A randomised trial comparing the levonorgestrel intrauterine system and thermal balloon ablation for heavy menstrual bleeding


Correspondence: Dr MC Sowter, Department of Obstetrics and Gynaecology, University of Auckland, Private Bag 92019, Auckland, New Zealand. Email martinsowter@xtra.co.nz


Objective  To compare the levonorgestrel intrauterine system (LNG-IUS) (Mirena®; Schering Co., Turku, Finland) and thermal balloon ablation (Thermachoice™; Gynecare Inc., Menlo Park, CA, USA) for the treatment of heavy menstrual bleeding.

Design  An open, pragmatic, prospective randomised trial.

Setting  A menstrual disorders clinic at National Women's Hospital, Auckland, New Zealand.

Population  Seventy-nine women with heavy menstrual bleeding randomised to the LNG-IUS (40 women) or the thermal balloon ablation (39 women).

Methods  Women were randomised to treatment with the LNG-IUS or thermal balloon ablation and followed up by a postal and telephone questionnaire.

Main outcome measures  Menstrual loss measured by a pictorial bleeding assessment chart (PBAC) at 3, 6, 12 and 24 months. Patient satisfaction, quality of life and menstrual symptoms were assessed by questionnaire administered at 3, 6, 12 and 24 months. Treatment side effects and treatment failures were also recorded.

Results  Both the treatments resulted in a significant reduction in PBAC scores. At 12 and 24 months, median PBAC scores were significantly lower in women treated with the LNG-IUS compared with women treated by thermal balloon ablation (11.5 versus 60.0 at 12 months [P= 0.002]; 12.0 versus 56.5 [P= 0.002] at 24 months). At 24 months, nine (35%) women still using the LNG-IUS had amenorrhoea compared with one (5%) woman successfully treated by thermal balloon ablation (P= 0.025). There were no significant differences in patient satisfaction between two treatments during follow up. Treatment failed in 11 (28%) women using the LNG-IUS and in 10 (26%) women treated with thermal balloon ablation. Overall, women in both groups showed an increased quality of life as a result of the treatment, with Short Form-36 scores increasing from 63.7 at randomisation to 76.1 at 24 months.

Conclusions  At 12 and 24 months of follow up, women with heavy menstrual bleeding treated with the LNG-IUS have significantly lower PBAC scores than women treated with thermal balloon ablation. Both the treatments resulted in a significant increase in overall quality of life, but there were no significant differences between either treatment in quality of life, patient satisfaction or the number of women requesting an alternative treatment during 24 months of follow up.


Hysteroscopic endometrial resection and ablation have an established role in the treatment of menorrhagia.1,2 However, a high level of operator skill is required to undertake such surgery safely and effectively. In recent years, a range of second-generation or global ablation devices have been developed that can be used by gynaecologists with limited hysteroscopic experience for the treatment of women with menorrhagia. These devices include thermal balloon ablation, microwave ablation, diffused laser energy ablation, bipolar impedance-controlled ablation, cryoablation and hot saline instillation.3–5 Thermal balloon ablation is the oldest of these global ablation therapies. Such a device was first described by Neuwirth et al.,6 and the first clinical trials of a balloon ablation device were reported in 1994.7 Both balloon ablation and other global ablation therapies have been compared in randomised trials with hysteroscopic resection or ablation,8–11 and very similar treatment results have been reported. It is likely that these newer ablation therapies will largely replace operative hysteroscopy for women seeking a surgical treatment for their menorrhagia.

Much less data are available comparing these devices with what is probably the most likely to be used alternative treatment, the levonorgestrel intrauterine system (LNG-IUS). The LNG-IUS is an intrauterine device that releases 20 micrograms of levonorgestrel every 24 hours over 5 years. It was developed in Finland during the 1980s and licensed for contraception in 1990. Worldwide, the number of current LNG-IUS users is more than 4 million,12 and the US food and drug administration approved the LNG-IUS for use as a contraceptive in 2001. The aim of this study was to compare a thermal balloon endometrial ablation device with the LNG-IUS in a prospective randomised trial.


