Clinica Ostetrica e Ginecologica I, Istituto Luigi Mangiagalli, University of Milano, Milan, Italy

FERTILITY AND STERILITY VOL. 80, NO. 2, AUGUST 2003 Copyright 2003 American Society for Reproductive Medicine Published by Elsevier Inc. Printed on acid-free paper in U.S.A. Comparison of a levonorgestrel-releasing intrauterine device versus expectant management after conservative surgery for symptomatic endometriosis: a pilot study Paolo Vercellini, M.D., Giada Frontino, M.D., Olga De Giorgi, M.D., Giorgio Aimi, M.D., Barbara Zaina, M.D., and Pier Giorgio Crosignani, M.D. Clinica Ostetrica e Ginecologica I, Istituto Luigi Mangiagalli, University of Milano, Milan, Italy Objective: To determine whether the frequency and severity of dysmenorrhea are reduced in women with symptomatic endometriosis in whom a levonorgestrel-releasing intrauterine device (Lng-IUD) is inserted after operative laparoscopy compared with those treated with surgery only. Design: Open-label, parallel-group, randomized, controlled trial. Setting: A tertiary care and referral center for patients with endometriosis. Patient(s): Parous women with moderate or severe dysmenorrhea undergoing first-line operative laparoscopy for symptomatic endometriosis. Intervention(s): Randomization to immediate Lng-IUD insertion or expectant management after laparoscopic treatment of endometriotic lesions. Main Outcome Measure(s): Proportions of women with recurrence of moderate or severe dysmenorrhea in the two study groups 1 year after surgery and overall degree of satisfaction with treatment. Result(s): Moderate or severe dysmenorrhea recurred in 2 of 20 (10%) subjects in the postoperative Lng-IUD group and 9/20 (45%) in the surgery-only group. Thus, a medicated device inserted postoperatively will prevent the recurrence of moderate or severe dysmenorrhea in one out of three patients 1 year after surgery. A total of 15/20 (75%) women in the Lng-IUD group and 10/20 (50%) in the expectant management group were satisfied or very satisfied with the treatment received. Conclusion(s): Insertion of an Lng-IUD after laparoscopic surgery for symptomatic endometriosis significantly reduced the medium-term risk of recurrence of moderate or severe dysmenorrhea. (Fertil Steril 2003; 80:305 9. 2003 by American Society for Reproductive Medicine.) Key Words: Endometriosis, dysmenorrhea, pelvic pain, IUD, progestins Received September 9, 2002; revised and accepted January 14, 2003. Prize paper presented at the 8th World Congress on Endometriosis, San Diego, California, February 24 27, 2002. Reprint requests: Paolo Vercellini, M.D., Clinica Ostetrica e Ginecologica I, Istituto Luigi Mangiagalli, Università di Milano, Via Commenda 12, 20122, Milan, Italy (FAX: 3902-50320252; E-mail: paolo. vercellini@unimi.it). 0015-0282/03/$30.00 doi:10.1016/s0015-0282(03) 00608-3 Dysmenorrhea is the most frequent complaint reported by women with endometriosis (1). Laparoscopic surgery is often the treatment of choice for symptomatic disease, but results are not always satisfactory and pain recurrence is common (2, 3). Postoperative medical therapy should theoretically induce resorption of microscopic foci and of lesions that could not be removed, reduce the risk of iatrogenic dissemination of endometriotic cells, and improve pain relief. However, the available evidence on this issue is contradictory, and it cannot be ruled out that hormones, which are not cytoreductive, produce a temporary effect limited to the period of amenorrhea that occurs because of anovulation and hypoestrogenism (4). Moreover, most of the drugs used postoperatively, including danazol and GnRH agonists, cause subjective and metabolic side effects, are costly, and should generally be withdrawn after a few months. Accordingly, the identification of safe alternatives to permit prolonged postsurgical medical treatment would be advantageous. Endometriosis is generally a localized disease but is currently managed with systemic medical therapies. The use of drugs administered locally and specifically aimed at pelvic organs could limit the metabolic impact without reducing antalgic efficacy. 305

