Key words
radiofrequency endometrial ablation - hypermenorrhea - menorrhagia - fibroids
Introduction
Heavy and prolonged menstrual bleeding is a common disorder and makes up one fourth
of the indications for hysterectomies [1].
The treatment spectrum for excessively heavy menstrual bleeding includes medical and
surgical therapeutic approaches. Drug treatments include the atrophying effect of
progestogens perorally or as an intrauterine pessary [2]. In the case of bleeding disorders due to a uterus myomatosus, the selective progestogen
receptor modulator ulipristal acetate was used until recently with high rates of amenorrhoea
[3], [4]. However, the new prescription of ulipristal acetate is no longer recommended in
Switzerland since February 2018 by the European Pharmacovigilance Risk Assessment
Committee (PRAC), since severe liver failure was described in four cases and in three
of the cases, a liver transplantation was needed [5]. According to the German Federal Institute for Drugs and Medical Devices (BfArM),
ulipristal acetate can currently be used, subject to risk conditions, such as regular
liver function tests before, after and during the treatment, provided that there is
no pre-existing impaired liver function [6]. Surgically, diagnostic as well as therapeutic curettage has been indicated, however
with poor long-term results regarding the amount of flow [7]. Up until 30 years ago, in the event of therapeutic failure, there was no other
surgical alternative than hysterectomy. In the 1980s, the methods of hysteroscopic
endometrial ablation using YAG laser, transcervical endometrial resection or “rollerball”
electrocoagulation were developed [8], [9]. However, these require visualisation of the uterine cavity and an experienced surgeon.
Eight randomised studies investigated the safety, efficacy and costs of endometrial
resection as an alternative method to hysterectomy in the treatment of bleeding disorders
[10], [11], [12]. Over the long term, it was seen that approximately 15% of the patients with endometrial
resection via first-generation methods still required a hysterectomy and patient satisfaction
4 months postoperatively was in favour of hysterectomy. The Cochrane analysis from
2016 concluded that endometrial resection, endometrial ablation and progestogen IUD
placement offer a less invasive and yet effective treatment option, in comparison
to hysterectomy [13].
In the 1990s, various non-hysteroscopic alternatives came onto the market. They are
fundamentally easier to use and significantly shorten the surgical time [14], [15], [16], [17]. The most frequently used methods of “blind” endometrial ablation are those of the
second generation: thermal ablation using an intrauterine balloon catheter (Thermachoice® and Cavaterm®) [18], [19], [20], microwave ablation (Microsulis Microwave Endometrial Ablation (MEA) system) [21], [22], bipolar radiofrequency ablation (NovaSure®) [15], diode laser ablation (ELITT™) [23], [24], cryoablation (HerOption™) [25] and hydrothermal ablation [24]. A list of first-generation and second generation techniques is given in [Table 1].
Table 1 1st and 2nd generation methods [32].
|
1st generation
|
2nd generation
|
|
Electrosurgery:
Laser
|
Thermal balloon:
-
Therma Choice®
-
Cavatherm®
-
(Vestablate®)
Hydrothermal ablation
Bipolar endometrial coagulation:
Microwave ablation (MEA)
Diode laser ablation (ELITT®)
|
These different technologies were compared with one another in various randomised
studies [19], [20], [22], [24]. Loffer and Grainger [20] were able to show, in a follow-up after 3 and 5 years, that endometrial ablation
using balloon hyperthermia (Thermachoice®) is equivalent to rollerball endometrial coagulation. Patient satisfaction as well
as the rate of amenorrhoea were evaluated for this purpose. Cooper et al. [22] demonstrated that the efficacy as well as patient satisfaction following transcervical
resection of the endometrium were significantly lower than after microwave ablation
[12]. A meta-analysis by Daniels et al. [26] was able to show that endometrial ablation using bipolar radiofrequency ablation
with regard to the amenorrhoea rate after 12 months is more effective than endometrial
ablation with intrauterine balloon catheters, hydrothermal ablation or cryoablation.
Microwave ablation was also superior to endometrial ablation with an intrauterine
balloon catheter as well as cryoablation, however not intrauterine laser ablation.