Women were recruited from March 1999 to July 2001 at National Women's Hospital in Auckland, New Zealand. The district health board ethics committee approved the trial protocol. Potential recruits were drawn from outpatient referrals to this hospital, with an invitation letter being sent to women who appeared to meet our inclusion criteria. Women were eligible for entry if they had self-described heavy menstrual bleeding, had completed their family, were 25–50 years old at initial assessment and had a regular cycle, with discrete episodes of menstruation occurring every 3–6 weeks. Exclusion criteria were ultrasound abnormalities (submucousal fibroids, intramural fibroids greater than 3 cm in diameter, large subserosal fibroids, endometrial polyps); laboratory abnormalities (follicle-stimulating hormone level [FSH] higher than 30 iu/l, adverse endometrial histology) and hysteroscopic abnormalities (submucousal fibroids, endometrial polyps), incidental adnexal abnormality on ultrasound, severe intermenstrual bleeding, severe dysmenorrhoea, severe premenstrual pain, chronic pelvic pain, medical contraindications to either study treatment, previous endometrial ablation or resection, uninvestigated postcoital bleeding and untreated abnormal cervical cytology.

Women giving a written informed consent were randomised to either LNG-IUS or thermal balloon ablation. computer-generated randomisation in blocks of 20 had been prepared prior to the commencement of the study and placed in consecutively numbered opaque envelopes. Blinding of patient and doctor to the treatment following the randomisation was not possible.

Treatments were performed in an outpatient setting during the first 10 days of the participant's menstrual cycle. Local anaesthetic (1 ml of 1% lignocaine and adrenaline 1 in 200 000) was injected into each quadrant of the cervix, the cervix was dilated if required and uterine cavity length measured with a uterine sound. All women underwent a diagnostic hysteroscopy using a 4-mm hysteroscope, with 0.9% saline solution to distend the uterine cavity prior to the insertion of either the LNG-IUS or the thermal balloon ablation. Women who could not tolerate hysteroscopy or treatment under local anaesthetic were rescheduled to have their hysteroscopy or treatment under a general anaesthetic on the next available theatre list. The LNG-IUS (Mirena®; Schering Co., Turku, Finland) was inserted as per manufacturers instructions.

All women being treated with thermal balloon ablation also had intravenous access established, and if required, 50–125 micrograms fentanyl citrate (Sublimaze®; Janssen-Cilag Ltd, North Ryde, NSW, Australia) and 1–3 mg midazolam (Hypnovel®; Roche Products (NZ) Ltd, Auckland, New Zealand) were administered. Women were given diclofenac 50 mg (Voltaren EC; Novartis Pharmaceuticals, Auckland, New Zealand) 1 hour before treatment. Thermal balloon ablation was undertaken using the Thermachoice I device (Gynecare Inc., Menlo Park, CA, USA) as per manufacturers instructions.

The objective of the study was to compare the efficacy of LNG-IUS and thermal balloon ablation for the management of heavy menstrual bleeding. The primary outcomes assessed were menstruation, by pictorial bleeding assessment chart (PBAC);13 quality of life, as measured by the Short Form-36 (SF-36) and patient satisfaction, by questionnaire at pre-treatment, 3, 6, 12 and 24 months. Haemoglobin and FSH levels were also measured at pre-treatment, 3, 12 and 24 months. Standardised sanitary products were issued for use during completion of PBAC questionnaires in order to allow for comparisons of the PBAC score to be made. An economic evaluation was also undertaken alongside this trial. These results will be presented in a future study.

Women had direct access to the research nurse throughout the study, and if necessary, alternative management options were discussed, selected and arranged. A treatment failure was deemed to have occurred whenever a major change in treatment was completed. For the LNG-IUS, this was confirmed expulsion, completed removal or the initiation of alternative therapy. For thermal balloon ablation, this was the initiation of medication or the completion of alternative surgery, such as hysterectomy. Treatments were otherwise described as ‘nonfailed’ but with varying levels of participant satisfaction.