An intrauterine device (IUD) releasing levonorgestrel (i.e., an Lng-IUD), a potent 19-nortestosterone derivative progestin, can induce amenorrhea with a different modality with respect to standard regimens. In fact, locally administered levonorgestrel has a profound effect on the endometrium, which becomes atrophic and inactive, although ovulation is usually not suppressed (5, 6). It has been demonstrated that insertion of an Lng-IUD may relieve the menstrual pain associated with recurrent endometriosis (7, 8). As no data are available on the postoperative use of this medicated device, we designed a pilot study to verify whether the frequency and severity of dysmenorrhea recurrence is reduced in women in whom an Lng-IUD is inserted immediately after laparoscopic surgery for endometriosis compared with women treated with laparoscopic surgery only. MATERIALS AND METHODS This open-label, parallel-group, randomized, controlled trial compared treatment with an Lng-IUD after laparoscopy for symptomatic endometriosis with postoperative expectant management. The main objective of the study was to evaluate the frequency and severity of dysmenorrhea 1 year after surgery. Dyspareunia and nonmenstrual pain as well as patient satisfaction were also assessed. Institutional review board approval was obtained, and the women gave their informed consent to the study. We considered parous women of 40 years who did not want children and who were undergoing first-line operative laparoscopy for symptomatic stage I IV endometriosis according to the revised American Fertility Society classification (9). Patients all reported disabling dysmenorrhea more than 6 months in duration. Patients were excluded if they had uterine or adnexal anomalies other than endometriosis (chronic pelvic inflammatory disease, leiomyomas, endometrial polyps, genital malformations, pelvic varices), used treatments for endometriosis other than nonsteroid antiinflammatory drugs in the 3 months before study entry (6 months for GnRH agonists), had contraindications to progestins, or were unwilling to tolerate menstrual changes. Each patient was asked to complete a questionnaire on the presence and severity of dysmenorrhea, deep dyspareunia, and nonmenstrual pelvic pain graded with a 0- to 3-point multidimensional categorical rating scale modified from the one devised by Biberoglu and Behrman (10), which defines dysmenorrhea according to loss of work efficiency and need for bed rest (absence of pain, 0; some loss of work efficiency, mild, 1; in bed part of 1 day, occasional loss of work, moderate, 2; in bed for 1 or more days, incapacitation, severe, 3); nonmenstrual pain according to various degrees of discomfort and use of analgesics (absence of pain, 0; occasional pelvic discomfort, mild, 1; noticeable discomfort for most of the cycle, moderate, 2; pain persisting during the cycle or requiring strong analgesics, severe, 3); and deep dyspareunia according to limitation of sexual activity (no discomfort, 0; tolerated discomfort, mild, 1; intercourse painful to the point of interruption, moderate, 2; intercourse avoided because of pain, severe, 3). The women were also requested to grade the severity of dysmenorrhea, nonmenstrual pelvic pain, and deep dyspareunia using a 100-mm visual analog scale, the left extreme of which indicates the absence of pain and the right one pain as bad as it could be; a score of 1 50 was considered mild pain, 51 80 moderate pain, and 81 100 severe pain. Subjects were recruited who had moderate or severe dysmenorrhea on both scales. Conservative surgery at laparoscopy was performed according to Cook and Rock (11) using mechanical instruments and electrosurgery only (2). Adhesions were sectioned with microscissors; the ovaries were completely mobilized; and endometriomas were evacuated, rinsed with normal saline, and excised by means of countertraction applied on the pseudocapsule and normal gonadal cortex with atraumatic microforceps. Hemostasis was achieved with limited application of bipolar current. After complete excision or coagulation of all endometriotic lesions, eligible subjects were randomized in a proportion of 1:1 to Lng-IUD insertion or expectant management. Treatment allocation was performed in accordance with a computer-generated randomization sequence using serially numbered, opaque, sealed envelopes. A medicated device that releases levonorgestrel 20 g/day over a period of 5 years (Mirena; Leiras Oy, Turku, Finland) was inserted immediately into the uterine cavity of