Ten-year follow-up of a randomised controlled trial comparing bipolar endometrial ablation with balloon ablation for heavy menstrual bleeding


Correspondence: Dr MC Herman, Department of Obstetrics and Gynecology, Máxima Medical Centre, PO Box 777, 5500 MB, Veldhoven, the Netherlands. Email m.herman@mmc.nl



Previously, we have reported that, at both 12 months and 5 years after treatment, bipolar endometrial ablation is superior to balloon ablation in the treatment of heavy menstrual bleeding. In this article, we evaluate the results at 10 years after these interventions.


Ten-year follow-up of a double-blind randomised controlled trial.


A teaching hospital in the Netherlands.


Premenopausal women suffering from heavy menstrual bleeding.


A follow-up questionnaire was sent to women 10 yearsafter randomisation for bipolar ablation and balloon ablation (2 : 1 ratio).

Main outcome measures

Amenorrhoea rates, re-intervention and patient satisfaction.


At 10 years of follow-up, the response rate was 69/83 (83%) in the bipolar group and 35/43 (81%) in the balloon group. Amenorrhoea rates were 50/69 (73%) in the bipolar group and 23/35 (66%) in the balloon group [relative risk, 1.1 (95% CI, 0.83–1.5)]. Further treatment following initial ablation was reported in 21 cases, 14 in the bipolar group and nine in the balloon group [relative risk, 0.9 (95% CI, 0.63–1.3)]. Eight of these women required further treatment after 5 years, including two hysterectomies. Patient satisfaction in the bipolar group was 81% (56/69) compared with 77% (27/35) in the balloon group [relative risk, 1.1 (95% CI, 0.82–1.2)].


Ten years after treatment, the superiority of bipolarablation over balloon ablation in the treatment of heavy menstrual bleeding was no longer evident.


Heavy menstrual bleeding (HMB) is a common gynaecological problem.[1] In 2004, we reported a randomised controlled trial comparing bipolar endometrial ablation (NovaSure®) and balloon ablation (Thermachoice®). The primary outcome measure was amenorrhoea and the secondary outcome measures were patient satisfaction, re-intervention and health-related quality of life (HRQoL). This trial concluded that bipolar endometrial ablation was more effective than balloon ablation in the treatment of HMB within a follow-up of 12 months.[2] At 5 years of follow-up, bipolar ablation remained superior to balloon ablation.[3] In the present study, we report the effectiveness of both treatments at 10 years after randomisation.


This report was based on a randomised controlled trial comparing the bipolar radiofrequency impedance-controlled endometrial ablation device (NovaSure) and the thermal balloon ablation device (Thermachoice), which was performed in the Maxima Medical Centre in the Netherlands between November 1999 and July 2001. This trial has been described in detail previously by Bongers and coworkers.[2, 3] Women with HMB as indicated on the pictorial chart described by Higham et al.,[4] with a minimum score of 150 points and no intra-cavitary pathology, were eligible for the trial.[2, 3] The trial included 126 women: 83 were allocated to the bipolar group and 43 to the balloon group. The trial flowchart is shown in Figure 1.

Figure 1.

Trial profile. †Number of women who died during the study follow-up.

The primary outcome for this long-term follow-up study was the amenorrhoea rate and secondary outcomes were patient satisfaction and re-intervention rate in keeping with our earlier reports. Postal questionnaires were sent from November 2010 until January 2011. The mean follow-up time was 9.8 years (range, 9–11 years), which we have approximated as a 10-year follow-up. If questionnaires were not returned after a reminder in 4 weeks, we tried to contact the patients by telephone. We performed telephone interviews. Instead of the pictorial chart, we asked the women about their menstruation in days, clots and dysmenorrhoea. In the case of missing addresses or telephone numbers, we contacted the women's general practitioners and also searched their medical records for information about re-interventions in other hospitals.


We analysed all women who were randomised according to the intention-to-treat principle. After the treatment of 44 women, a technical failure in the bipolar ablation device was detected, which was solved by replacement of the bipolar device.[2] We therefore decided to perform a second analysis that was limited to the women who were treated after replacement of the bipolar ablation device.

