Comparison of ovarian cyst formation in women using the

Ultrasound Obstet Gynecol 2002; 20: 381 385 Comparison of ovarian cyst formation in women using the Blackwell Science, Ltd levonorgestrel-releasing intrauterine system vs. hysterectomy P. INKI*, R. HURSKAINEN **, P. PALO*, E. EKHOLM*, S. GRENMAN*, A. KIVELÄ, E. KUJANSUU, J. TEPERI**, M. YLISKOSKI and J. PAAVONEN Departments of Obstetrics and Gynecology, University Hospitals of *Turku, Helsinki, Oulu, Tampere and Kuopio, Finland, and **National Research and Development Center for Welfare and Health (STAKES), Helsinki, Finland KEYWORDS: Levonorgestrel-releasing intrauterine system, Menorrhagia, Ovarian cyst, Ultrasound ABSTRACT Objective To analyze the effect of the levonorgestrelreleasing intrauterine system (LNG-IUS) on ovarian cyst formation, endometrial thickness and the size of uterus and uterine fibroids by ultrasonography. Subjects and methods This was a prospective, randomized trial comparing the LNG-IUS and hysterectomy in 236 women (age range 35 49 years) referred for menorrhagia. Transvaginal ultrasound examination was used to study the presence of ovarian cysts, uterine size, endometrial thickness, and the size of the uterus and uterine fibroids during a 12- month follow-up period. Results At baseline examination, 12 ovarian cysts were detected, eight in the LNG-IUS group and four in the hysterectomy group. During the follow-up period, 14 new cysts had emerged at 6 months and 14 new cysts had emerged at 12 months in the LNG-IUS group, whereas the corresponding figures in the hysterectomy group were three and eight, respectively. All but one of the 14 new cysts (94.1%) detected at 6 months in the LNG-IUS group resolved spontaneously, whereas two out of the eight cysts detected at the baseline examination persisted for 12 months. Three cysts were removed at operation. The relative risk of the occurrence of ovarian cysts was significantly higher in women with LNG- IUS, compared with women who underwent hysterectomy. LNG-IUS did not affect the size of the uterus or uterine fibroids, but it was associated with a decrease in endometrial thickness. The occurrence of cysts did not correlate with age or follicle stimulating hormone levels, but a weak positive correlation between the occurrence of cysts and the presence of irregular bleeding was observed. Conclusions LNG-IUS use in the treatment of menorrhagia was associated with the development of ovarian cysts, but these were symptomless and showed a high rate of spontaneous resolution. LNG-IUS did not affect the size of the uterus or the size of uterine fibroids, but decreased the thickness of the endometrium. INTRODUCTION The levonorgestrel-releasing intrauterine system (LNG-IUS) is an effective treatment for menorrhagia, with relatively few side-effects 1 3. However, previous studies have shown an increased incidence of ovarian cysts associated with LNG- IUS use 4 7. Earlier studies have not described the natural evolution of the cysts and the follow-up has tended to be short (3 months). 7 In addition, the effect of LNG-IUS on uterine size or the size of uterine fibroids has not been evaluated. Furthermore, it is not known whether hysterectomy has any effect on the occurrence of ovarian cysts in fertile aged women, although the effect of hysterectomy on ovarian function has been extensively studied 8 10. The purpose of this study was to analyze the occurrence and natural history of ovarian cysts during treatment with LNG-IUS, compared with that following hysterectomy, as well as the effect of LNG-IUS on uterine size and the size of uterine fibroids in a randomized trial of women referred for menorrhagia. METHODS The study population consisted of 236 women participating in a randomized trial comparing LNG-IUS and hysterectomy in the treatment of menorrhagia. The patient characteristics, trial profile, inclusion and exclusion criteria have been described in detail 1. In brief, the mean age of the women was 43 years (range 35 49 years), all had regular menstrual cycles, had completed family size, and were referred to any of the five university hospitals in Finland because of menorrhagia. Women with large enough fibroids to cause bowel or Correspondence: Dr Pirjo Inki, Department of Obstetrics and Gynecology, University Central Hospital of Turku, PO Box 52, 20521 Turku, Finland (e-mail: pirjo.inki@tyks.fi) Accepted 4-6-02 ORIGINAL PAPER 381