patients allocated to the adjuvant therapy arm. After operative laparoscopy, the patients underwent follow-up visits every 3 months, at which time a gynecologic examination was performed, a history of the pattern of uterine bleeding was taken, and variations in pain symptoms intensity were recorded. When the Lng-IUD prevented regular flows, pain during erratic bleeding episodes was considered as dysmenorrhea. Subjects were withdrawn from the study if after surgery they used any type of hormone therapy that could affect the pain symptoms associated with endometriosis. At the 12- month evaluation, the women were requested to rate their overall degree of satisfaction with the treatment (very satisfied, satisfied, uncertain, dissatisfied, or very dissatisfied). The rates of recurrence of moderate to severe menstrual pain in the two study arms were compared using Fisher s exact test. Variations in pelvic symptoms were evaluated by computing pain scores on the multidimensional categorical rating and visual analog scales and comparing differences between baseline and final median scores in the two groups with the Mann-Whitney U test. When appropriate, the 95% confidence interval (95% CI) was calculated for the observed 306 Vercellini et al. Medicated IUD after surgery for endometriosis Vol. 80, No. 2, August 2003

TABLE 1 Age, parity, and endometriosis stage at previous surgery by treatment allocation. Postoperative Lng-IUD group n (%) differences. All statistical tests were two sided. P.05 was considered statistically significant. RESULTS Conservative surgery only group n (%) Age (y) 30 6 (30) 7 (35) 30 14 (70) 13 (65) Parity 1 11 (55) 12 (60) 2 9 (45) 8 (40) Disease stage a I 2 (10) 1 (5) II 2 (10) 4 (20) III 9 (45) 6 (30) IV 7 (35) 9 (45) Note: Lng-IUD levonorgestrel-releasing intrauterine device. a According to the revised American Fertility Society Classification (9). Vercellini. Medicated IUD after surgery for endometriosis. Fertil Steril 2003. Seventy-seven women evaluated at our endometriosis outpatient clinic were eligible for the study, but 32 declined randomization and five were lost to follow-up. Twenty of the remaining patients were allocated to Lng-IUD insertion after laparoscopic treatment, and 20 to conservative surgery only. Baseline clinical characteristics of the women enrolled in the trial are shown in Table 1. The distribution of the study variables was similar in both groups. Displacement of the Lng-IUD was observed in one woman 5 months after insertion, based on visualization at specular examination of the caudal extremity of the Lng- IUD vertical arm at the external cervical os. One subject in each group was lost to follow-up (at 9 months in the Lng- IUD group and at 7 months in the surgery-only group). These patients were classified as treatment failures in the evaluation both of postoperative dysmenorrhea rate, the primary end point, and of satisfaction with treatment. However, they were excluded from analysis of symptoms scores and menstrual pattern variations because complete information from pain diaries was not available. At the 12-month evaluation, amenorrhea was reported by five (28%) of the remaining 18 women in the Lng-IUD 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) and 1 point (1 2) in the postoperative Lng-IUD group and by 30 (25 40) and 1 (0 2) in the surgery-only group (P.012 and.021, respectively, Mann-Whitney U test; Table 2). 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.03, Fisher s exact test; relative risk.22; 95% CI,.05 90). The absolute risk reduction of dysmenorrhea recurrence in subjects undergoing Lng-IUD insertion compared with those allocated to expectant management was 35% (95% CI, 9% 61%). This means that an Lng-IUD inserted postoperatively will prevent the recurrence of moderate or severe dysmenorrhea in one out of three patients (95% CI, 2 11) 1 year after surgery, with a relative risk reduction of 78%. Dyspareunia and nonmenstrual pain scores were also reduced to a greater extent with the postoperative use of Lng-IUD (Table 2). One or more side effects were reported by eight of the 20 patients allocated to Lng-IUD insertion (bloating in six, weight gain in six, headache in three, seborrhea and acne in two, breast tenderness in one, decreased libido in one, and pelvic pain in one). Side effects were deemed