For the study to have 80% power to detect a 50-point difference in PBAC score, it was calculated that 30 women would be required for each treatment arm. Chi-square test, t test and Wilcoxon test were used for statistical analysis. This was based on a review of earlier studies of the LNG-IUS and other ablation therapies, which suggested that this was likely to represent a clinically meaningful difference for women who believed that they had problematic menstrual bleeding.14–16


Between March 1999 and July 2001, 177 women were assessed in the menstrual disorders clinic. Seventy-one women (40% of women assessed) were excluded from the study, and 23 of the 106 eligible women (22%) declined randomisation. These 71 women were excluded because they did not have heavy menstrual bleeding (23 women); had not completed their family (23 women); had ultrasound abnormalities (17 women—submucous fibroids, 3 women—endometrial polyp, 1 woman—adnexal mass) or had a laboratory abnormality (three women—FSH level higher than 30 iu/l, one woman—endometrial hyperplasia).

Eighty-three women were randomised, but four women were excluded after randomisation when a hysteroscopic abnormality (submucous fibroids) was found. Forty women were treated with the LNG-IUS, and 39 women were treated by thermal balloon ablation. A trial profile is shown in 1Figure 1. One woman randomised to the thermal balloon ablation group was treated under general anaesthesia after not tolerating the initial part of the balloon ablation procedure. All the initial hysteroscopy procedures and other treatments were undertaken in an outpatient clinic setting.

Figure 1.

Trial profile.

There were no significant differences in patient characteristics following randomisation (1Table 1). Pre-treatment FSH and haemoglobin levels were also similar in the LNG-IUS and thermal balloon ablation groups: FSH, 8.1 versus 8.2 iu/l; haemoglobin, 129 versus 129 g/l.

Table 1.  Baseline characteristics of randomised women prior to treatment. Values are given as n (%) or mean (SD)
CharacteristicsRandomised to LNG-IUS (n= 42)Randomised to TBA (n= 41)
  • TBA, thermal balloon ablation. Body mass index = weight (kg)/height (m2).

  • *

    Only women housebound or suffering disturbed night were included.

Age (years)
 <407 (16.7)13 (31.7)
 40–4421 (50.0)16 (39.0)
 45–4914 (33.3)12 (29.3)
 European31 (73.8)26 (63.4)
 Maori3 (7.1)9 (22.0)
 Other8 (19.1)6 (14.6)
Body mass index28.8 (8.0)29.7 (5.4)
Nulliparous1 (2.4)0 (0)
Sterilisation or vasectomy23 (54.8)17 (41.5)
Sexually active35 (87.5)31 (75.6)
Days of bleeding7.4 (3.3)7.0 (2.6)
Length of cycle (days)25.5 (3.7)24.9 (3.5)
Number of days of heavy bleeding3.8 (2.0)3.4 (1.3)
Number of days of painful bleeding3.8 (1.8)4.1 (2.6)
Unable to leave house on heaviest days27 (65.9)31 (75.6)
Number of days housebound*1.6 (1.4)1.78 (1.4)
Sleep disturbed39 (95.1)39 (95.1)
Number of nights disturbed*2.9 (1.7)2.3 (1.1)
PBAC score490 (419)502 (422)
Endometrial thickness at time of treatment (mm)8.2 (3.3)7.9 (2.7)
Cavity length (cm)8.5 (0.9)8.4 (0.86)

Objectively measured menstrual blood loss assessed by PBAC score was significantly reduced in both the treatment groups compared with pre-treatment scores. At 3 and 6 months, there were no significant differences in PBAC scores between the two groups. At 12 and 24 months, PBAC scores were significantly lower in women treated with the LNG-IUS compared with women treated with thermal balloon ablation (2Table 2). A similar trend was seen in amenorrhoea rates, although these results only became significant at 24 months of follow up, with nine (35%) of the successfully treated women having amenorrhoea on PBAC scores in the LNG-IUS group and one (5%) woman in the thermal balloon ablation group (P= 0.025).