urinary symptoms, with submucous fibroids, lack of indication for hysterectomy, metrorrhagia as a main complaint, previous treatment failure with LNG-IUS, severe depression, history of malignancies, uterine malformation, or with ovarian cysts exceeding 55 mm in diameter, or with adnexal tumors regardless of the size, were excluded 1. At enrollment into the study, follicle stimulating hormone (FSH) levels were measured in each patient using an immunofluorometric method (Wallac, Turku, Finland). Patients were randomized using numbered, opaque, sealed envelopes to receive either LNG-IUS (n = 119) or hysterectomy (n = 117). LNG-IUS (Mirena, Leiras, Turku, Finland) was inserted during the randomization visit in 117 of the 119 women randomized for the LNG-IUS group. LNG-IUS insertion was unsuccessful in two of the 119 women because of cervical stricture or the presence of a submucous fibroid (not detected at ultrasonography). Hysterectomy was performed abdominally (20%), vaginally (28%), or laparoscopically (52%) in 107 women. In 10 women, hysterectomy was not performed due to decreased blood loss or changes in family or employment circumstances (as described previously) 1. Transvaginal ultrasound examinations were performed at baseline, at 6-month and at 12- month follow-up visits by eight experienced gynecologists. The examinations were performed using real-time linear array ultrasound machines equipped with a high frequency (5.0 7.5 MHz) endovaginal convex probe (Toshiba SSA 270 sonolayer, Tokyo, Japan). The size of the uterus (in sagittal plane and coronal plane) as well as the thickness of the endometrium (double-layer, measured from above the LNG- IUS) were noted. The size of the ovaries in two planes was also measured. The presence and location of any uterine fibroids exceeding 20 mm in diameter was also recorded and measured in two dimensions. Ovarian cysts exceeding 30 mm in size were recorded as cysts. A cyst was considered persistent if it was observed in the same ovary in the same patient at subsequent ultrasound examinations. The presence of septa, papillary formations, or inner echoes was noted. Additional follow-up visits or laparoscopy were scheduled if necessary. Statistical analysis Mean values and standard deviations were calculated for the ultrasonic measurements. These were normally distributed. The statistical differences between the groups were calculated using Student s t-test. The method of comparison of two frequencies was used to compare the incidences of cysts between the groups. A P-value of equal or less than 0.05 was considered as statistically significant. Chi-squared test was used to test the correlation between various patient characteristics and the occurrence of cysts in the LNG-IUS and hysterectomy groups. RESULTS At 6 months, ultrasound examination was performed in 97 women with the LNG-IUS in situ. Of the 117 women in the LNG-IUS group, nine women had had hysterectomy, 10 women no longer had the LNG-IUS, and one was lost to follow-up. At 12 months, ultrasound examination was performed in 79 women still using LNG-IUS. A total of 24 women had had hysterectomy, 10 women no longer had the LNG-IUS, and three women were lost to follow-up. Twentyeight (29%) of the 97 women with the LNG-IUS in situ after 6 months reported amenorrhea or oligomenorrhea and 41 (42%) reported intermenstrual bleeding. The corresponding figures at 12 months were 41 (51%) and 26 (32%) of 81 women, respectively 1. The most common reason for removal of the LNG-IUS was irregular or heavy bleeding, or pain. In only one patient was the presence of an ovarian cyst an additional reason for removal of the LNG-IUS. Hysterectomy was performed in 107 of the 117 women in the hysterectomy group. Bilateral oophorectomy was performed in five women and unilateral oophorectomy in seven women. In the hysterectomy group, 101 women who had at least one intact ovary attended the 6-month and 12-month follow-up visits. Their ultrasound findings were compared with those of 97 and 79 women with LNG-IUS in situ at 6 months and 12 months. At