tolerable by the women, and removal of the IUD was not necessary except in the case of the displaced device. At 12 months, six (30%) patients in the surgery plus Lng-IUD group were very satisfied with the treatment received, nine (45%) were satisfied, two (10%) uncertain, one (5%) dissatisfied, and two (10%) very dissatisfied compared with three (15%), seven (35%), three (15%), five (25%), and two (10%) in the laparoscopic surgery-only group. Overall, 75% of subjects in the surgery plus Lng-IUD group were satisfied or very satisfied after 1 year of treatment compared with 50% in the surgery-only group. DISCUSSION In this pilot study, the use of an Lng-IUD after conservative surgery for symptomatic endometriosis significantly reduced the medium-term risk of recurrence of moderate or severe dysmenorrhea and offered a higher degree of patient satisfaction. These results are probably due to the amenorrhea or hypomenorrhea associated with endometrial atrophy induced in most women by the locally released levonorgestrel (6). In fact, we had previously demonstrated a major decrease in uterine bleeding as well as in pain at menstruation in women with endometriosis treated with the Lng-IUD (7). However, a general effect secondary to uterine absorption of the progestin cannot be excluded, especially considering the typical side effects reported by some women using the medicated device (5). Nevertheless, toxicology studies suggest that local delivery of progestins in depot formulations is safe and generally well tolerated and may offer advantages over systemic administration (12). Also, deep-thrust dyspareunia was relieved to a greater extent in the postoperative Lng-IUD group in comparison with the surgery-only group. The effect of the Lng-IUD on FERTILITY & STERILITY 307

TABLE 2 Pain symptom scores in patients with endometriosis before and 12 months after laparoscopic surgery according to treatment allocation and pain scale. a Postoperative Lng-IUD group Conservative surgery only group Visual analog Multidimensional Visual analog Multidimensional Dysmenorrhea (n 18) b (n 19) b Baseline values 79 (65 83) 2 (1 2) 77 (59 83) 2 (1 2) 12-month values 22 (12 39) 1 (0 1) 41 (21 58) 1 (1 1) Median reduction c 50 (35 65) 1 (1 2) 30 (25 40) d 1(0 2) e Deep dyspareunia f (n 9) (n 8) Baseline values 52 (30 69) 1 (0 1) 55 (35 71) 1 (0 2) 12-month values 16 (12 33) 0 (0 1) 34 (20 44) 1 (1 1) Median reduction c 31 (20 45) 1 (1 1) 15 (10 40) 1 (0 1) Nonmenstrual pain (n 5) (n 7) Baseline values 53 (40 62) 1 (1 1) 49 (25 70) 1 (0 2) 12-month values 31 (20 48) 1 (0 1) 36 (21 45) 1 (0 1) Median reduction c 17 (8 17) 0 (0 1) 10 (7 14) 0 (0 1) Note: Lng-IUD levonorgestrel-releasing intrauterine device. a Values are medians with interquartile ranges in parentheses. b n no. of subjects reporting the symptom. c Figures are baseline values minus 6-month values. d The between-group difference is statistically significant, P.012, Mann-Whitney U test. e The between-group difference is statistically significant, P.021, Mann-Whitney U test. f One patient in the postoperative Lng-IUD group and two in the conservative surgery only group reported no sexual activity. Vercellini. Medicated IUD after surgery for endometriosis. Fertil Steril 2003. pain at intercourse has already been demonstrated by Fedele et al. (8) and could be due to a receptor-mediated action of locally released levonorgestrel on deep endometriotic foci adjacent to the isthmus and uterine cervix. Differences in nonmenstrual pain were limited and of questionable clinical importance. This is not unexpected as ovulation, the main cause of the above symptom (13), is usually not suppressed in women with an Lng-IUD. The small sample size undoubtedly limits the strength of our findings. Because no literature data were available on which to base a predictable additional effect of the medicated IUD in preventing menstrual pain after conservative surgery, a preplanned power calculation would have been difficult to perform. However, our trial was designed as a pilot study to identify major differences in symptom relief. The number of subjects recruited was enough to identify as statistically significant the observed 35% absolute risk reduction in dysmenorrhea recurrence. Another potential bias of the study is the open-label design. Double blinding would