The intrauterine laser ablation, by contrast, showed higher 12-month amenorrhoea rates
than microwave ablation, endometrial ablation with intrauterine balloon catheters,
cryoablation and hydrothermal ablation. However, there is a lack of data for a comparison
of laser ablation with bipolar radiofrequency ablation.
The objective of this prospective cohort study was to identify prognostic factors
for successful treatment of abnormal menstrual bleeding using bipolar radiofrequency
endometrial ablation (NovaSure®) in a multivariate analysis.
Materials and Methods
All patients who were scheduled for bipolar radiofrequency endometrial ablation in
the electrosurgery programme between January 2009 and June 2016 at the Frauenfeld
Cantonal Hospital, were prospectively surveyed. The study was approved by the local
ethics committee (reference #01.53.01) and corresponding informed consent forms from
the patients are on hand.
The data collected preoperatively were age, parity, body mass index (BMI) as well
as the findings collected by one of the authors of the transvaginal ultrasound examination
with measurement of the uterus, its dimensions, and measurement of any masses in the
uterine wall or intracavitary space. The preoperative examination was performed with
the aim of excluding patients with an excessively large uterine cavity (length: maximum
6.5 cm, minimum 4 cm, width: maximum 4.5 cm, minimum 2.5 cm). In addition, patients
with an apparent deformation of the cavity due to intramural masses, endometrial polyps,
uterine septae or a bicornuate uterus were excluded. The endometrial ablation with
the NovaSure® system was performed under anaesthesia (laryngeal mask, intubation anaesthesia, spinal
anaesthesia). The mean duration of surgery was 20 minutes. After hysteroscopy and
curettage of the corpus were performed, the surgical technique was performed analogously
to the manufacturerʼs guidelines (NovaSure®, Hologic Inc., Marlborough, MA, USA).
Intraoperative procedure
The cervical canal and the cavity were initially measured using the hysterometer and
subsequent dilation of the cervical canal using Hegar dilators up to 8 mm. Then hysteroscopy
with curettage of the endometrium was performed.
The length of the cavity on the NovaSure® disposable instrument is adjusted by subtracting the cervical canals from the hysterometer,
whereby a max. length of 6.5 cm is possible. After introduction of the disposable
instrument into the cavity, the introducer sheath is withdrawn so that the gold mesh
electrode is exposed in the cavity. By advancing the stamp as well as horizontal and
vertical movements of the Novasure® disposable instrument, an optimal triangular expansion of the frame on which the
gold mesh is attached can take place. The width of the mesh between the two interstitial
portions of the tube should be a minimum of 2.5 cm and a maximum of 5 cm; this can
be read directly in the handle of the instrument. The width and length of the uterine
cavity determined are now set on the device and the flow of energy is adjusted individually
for each patient. To check the intactness of the cavity and for approximation of the
endometrium to the mesh, a vacuum is generated. The bipolar radiofrequency ablation
is performed under vacuum aspiration of the uterine wall and blood and secretions
during the coagulation are also aspirated from the cavity using the vacuum. After
removal of the gold mesh, a control hysteroscopy with photo documentation of the uterine
cavity is performed.
During the follow-up examinations, patients were asked about the heaviness of the
current menstrual flow: no flow, spotting, light menstrual flow (maximum of 3 pads/tampons
per day), moderate menstrual flow (maximum of 6 pads/tampons per day) and excessive
menstrual flow (> 6 pads/tampons per day). The treatment was classified as successful
if spotting or amenorrhoea occurred postoperatively. The follow-ups took place according
to the attending physician. For study participants who did not wish to undergo a follow-up
examination by a study doctor (n = 106), the follow-up data were queried annually,
which is also why no quality-of-life data could be collected.
Statistics
The evaluations and statistical analyses were performed with the statistics software
Stata 13 (StataCorp LLC, College Station, Texas, USA). For all continuous variables,
the normality test was performed. The multivariate analysis which includes the variables
of age, BMI, parity, hysterometer and intramural masses was performed using a logistical
regression analysis, whereby the method of gradual backward elimination was used.
An alpha value of less than 0.05 was defined as significant.
Results
Between January 2009 and June 2016, a total of 207 women were included in the study.