baseline, the size of the uterus, the presence and size of uterine fibroids (Table 1), or other patient characteristics, such as age, body mass index (BMI), or parity did not differ between the two groups. The overall incidence of ovarian cysts was 5.1% (12 of 236). At 6 months, the ultrasound examination revealed ovarian cysts in 20 patients, 17 in the LNG-IUS group and three in the hysterectomy group (Figure 1, Table 2). Thus, the overall incidence of ovarian cysts was 17.5% in the LNG-IUS group and 3.0% in the hysterectomy group. Out of 12 cysts initially detected, seven had spontaneously resolved (four in the LNG- IUS group and three in the hysterectomy group), two had been removed by operation (one in both groups), and three persisted (all in the LNG-IUS group). The cysts removed were both endometriomas. Overall, 17 new cysts were detected (14 in the LNG-IUS group and three in the hysterectomy group) (Figure 1). The cysts in the LNG-IUS group were somewhat larger (mean diameter 41 mm; range 31 74 mm) than those in the hysterectomy group (mean diameter 34 mm; range 32 37 mm, difference not significant). None of the women had more than one cyst. All except two cysts showed no inner echoes, and seven cysts showed a septum or septa. At 12 months, altogether 25 patients had an ovarian cyst, 17 in the LNG-IUS group and eight in the hysterectomy group (Figure 1, Table 2). The overall incidence of ovarian cysts was 21.5% in the LNG-IUS group and 8.0% in the hysterectomy group. Of the 20 cysts found at the 6-month follow-up, 17 had resolved (14 in the LNG-IUS group and Table 1 Selected ultrasound characteristics of the study population at baseline examination. Mean values are shown in mm Character LNG-IUS group (n = 119) Uterine thickness 49.5 48.6 Uterine width 58.1 59.1 Endometrial width 8.2 7.9 Occurrence of fibroids (%) 31.9 29.9 Average size of fibroids 28.0 29.4 Hysterectomy group (n = 117) 382 Ultrasound in Obstetrics and Gynecology

Figure 1 Occurrence of ovarian cysts during LNG-IUS treatment (a) or after hysterectomy (b) for menorrhagia. Number of cysts at baseline ultrasound examination is shown in white columns. New cysts are shown in gray columns at 6 months, and in black at 12 months. three in the hysterectomy group), none had been removed by operation, and three cysts persisted (these were all in the LNG-IUS group and two had already been identified at baseline). A total of 22 new cysts were detected; 14 in the LNG-IUS group and eight in the hysterectomy group. All except two patients had single cysts. The size of the cysts was equal in both groups (mean diameter in both groups 34 mm; range 23 66 mm in LNG-IUS group and 26 57 mm in hysterectomy group). In addition, one patient in the LNG- IUS group had an adnexal cyst that had persisted from the baseline. This was a benign paraovarian cyst subsequently removed. In the LNG-IUS group, two cysts persisted from the baseline examination to the 12-month follow-up visit. In one of these patients, ultrasound examination was suggestive of a hydrosalpinx, and did not require surgery. The other patient s cyst was later removed and was found to be benign and nonneoplastic. Only one out of 17 cysts detected at the last followup visit had developed during the follow-up (Figure 1a). The cysts that appeared during the follow-up remained symptomless. During the use of LNG-IUS, altogether 28 cysts emerged, compared with 11 in the hysterectomy group (P = 0.0012). The relative risk for ovarian cysts among patients treated with LNG-IUS was 5.9 (OR 6.9; 95% confidence intervals 1.96, 24.5) at 6 months, and 2.7 (OR 3.15; 95% confidence intervals 1.28, 7.76) at 12 months. The occurrence of ovarian cyst did not correlate with age or FSH levels (data not shown). In the LNG-IUS group, there was a correlation between cysts and irregular bleeding both at 6-month (χ 2 = 3.9, P = 0.05) and at 12-month followup (χ 2 = 8.4, P = 0.004). In addition, the occurrence of cysts correlated also with the presence of amenorrhea at 12 months (χ 2 = 5.0, P = 0.03) but not at 6 months (χ 2 = 1.8, P = 0.4). At baseline ultrasound examination, the size of the uterus and the thickness of the endometrium was similar in both groups (Table 1). The size of the uterus changed during LNG-IUS treatment, whereas the endometrial lining became thinner (Figure 2). Baseline ultrasound examination found fibroid(s) in 38 out of 119 patients (31.9%) in the LNG-IUS arm. At the 6-month follow-up visit, uterine fibroids were found in 19 out of 98 Figure 2 Size of uterus and fibroids and thickness of endometrium during LNG-IUS treatment for menorrhagia. Ultrasonic measurements (mean values in mm and SD) of uterine thickness (a), width (b), average diameter of fibroid(s) (c) and endometrial thickness (d) are shown. Ultrasound in Obstetrics and Gynecology 383