have been preferable but was not feasible owing to local organizational and bureaucratic requirements. Furthermore, insertion of an inert, nonmedicated, simple Silastic IUD was deemed unethical because it could have increased the amount of blood flow and pain at menstruation, exposing women to potential morbidity without any therapeutic advantage. Completion of pain diaries and questionnaires by the women themselves should have avoided undue influences by the investigators. Pain variation was assessed with two types of scales: a visual analog scale that estimates the subjective perception of the symptom and a multidimensional categorical rating scale that reflects also the functional impairment associated with dysmenorrhea. In addition, the degree of satisfaction with treatment was evaluated to include recognition of the patient s point of view and to give adequate weight to side effects that may have a major impact on the health-related quality of life. Except in one case requiring removal because of displacement, the Lng-IUD was generally well tolerated. The limited severity of the side effects reported may be due to the lower progestin plasma levels detected in medicated IUD wearers compared with subjects taking the drug orally or SC (5, 6, 12). Insertion of a medicated device after conservative surgery for endometriosis may constitute an innovative, effective, safe, and convenient adjuvant treatment for the long-term reduction of risk of dysmenorrhea recurrence. As hormonal therapies are symptomatic, the possibility of prolonging treatment indefinitely would be a major achievement in the management of women with endometriosis. However, further trials are needed to verify whether the good results observed are maintained during the entire 5-year period of efficacy of the system. Moreover, our study was conducted 308 Vercellini et al. Medicated IUD after surgery for endometriosis Vol. 80, No. 2, August 2003

on a selected population, namely, parous women not wanting to conceive with dysmenorrhea as their only or main symptom. The Lng-IUD may not demonstrate the same antalgic activity or could constitute an inappropriate choice in other clinical settings. References 1. Vercellini P, Trespidi L, De Giorgi O, Cortesi I, Parazzini F, Crosignani PG. Endometriosis and pelvic pain: relation to disease stage and localization. Fertil Steril 1996;65:299 304. 2. Crosignani PG, Vercellini P, Biffignandi F, Costantini W, Cortesi I, Imparato E. Laparoscopy versus laparotomy in conservative surgical treatment for severe endometriosis. Fertil Steril 1996;66:706 11. 3. Vercellini P, De Giorgi O, Pisacreta A, Pesole A, Vicentini S, Crosignani PG. Surgical management of endometriosis. Baillière s Clin Obstet Gynaecol 2000;14:501 23. 4. Vercellini P, Frontino G, De Giorgi O, Pietropaolo G, Pasin R, Crosignani PG. Pre- and post-surgical management of endometriosis. Obstet Gynecol Clin N Am 2003;30:163 80. 5. Nilsson CG, Lahteenmaki T, Robertson DN, Luukkainen T. Plasma concentrations of levonorgestrel as a function of the release rate of levonorgestrel from medicated intrauterine devices. Acta Endocrinol 1980;93:380 4. 6. Odlin V. Long-term experience of a levonorgestrel-releasing intrauterine system. Eur J Contracept Reprod Health Care 1996;1:319 23. 7. Vercellini P, Aimi G, Paonazza S, De Giorgi O, Pesole A, Crosignani PG. A levonorgestrel-releasing intrauterine system for the treatment of dysmenorrhea associated with endometriosis: a pilot study. Fertil Steril 1999;72:505 8. 8. Fedele L, Bianchi S, Zanconato G, Portuese A, Raffaelli R. Use of a levonorgestrel-releasing intrauterine device in the treatment of rectovaginal endometriosis. Fertil Steril 2001;75:485 8. 9. The American Fertility Society. Revised American Fertility Society classification of endometriosis: 1985. Fertil Steril 1985;43:351 2. 10. Biberoglu KO, Behrman SJ. Dosage aspects of danazol therapy in endometriosis: short-term and long-term effectiveness. Am J Obstet Gynecol 1981;139:645 50. 11. Cook AS, Rock JA. The role of laparoscopy in the treatment of endometriosis. Fertil Steril 1991;55:663 80. 12. Jordan A. Toxicology of progestogens of implantable contraceptives for women. Contraception 2002;65:3 8. 13. Vercellini P. Endometriosis: what a pain it is. Semin Reprod Endocrinol 1997;15:251 61. FERTILITY & STERILITY 309