A bipolar radiofrequency endometrial ablation was planned for these women. Seven women
were unable to undergo the planned radiofrequency ablation for technical reasons (failure
to generate a vacuum). In another 5 cases, the procedure could not be performed for
anatomical reasons (cavity length less than 4 cm and/or cavity width less than 2.5 cm).
In these cases, bipolar endometrial resection with the resectoscope was performed
under the same anaesthesia. Radiofrequency ablation was performed in 195 patients
and 187 patients came to follow-up examinations.
Patients
The mean age of the patients was 44 years, the mean parity was 2 and the median BMI
was 23.5 kg/m2 ([Table 2]). The length of the uterine cavity, measured using a hysterometer, was 8.7 cm (SD
± 1.1). In 31 of 195 patients (15.9%) an intramural mass was diagnosed preoperatively
on the transvaginal ultrasound, consistent with a uterine fibroid.
Table 2 Patient characteristics and follow-up observation period (n = 195).
|
Variable
|
Values
|
|
* Average (standard deviation; distribution), † Median (interquartile range; distribution), § Data missing for 8 patients
|
|
Age – years
|
44 (± 5.3; 25 – 55)*
|
|
Hysterometer – cm
|
8.7 (± 1.14; 5 – 12)*
|
|
BMI – cm/m2
|
23.5 (21 – 27; 16.6 – 47.2)†
|
|
Parity
-
Para 0 (10.1%)
-
Para 1 (12.8%)
-
Para 2 (44.7%)
-
Para 3 (32.4%)
|
2 (2 – 3; 0 – 3)†
|
|
Suspected fibroid on ultrasound n (%)
|
31 (15.90)
|
|
Diameter of intramural mass – mm †
|
15 (10 – 24; 5 – 50)†
|
|
Follow-up period – months †
|
17.5 (4.5 – 34.9; 1 – 82)†
|
Outcome
With a median follow-up period of 17.5 months, 86% of patients reported amenorrhoea
or cyclical spotting: 88 patients (47.1%) indicated amenorrhoea, 77 (41.2%) spotting,
16 (8.6%) normally heavy menstrual bleeding and 6 (3.2%) persistent hypermenorrhoea.
Two patients (1%) had to be treated postoperatively with antibiotics due to a pelvic
infection.
During the follow-up period, 10 patients (5%) underwent a hysterectomy. The reasons
for this were as follows: 3 patients had pelvic organ prolapse with unremarkable histology
of the uterus (18 – 48 months after endometrial ablation), one patient had persistent
lower abdominal pain with adenomyosis and 6 patients had persistent uterine bleeding
(two with uterus myomatosus, one with adenomyosis, one with atypical endometrial hyperplasia
and two with unremarkable histology).
In the multivariate analysis, the preoperative presence of an intramural mass was
significantly associated with therapeutic failure among the 25 – 45-year-old patients,
although its diameter was only a median of 15 mm ([Tables 3] and [4]).
Table 3 Success rate in the setting of fibroid suspected on ultrasound (n = 187).
|
Fibroid
|
≤ 45 years (n = 111)
|
> 45 years (n = 76)
|
Total* (n = 187)
|
|
|
# Success
|
%
|
# Success
|
%
|
# Success
|
%
|
|
* Data for 8 patients missing, success defined as postoperative amenorrhoea or spotting
and no repeat surgery
|
|
No
|
83/94
|
88.3
|
55/63
|
87.3
|
138/157
|
87.9
|
|
Yes
|
12/17
|
70.6
|
11/13
|
84.6
|
23/30
|
76.7
|
|
Total
|
95/111
|
85.6
|
66/76
|
86.8
|
161/187
|
86.1
|
Table 4 Failure depending on age, presence of fibroids, parity, BMI and the hysterometer,
multivariate analysis of the risk factors (n = 187§).