Table 2.  Menstrual status at randomisation and at 3-, 6-, 12- and 24-month follow-up periods (excluding treatment failures)
 LNG-IUS (n= 40)TBA (n= 39)P value
 n (%)Mean (SD)Mediann (%)Mean (SD)Median
  1. TBA, thermal balloon ablation.

3 months
 Amenorrhoea2 (5.6) 5 (16.1) 0.236
 Total PBAC score 125.0 (198.5)52.0 220.8 (438.5)75.00.452
6 months
 Amenorrhoea3 (9.4) 1 (3.1) 0.613
 Total PBAC score 72.1 (118.6)32.0 107.5 (135.4)52.50.080
12 months
 Amenorrhoea6 (20) 2 (6.9) 0.254
 Total PBAC score 41.1 (86.5)11.5 94.7 (112.0)60.00.002
24 months
 Amenorrhoea9 (35) 1 (5) 0.025
 Total PBAC score 20.6 (28.8)12.0 75.4 (91.1)56.50.002

Subjectively measured menstrual symptoms included the number of days of heavy bleeding and the number of days the women were unable to leave home or had sleep disturbed by heavy bleeding (3Table 3). In both groups by 3 months post-treatment, there was a significant reduction in the number of days of heavy bleeding and the proportion of women unable to leave home or experiencing disturbed sleep compared with pre-treatment. There was no significant difference between the two treatment groups in these symptoms over 24 months of follow up.

Table 3.  Menstrual symptoms at randomisation and at 3, 12 and 24 months (excluding treatment failures)
 LNG-IUS (n= 40)TBA (n= 39)
 n (%)Mean (SD)Mediann (%)Mean (SD)Median
3 months3639
 Number of days of heavy bleeding 0.4 (1.0)0 1.2 (1.3)1.0
 Unable to leave house on heaviest days2 (6) 0
 Number of days unable to leave house 0.1 (0.2)0 00
 Sleep disturbed4 (11) 5 (13) 
 Number of nights disturbed 0.3 (0.9)0 0.2 (0.5)0
12 months3132
 Number of days of heavy bleeding 0.5 (1.3)0 1.0 (1.2)0
 Unable to leave house on heaviest days01 (4) 
 Number of days unable to leave house 00 0.1 (0.4)0
 Sleep disturbed1 (3) 4 (13) 
 Number of nights disturbed 0.3 (1.5)0 0.3 (0.9)0
24 months2620
 Number of days of heavy bleeding 0.3 (0.8)0 0.6 (1.1)0
 Unable to leave house on heaviest days00
 Number of days unable to leave house 00 00
 Sleep disturbed01 (5) 
 Number of nights disturbed 00 00

Women were also asked whether they felt that their menstrual symptoms had been successfully treated and whether they would recommend their treatment to a friend. At both 12 and 24 months, there were no significant differences in the number of women who felt that their treatment was definitely a success: 31 women (89%) in the LNG-IUS group and 29 (83%) in the thermal balloon ablation group at 12 months (P= 0.29); 34 women (92%) in the LNG-IUS group and 25 (81%) in the thermal balloon ablation group at 24 months (P= 0.21). Similarly, the number of women who would definitely recommend their treatment to a friend with menorrhagia was not significantly different in each group: 30 women (86%) in the LNG-IUS group and 26 (74%) in the thermal balloon ablation group at 12 months (P= 0.29); 35 women (95%) in the LNG-IUS group and 26 (84%) in the thermal balloon ablation group at 24 months (P= 0.20).