Women who underwent a hysterectomy after ablation therapy were analysed as nonamenorrhoeic and as being dissatisfied with their ablation treatment. We analysed the presence of clots and dysmenorrhoea only for those women who had not undergone a hysterectomy. Time to hysterectomy and time to amenorrhoea were compared by Kaplan–Meier analysis. Dichotomous outcomes were compared by calculating the relative risk (RR) and its 95% confidence interval (95% CI). Logistic regression analysis was used to predict the outcome of amenorrhoea.


In November 2010, we sent 125 questionnaires, as one of the 126 women had indicated that she no longer wished to participate at 5 years of follow-up; 96 women returned their questionnaires; eight women participated in a telephone interview; two women had died. Overall, 22 women, 14 in the bipolar group and eight in the balloon group, were lost to follow-up, resulting in 10-year follow-up rates of 83% (n = 69) in the bipolar group and 81% (n = 35) in the balloon group. The baseline characteristics showed a median age of 54 years and a difference only in the position of the uterus; there were more women with a retroverted uterus in the bipolar group (16% versus 9%).


At 10 years of follow-up, the number of women with amenorrhoea in the bipolar ablation group was 50/69 (73%) compared with 23/35 (66%) in the balloon group [RR, 1.1 (95% CI, 0.83–1.5)]. After exclusion of the women who were randomised before the failure in the bipolar device was detected, amenorrhoea rates in both the bipolar and balloon groups were somewhat higher: 38/47 (81%) compared with 16/21 (76%) [RR, 1.1 (95% CI, 0.81–1.4)]. Table 1 shows the percentage of women with amenorrhoea and the presence of dysmenorrhoea and clots after bipolar ablation and balloon ablation. Figure 2 shows the time to amenorrhoea after bipolar and balloon ablation.

Table 1. Comparison of bipolar and balloon ablation: results at 10 years of follow-up
 Bipolar A groupa (n = 69) (%)Balloon A groupa (n = 35) (%)RR (95% CI)Bipolar B groupa (n = 47) (%)Balloon B groupa (n = 21) (%)RR (95% CI)
  1. CI, confidence interval; RR, relative risk.

  2. a

    Group A, analysis of all randomised women. Group B, analysis limited to those women who were randomised after the failure of the Novasure generator had been corrected.

5 years39/81 (48)12/39 (32)1.6 (0.93–2.6)34/53 (64)9/24 (37)1.7 (0.98–3.0)
10 years50 (73)23 (66)1.1 (0.83–1.5)38 (81)16 (76)1.1 (0.81–1.4)
Presence of dysmenorrhoea
5 years10/68 (14)7/29 (25)0.61 (0.26–1.4)6/45 (13)4/16 (25)0.53 (0.17–1.7)
10 years6/59 (10)4/30 (13)1.0 (0.88–1.2)2/43 (5)1/19 (5)1.0 (0.89–1.0)
Presence of clots
5 years6/63 (9)10/29 (32)0.28 (0.11–0.69)3/43 (6)7/18 (33)0.18 (0.05–0.62)
10 years4/59 (7)3/30 (10)0.7 (0.16–2.8)2/43 (5)2/19 (11)0.44 (0.067–2.9)
Figure 2.

Kaplan–Meier curves showing the cumulative amenorrhoea rate (log rank, 0.16) and cumulative hysterectomy rate (log rank, 0.91) after bipolar and balloon ablation.

By stratifying the analysis for the median age of 54 years, the amenorrhoea rates were 67% (35/52) in the <54-year age group and 73% in the >54-year age group (38/52) [RR, 0.9; (95% CI, 0.73–1.2)]. The logistic regression analysis showed a significant effect of age on amenorrhoea (= 0.001), but no interaction between age and treatment (P = 0.56).


There were 23 women who required further treatment, 14 in the bipolar group and nine in the balloon group [RR, 0.9 (95% CI, 0.63–1.3)]. Ten women in the bipolar group underwent a hysterectomy compared with five in the balloon group [RR, 1.0 (95% CI, 0.69–1.49)]. When we limited this analysis to women who were included after repair of the failure, these rates were five and two, respectively [RR, 1.0 (95% CI, 0.63–1.7)]. Since the 5-year follow-up, only two more women in the bipolar group and none in the balloon group underwent a hysterectomy. The lifetime table analysis is shown in Figure 2. In the bipolar group, one woman received re-ablation with the bipolar device, and one received a Mirena® intrauterine device. In the balloon group, four women received hormonal therapy. All these treatments were performed at 5 years after the initial treatment.