Table 2 Occurrence of ovarian cysts by ultrasound examination in patients treated for menorrhagia by levonorgestrel-releasing intrauterine system (LNG-IUS) or hysterectomy Group patients (19.4%), and at the 12-month follow-up in 16 out of 82 (19.5%). During the follow-up period, there was no marked change in the size of the fibroids (Figure 2). The lower incidence of fibroids at the follow-up visits can be at least partly explained by the fact that 46% of patients initially randomized into the LNG-IUS group and who subsequently had hysterectomy, had fibroids detected at the baseline ultrasound examination (data not shown). Thus, the likelihood of having the uterus removed was higher in patients who initially had fibroids. The average size of the fibroids was also somewhat larger (mean diameter, 34.3 mm) in these patients, compared with the average size of fibroids (mean diameter 28.7 mm) in the whole study population, but this difference was not statistically significant (data not shown). DISCUSSION Ovarian cysts present Baseline 6 months 12 months n/total (%) n/total (%) n/total (%) LNG-IUS 8/119 6.7 17/97 17.5 17/79 21.5 Hysterectomy 4/117 3.4 3/101 3.0 8/101 8.0 Total 12/236 5.1 20/198 10.1 25/179 14.0 We found that ovarian cysts appeared more frequently in menorrhagic women using LNG-IUS than in women who underwent hysterectomy. However, these were symptomless, relatively small in size, and showed a high rate (94%) of spontaneous resolution. The occurrence of ovarian cysts did not correlate to age or the level of FSH, whereas a weak correlation was observed between the occurrence of ovarian cysts and irregular bleeding. LNG-IUS did not affect the size of the uterus or the size of fibroids. The endometrium became thinner during LNG-IUS treatment, compared with the baseline ultrasound examination. At the baseline ultrasound examination, the overall occurrence of ovarian cysts was 5.1%, which is comparable with the incidence of 6.6% found by Borgfeldt and Andolf 7 in a random sample of women of 25 40 years of age. The incidence of ovarian cysts in previously published studies has been highly variable, depending on the inclusion criteria. If postmenopausal women are studied, the inclusion of small cysts less than 10 mm is justified. However, ovarian follicles can occasionally increase to 30 mm in size, although the upper limit of an ovarian follicle remains disputable (as previously noted by Borgfeldt and Andolf 7 ). Therefore, only cysts exceeding 30 mm were considered in our study. Progestin-only oral contraceptives and LNG-IUS have both been associated with increased occurrence of ovarian cysts. The incidence of ovarian cysts has been in the range of 8% 31% in previous studies 5 7,11. With the exception of one study 11, the incidence has been lower than in the present study. The women have also been younger (mean age 31 33 years) than in our study 5,7. In the present study the mean age was 43 years. Furthermore, earlier studies have been observational with no or relatively short (up to 3 months) follow-up, whereas in the present study women were followed up to 12 months. We found that ovarian cysts occurred in 17.5% of women using LNG-IUS, compared with 3.0% in women subjected to hysterectomy at 6 months. The corresponding figures at 12 months were 21.5% and 8.0%, respectively. The precise mechanism by which the ovarian cysts are caused by LNG-IUS is not known, but earlier studies have shown disturbances in the normal growth and rupture of follicles during LNG-IUS use 12. Most of the menstrual cycles seem to be ovulatory according to normal progesterone levels observed in the luteal phase 12,13. Accordingly, we found a