Treatment failures were assessed at 3, 6, 12 and 24 months. In the LNG-IUS group, the 11 treatment failures included the following: by 3 months, one LNG-IUS expelled and two removed because of pain; by 6 months, no further treatment failures; by 12 months, one woman with menorrhagia, two LNG-IUS removed because of troublesome, unscheduled bleeding and two further LNG-IUS expelled and by 24 months, one woman with menorrhagia, one with actinomycoses and one LNG-IUS expelled. In the thermal balloon ablation group, the ten treatment failures included the following: at 3 months, no treatment failures; at 6 months, two women with menorrhagia; at 12 months, two women with menorrhagia and at 24 months, five women with menorrhagia and one with dysmenorrhoea. Among the 11 treatment failures in the LNG-IUS groups, three women had a thermal balloon ablation, one had a microwave endometrial ablation, one was prescribed the oral contraceptive pill, one was prescribed tranexamic acid, three underwent hysterectomy and two had no further treatment. Among the ten treatment failures in the thermal balloon ablation group, five women were treated with the LNG-IUS and five underwent hysterectomy.

There were no serious complications in either treatment group. One woman had her LNG-IUS removed at 19 months when actinomycoses were cultured. In the thermal balloon ablation group, five women were prescribed postoperative antibiotics for possible endometritis.

Quality of life was assessed using the SF-36 at randomisation, 3, 12 and 24 months. Mean SF-36 scores are shown in 4Table 4 overall and for each treatment. For women who failed the treatment, the SF-36 at the time of last assessment was used in subsequent assessments as a proxy for their quality of life had they continued with treatment. The results suggest that there was a significant increase in quality of life between randomisation and 3 months for both the LNG-IUS (t= 4.83) and the thermal balloon ablation (t= 5.44) groups; between randomisation and 12 months for both the LNG-IUS (t= 5.36) and the thermal balloon ablation (t= 4.43) groups and between randomisation and 24 months for both the LNG-IUS (t= 4.57) and the thermal balloon ablation (t= 3.34) groups. There were no significant differences between treatments at any of the assessments.

Table 4.  Quality of life (SF-36) at randomisation and at 3, 12 and 24 months (including treatment failures). Values are given as mean (SD)
SF-36 scoresOverall (n= 79)TBA (n= 39)LNG-IUS (n= 40)
  1. TBA, thermal balloon ablation.

Randomisation63.7 (19.0)63.7 (14.4)63.7 (22.7)
3 months77.9 (13.7)78.2 (13.7)77.7 (17.0)
12 months78.1 (16.8)76.9 (16.8)79.3 (16.5)
24 months76.2 (18.6)74.9 (18.8)77.5 (20.1)


Both the treatments compared in this study are well established for the treatment of menorrhagia, but there have been few direct comparisons of the two techniques, and to date, only two other randomised trials comparing the LNG-IUS with balloon ablation have been published. Soysal et al.17 randomised 72 women to treatment with either the LNG-IUS or the thermal balloon ablation with the Thermachoice device. At 12 months post-treatment, the reduction in PBAC scores was significantly greater in the thermal balloon ablation group compared with the LNG-IUS group (a reduction of 388.2 [SD 21] versus 343 [SD 27]; P < 0.001). Patients treated by thermal balloon ablation also reported significantly fewer adverse effects and higher health-related quality of life (assessed by SF-36) in physical role functioning. In another study, Barrington et al.18 compared the LNG-IUS with the Thermachoice device in 50 women using PBAC scores at pre-treatment and at 6 months post-treatment. At 6 months, PBAC scores were not significantly different in the two groups. Three of the 25 women treated with the LNG-IUS requested removal or hysterectomy, and 5 of the 25 women treated with thermal balloon ablation requested hysterectomy. A third multicentre randomised study in Great Britain, the SMART trial, was abandoned due to poor recruitment, a problem not experienced in this New Zealand study.19,20

