choice in different clinical settings, and comparative trials with other treatment options are also needed to confirm the effect on organic pain symptoms, such as dyspareunia and dyschezia.
Future perspective
Although the data regarding the efficacy of the LNG-IUS for the treatment of endometriosis seem encouraging to date, further trials are needed to verify whether the good results observed are maintained during the entire 5-year period of efficacy. Moreover, effects on specific forms or symptoms associated with the disease are still to be verified. Since the LNG-IUS does not inhibit ovulation, recurrence of ovarian endometriomas should be considered as an area of drug evaluation. Future efforts should be directed toward large, randomized, well-controlled studies so that the efficacy of the LNG-IUS with regard to recurrence, symptoms and disease signs can be evaluated in respect to other medical therapies for the disease.
Executive summary
The biological bases for levonorgestrel-releasing intrauterine system (LNG-IUS) use include the induction of endometrial glandular atrophy and extensive decidual transformation of the stroma, the downregulation of endometrial cell proliferation and the increase in apoptotic activity.
Anovulatory rates of 70–85% have been reported during the first 4 months after device insertion; thereafter, ovulation is suppressed in only 15–40% of users in relation to length of use.
The LNG-IUS has proven effective in relieving pelvic pain symptoms caused by peritoneal and rectovaginal endometriosis, and in reducing the risk of recurrence of dysmenorrhea after conservative surgery.
After the first year of LNG-IUS use, a 70–90% reduction in menstrual blood loss is observed; few women report intermenstrual bleeding and approximately 20–30% report amenorrhea.
The intrauterine administration of LNG, which potentially leads to pelvic concentrations greater than plasma levels, could result in superior effectiveness with limited adverse effects and could increase patients' compliance on a long-term basis.
Large, randomized, well-controlled studies are necessary to confirm the long-term efficacy of the LNG-IUS with regard to recurrence, symptoms and disease signs.