Of the 10 hysterectomies that were performed in the bipolar group, five were performed because of persisting HMB. Other reasons were dysmenorrhoea (n = 1), cyclic abdominal pain as a result of haematometra resulting from cervical stenosis (n = 1), atypia of the endometrium (n = 1), abdominal pain (n = 1) and a myoma nascens (n = 1). Histopathology in the bipolar group showed adenomyosis (n = 2), fibroids (n = 3), a combination of adenomyosis and fibroids (n = 2) and no abnormalities (n = 2), and one missing pathological examination.

The five hysterectomies in the balloon group were performed for HMB (n = 3), abdominal pain of unknown cause (n = 1) and pelvic organ prolapse (n = 1). Histopathology showed fibroids (n = 1), a combination of adenomyosis and fibroids (n = 1) and no abnormalities (n = 3).


Figure 3 shows the satisfaction rates at 10 years of follow-up. Patient satisfaction in the bipolar group is 81% (56/69) compared with 77% (27/35) in the balloon group [RR, 1.1 (95% CI, 0.82–1.2)]. If we exclude women treated before the detection of the bipolar device defect, these rates were 86% (42/49) and 85% (17/20), respectively [RR, 1.1 (95% CI, 0.81–1.3)]. More women were completely satisfied in the bipolar group compared with the balloon group: 73.9% (51/69) and 48.6% (17/35), respectively [RR, 1.1 (95% CI, 1.1–2.2)].

Figure 3.

Patient satisfaction at 10 years after treatment.


Main findings

In this follow-up study, we compared the treatment of two second-generation endometrial ablation techniques in women with HMB, 10 years after treatment. In view of the 1-year and 5-year follow-up results, the 10-year follow-up results indicated an increase in amenorrhoea rates for both ablation groups from 48% and 32% at 5 years to 73% and 66% at 10 years in the bipolar and balloon groups, respectively [RR, 1.1 (95% CI, 0.83–1.5)]. Although the amenorrhoea rates were still higher in the bipolar group, they were neither statistically nor clinically significant. A high satisfaction level was found for both groups, with a significantly higher rate of completely satisfied women in the bipolar group.

Only eight additional re-interventions were required after 5 years of follow-up and no significant difference in re-intervention was found between the two groups. The Kaplan–Meier curve shows that, for both the ablation procedures, hysterectomies were mainly performed during the first 5 years after treatment. Most women had a hysterectomy because of persisting HMB. The pathology of hysterectomy specimens showed a variety of conditions, but no differences between the groups.

Strengths and weaknesses

This is one of the few 10-year follow-up studies in endometrial ablation and the follow-up rates are high. Long-term follow-up data are important to better inform women of the likely outcomes and to guide clinicians. However, many women became menopausal during follow-up, so that the relevance of amenorrhoea as a primary outcome measure is questionable. The effect of menopause in many participants may explain the lack of difference in clinical outcomes at 10 years in contrast with those observed in our earlier reports at 1 and 5 years of follow-up.[2, 3] The generalisability of our findings is limited because the balloon ablation device (Thermachoice I) used in our trial now no longer exists, being superseded by newer versions.


The increase in the amenorrhoea rate can be explained by the fact that most women became postmenopausal during the follow-up period. After stratifying for age, we still found comparable amenorrhoea rates in the two groups, and no interaction between age and treatment was observed. Nevertheless, this study provides more information 10 years after treatment. Although amenorrhoea was the primary outcome in the initial study, for the long-term follow up, the satisfaction rate and the hysterectomy rate are more useful for the evaluation of these ablation techniques. As we found that most re-interventions took place during the first 5 years after treatment, we conclude that amenorrhoea or satisfaction after short-term follow up can help to predict long-term success in terms of the need for re-intervention.

It was reassuring to note that satisfaction with treatment remained high 10 years following initial endometrial ablation for HMB. Only a few published trials have reported satisfaction rates. Compared with the 10-year follow-up study of Sambrook et al.,[5] which describes total or general satisfaction levels of 50–60%, both groups in this trial were found to be doing even better.