No medical treatment or conservative surgery can rival hysterectomy in achieving 100% cessation of bleeding; yet, randomised studies reporting on the quality of life suggest that women value the improvements in menstruation reported with LNG-IUS in spite of continuing bleeding.21 The largest randomised study to date comparing hysterectomy with LNG-IUS has recently reported 5 years of follow up.12,22 Among 119 women randomised to LNG-IUS, 50 (42%) eventually underwent hysterectomy, so the LNG-IUS is clearly not a completely effective treatment for menorrhagia. Despite this relatively high proportion of women eventually requesting hysterectomy, satisfaction rates were similar in both groups and the discounted direct and indirect costs associated with the LNG-IUS remained significantly lower than those associated with hysterectomy at $2817 (95% CI $2222–$3530) versus $4660 (95% CI $4014–$5180). Arguably, women randomised in this study were willing to consider hysterectomy at trial entry and so were a different population from those entering a study comparing an ablation therapy with the LNG-IUS.

Five randomised controlled trials have compared endometrial resection with hysterectomy.23 A systematic review of the five trials reported that both procedures are effective and their satisfaction rates are high. Although hysterectomy is associated with a longer operating time, a longer recovery period and higher rates of postoperative complications, it offers permanent relief from heavy menstrual bleeding. The cost of endometrial destruction is significantly lower than the cost of hysterectomy, but since re-treatment is often necessary, the cost difference narrows over time. A number of randomised trials have compared thermal balloon ablation9,15,24 and other second-generation devices25 with hysteroscopic endometrial resection or ablation and not shown any significant differences in effectiveness. No randomised trials has compared second-generation devices with hysterectomy. The variety of second-generation devices that have been developed and the apparent effectiveness of the LNG-IUS make such a comparison less clinically useful than comparisons between individual second-generation ablation devices or with the LNG-IUS.

Some devices such as microwave endometrial ablation (MEA™; Microsulis, Portsmouth, UK)26,27 may be more versatile than thermal balloon ablation in that larger or distorted uterine cavities can be treated. It is also possible that some newer devices will result in higher rates of amenorrhoea and less menstrual blood loss than thermal balloon ablation. NovaSure (Cytyc, Palo Alto, CA, USA),10,28 a global ablation device that uses bipolar electrosurgical energy, has been shown in a randomised trial to produce a significantly higher rate of postoperative amenorrhoea over 12 months of follow up than Cavaterm (Wallsten Medical SA, Morges, Switzerland),29 a thermal balloon ablation device (16 of 37 women [43%] versus 2 of 18 women [11%]). Novasure has also recently been compared with Thermachoice in a larger, double blind randomised trial, and at 12 months of follow up, amenorrhoea rates were 43% in the Novasure group (34 of 83 women) compared with 8% (3 of 43 women) in the Thermachoice group (P < 0.01), with significantly higher rates of postoperative satisfaction also reported.30 Refinements in the Thermachoice device including circulating the fluid within the balloon and the development of other balloon devices may improve amenorrhoea rates and reduce objectively measured postoperative menstrual blood loss, although to date, only one small trial has compared two different balloon devices and not demonstrated any difference in effectiveness.31 The reversibility of treatment with LNG-IUS, its contraceptive efficacy and the opportunity it provides to maintain future fertility will make it an attractive alternative to ablation therapy even if other devices do prove to be more effective than thermal balloon ablation.

Importantly, this study demonstrates that 24 months of follow up, treatment with the LNG-IUS is associated with lower PBAC scores and a higher amenorrhoea rate than thermal balloon ablation. Most randomised and observational studies of new endometrial ablation therapies have a relatively short duration of follow up of no more than 12 months, and this study highlights the importance of a long duration of follow up when the LNG-IUS and the ablation therapy are compared.

In conclusion, this study has shown that the LNG-IUS results in a significantly lower objectively measured blood loss and higher amenorrhoea rates than thermal balloon ablation over 24 months of follow up, but both the treatments are associated with similarly high levels of patient satisfaction and quality of life. However, following treatment with both LNG-IUS and thermal balloon ablation, at least 25% of women with a normal-sized uterus and menorrhagia will have requested an alternative treatment within 2 years. Neither treatment is clearly superior to the other, and the treatment chosen should be based on the individual preferences of each woman.