weak positive correlation between the occurrence of ovarian cysts and irregular bleeding. The vast majority of the ovarian cysts were functional showing a high rate of spontaneous resolution, and were symptom-free. All but one of the 17 cysts (94.1%) that emerged during the follow-up disappeared spontaneously during the next 6 months. This finding is consistent with the study of Borgfeldt and Andolf 7, showing a high (82%) rate of spontaneous resolution during a short follow-up period of only 3 months. In contrast, two out of eight cysts (25%) detected at the baseline examination persisted for 12 months. If therefore a cyst is observed prior to the insertion of LNG- IUS, it has a lower probability of spontaneous resolution, compared with those appearing during the use of LNG-IUS. The LNG-IUS use did not change the size of the uterus or uterine fibroids. Patients who were found to have uterine fibroids at the baseline ultrasound examination were more likely to undergo hysterectomy during LNG-IUS use. As expected from previous studies, the endometrial lining became thinner during LNG-IUS use 5,14. In conclusion, use of LNG-IUS in the treatment of menorrhagia was associated with a relatively high rate of ovarian cysts. However, the cysts were symptomless and showed a high rate of spontaneous resolution. Therefore, in our opinion, no routine ultrasound screening is necessary of women using LNG-IUS. LNG-IUS did not affect the size of uterus or uterine fibroids, whereas the endometrial lining became thinner. REFERENCES 1 Hurskainen R, Teperi J, Rissanen P, Aalto A-M, Grenman S, Kivelä A, Kujansuu E, Vuorma S, Yliskoski M, Paavonen J. Randomised trial of levonorgestrel releasing intrauterine system or hysterectomy for the treatment of menorrhagia: quality of life and cost effectiveness. Lancet 2001; 357: 273 7 2 Lähteenmäki P, Haukkamaa M, Puolakka J, Riikonen U, Sainio S, Suvisaari J, Nilsson CG. Open randomised study of use of levonorgestrel releasing intrauterine system as alternative to hysterectomy. BMJ 1998; 316: 1122 6 3 Barrington JW, Bowen-Simpkins P. The levonorgestrel intrauterine system in the management of menorrhagia. Br J Obstet Gynaecol 1997; 104: 614 6 4 Barbosa I, Bakos O, Olsson SE, Odlind V, Johansson ED. Ovarian function during use of a levonorgestrel-releasing IUD. Contraception 1990; 42: 51 66 5 Pakarinen P, Suvisaari J, Luukkainen T, Lähteenmäki P. Intracervical and fundal administration of levonorgestrel for contraception: 384 Ultrasound in Obstetrics and Gynecology

endometrial thickness, patterns of bleeding, and persisting ovarian follicles. Fertil Steril 1997; 68: 59 64 6 Robinson GE, Boubds W, Kubba AA, Adams J, Guillebaud J. Functional ovarian cysts associated with the levonorgestrel-releasing intrauterine device. Br J Fam Plann 1989; 14: 131 2 7 Borgfeldt C, Andolf E. Transvaginal sonographic ovarian findings in a random sample of women 25 40 years old. Ultrasound Obstet Gynecol 1999; 13: 345 50 8 Metcalf MG, Braiden V, Livesey JH. Retention of normal ovarian function after hysterectomy. J Endocrinol 1992; 135: 597 602 9 Dogan MM, Basaran Z, Ekici E, Basaran O, Ozcan T, Gokay Z, Gokmen O. Effect of hysterectomy on ovarian blood supply and function. J Obstet Gynecol 1998; 18: 263 6 10 Derksen JGM, Brolmann HAM, Wiegerinck MAHM, Vader HL, Heintz APM. The effect of hysterectomy and endometrial ablation on follicle stimulating hormone (FSH) levels up to 1 year after surgery. Maturitas 1998; 29: 133 8 11 Järvelä I, Tekay A, Jouppila P. The effect of a levonorgestrelreleasing intrauterine system on uterine artery blood flow, hormone concentrations and ovarian cyst formation in fertile women. Hum Reprod 1998; 13: 3379 83 12 Barbosa I, Olsson SE, Odlind V, Goncalves T, Coutinho E. Ovarian function after seven years use of a levonorgestrel IUD. Adv Contracept 1995; 11: 85 95 13 Nilsson CG, Lähteenmäki P, Luukkainen T. Ovarian function in amenorrheic and menstruating users of a levonorgestrel-releasing intrauterine device. Fertil Steril 1984; 41: 52 5 14 Silverberg SG, Haukkamaa M, Arko H, Nilsson CG, Luukainen T. Endometrial morphology during long-term use of levonorgestrelreleasing devices. Int J Gynecol Pathol 1986; 5: 235 41 Ultrasound in Obstetrics and Gynecology 385