The superiority of bipolar ablation over balloon ablation in the treatment of HMB, which was observed at 1 and 5 years after initial treatment, was no longer evident after 10 years.

Disclosure of interests

No potential conflict of interest exists for any of the authors.

Contribution to authorship

All authors participated in the design, implementation and writing of the manuscript. All authors saw and approved the final version.

Details of ethics approval

Approval to conduct this study was obtained from the Maxima Medical Centre Ethics Committee in 1998. Reference number 9921.


The NovaSure® devices were provided free of charge by Novacept (Palo Alto, CA, USA). The Thermachoice® devices were discounted. The study has not been otherwise supported. Data collection, data analysis and the writing up of this study were undertaken independent of Novacept or Thermachoice.



Commentary on ‘Ten-year follow-up of a randomised controlled trial comparing bipolar endometrial ablation with balloon ablation for menorrhagia’

Second-generation endometrial ablation has superseded hysteroscopic resection and ablation because of improved feasibility, safety and equivalent effectiveness (Daniels et al., BMJ 2012;344:e2564.). However, head-to-head comparisons between technologies are lacking. This randomised controlled trial compared two commonly used ablative techniques. The 1- and 5-year data found amenorrhoea rates to be higher in women treated with radiofrequency ablation (RFA) as opposed to thermal balloon ablation (TBA) (Bongers et al., BJOG 2004;111:1095–102 and Kleijn et al., BJOG 2008;115:193–98). However, although amenorrhoea is an easily measured endpoint, its use as a primary outcome is questionable. This is because amenorrhoea is not always desired by women and, indeed, may be detrimental to the patient's psyche. The aim of treatment should be to improve health-related quality of life (HRQL), regardless of whether amenorrhoea results. Furthermore, the mean age of the women was 43 years at inception and therefore, at 10 years, many would be postmenopausal, which is likely to explain the observed increase in amenorrhoea rates over time in both treatment groups and the loss of superiority of RFA over TBA previously observed at 1 and 5 years. There was no significant difference in the amenorrhoea rates (69% versus 67%) when comparing the RFA group with the TBA group at 10 years, whereas the amenorrhoea rates were 41% versus 7%, respectively, at 1 year and 47% versus 30%, respectively, at 5 years.

Amenorrhoea may be a suitable short-term outcome but, for long-term follow-up, other measures, including HRQL, satisfaction and hysterectomy rates, are more useful endpoints to compare. The authors of the current study have recognised this and reported patient satisfaction rates of around 80% at 10 years and re-intervention rates. Of the re-interventions, two additional hysterectomies were observed in the RFA group after 5-year follow-up and none in the TBA group, representing an increase in the whole cohort from 13 (10%) hysterectomies after 5 years to 15 at 10 years (12%) of follow-up. All of the hysterectomies associated with bleeding in both groups occurred in the first 5 years.

However, the current trial does not report HRQL data. In a study comparing four endometrial ablation techniques in 139 women, Abbott et al. (J Am Assoc Gynecol Laparosc 2003;10:491–5) found that quality of life, including pleasure, habit and discomfort scores measured preoperatively using standard evaluation tools, and sexuality questionnaires were substantially improved 12 months after treatment. No significant differences were found in quality-of-life scores between women after endometrial ablation and the general population without abnormal bleeding. An overall, equivalent improvement in HRQL has been reported in another randomised controlled trial comparing RFA and TBA, but in an outpatient setting (Clark et al., Obstet Gynecol 2011;117:109–18).

The authors of the present study deserve much credit for conscientiously following such a very large percentage of their original patients, although some observer bias may have been introduced through the use of telephone follow-up when nonresponse from postal questionnaires was encountered. Long-term follow-up data from well-conducted studies, such as this, are often unavailable. Data from this trial will help physicians to counsel their patients about short- and long-term outcomes and, in particular, to not select ablative devices based on short-term amenorrhoea rates alone.

Disclosure of interests

I am a consultant for Conceptus, Boston Scientific and Aegea Medical. I serve on the Editorial Board of the Journal of Minimally Invasive Gynecology.

  • MH Glasser

  • Department of Obstetrics and Gynecology, Kaiser Permanente Medical Center, San Rafael, CA, USA