|
Risk factors
|
≤ 45 years (n = 111)
|
> 45 years (n = 76)
|
|
|
Hazard ratio
|
p-value*
|
95% confidence interval
|
Hazard ratio
|
p-value*
|
95% confidence interval
|
|
* p-value: multivariate logistical regression model, § Data missing for 8 patients
|
|
Fibroid
|
3.699
|
0.036
|
1.089 – 12.570
|
0.662
|
0.729
|
0.064 – 6.816
|
|
Parity
|
1.038
|
0.903
|
0.570 – 1.889
|
1.849
|
0.122
|
0.849 – 4.025
|
|
BMI
|
1.063
|
0.239
|
0.960 – 1.176
|
1.050
|
0.333
|
0.952 – 1.158
|
|
Hysterometer
|
0.984
|
0.949
|
0.595 – 1.627
|
0.767
|
0.464
|
0.377 – 1.560
|
Accordingly, the risk of a hysterectomy performed during the follow-up period was
increased 8-fold in the multivariate analysis in the younger patients if intramural
masses could be identified preoperatively ([Table 5]).
Table 5 Likelihood of hysterectomy depending on age and risk factors for therapeutic failure
in multivariate analysis (n = 187§).
|
Risk factors
|
≤ 45 years (n = 111)
|
> 45 years (n = 76)
|
|
|
Hazard ratio
|
p-value*
|
95% confidence interval
|
Hazard ratio
|
p-value*
|
95% confidence interval
|
|
* p-value: multivariate logistical regression model, § Data missing for 8 patients
|
|
Fibroid
|
7.872
|
0.033
|
1.176 – 52.701
|
2.252
|
0.540
|
0.168 – 30.254
|
|
Parity
|
1.241
|
0.643
|
0.498 – 3.092
|
1.267
|
0.673
|
0.422 – 3.806
|
|
BMI
|
1.065
|
0.455
|
0.903 – 1.256
|
1.000
|
0.991
|
0.851 – 1.178
|
|
Hysterometer
|
1.087
|
0.824
|
0.521 – 2.266
|
0.903
|
0.841
|
0.334 – 2.446
|
In 12 patients (6.5%), a progestogen IUD was placed intraoperatively additionally
for contraception. Of these patients, 9 had amenorrhoea postoperatively and 3 had
spotting. Only 6 of these patients were younger than 45. Excluding these 12 patients
in the multivariate analysis did not demonstrate any change in the result. Because
of the small number of cases of this collective, an analysis of the success rate was
performed with a 2 : 1 paired sample according to age and intramural mass. No significant
superiority of the combined therapy IUD plus endometrial ablation versus endometrial
ablation alone was seen (95% CI 0.932 – 1.546; p = 0.157).
Discussion
In our prospective cohort study, bipolar radiofrequency ablation of the endometrium
led to amenorrhoea or cyclical spotting in 86% of the patients. The amenorrhoea rate
of 47.1% in our study was comparable with the result (43.8%) of the meta-analysis
from 2012 from Daniels et al. [26]. Newer publications show amenorrhoea rates between 45 – 56% ([Table 6]).
Table 6 Results of studies with bipolar radiofrequency ablation.
|
Study
|
n
|
Comparison
|
End points
|
Follow-up period (months)
|
Results
|
|
Abbott et al. 2003 [33]
Randomised study
|
55
|
Radiofrequency ablation vs. Thermal balloon (Cavaterm)
|
Amenorrhoea rate after 12 months
Pain 4 hours postoperatively
|
12 (postop.)
|
Amenorrhoea: 43% vs. 12% (p = 0.04)
Pain 48% vs. 78% (p = 0.01)
|
|
Bongers et al. 2004 [29]
Kleijn et al. 2008 [30]
Randomised study
|
126
|
Radiofrequency ablation vs. Thermal balloon (Thermachoice)
|
Amenorrhoea rate Patent satisfaction
|
12 (postop.)
|
Amenorrhoea: 43% vs. 8% (p < 0.001)
Satisfaction 90% vs. 79% (p = 0.003)
|
|
Amenorrhoea rate Hysterectomy rate
Quality of life
|
60 (postop.)
|
Amenorrhoea: 48% vs. 23% (p < 0.001)
Hysterectomy rate 9.9% vs. 12.9%, HR 1.2
Quality of life same (p = 0.73)
|
|
Clark et al. 2011 [17]
Randomised study
|
81
|
Radiofrequency ablation vs. Thermal balloon
|
Amenorrhoea
Duration of surgery
|
6 (postop.)
|
Amenorrhoea 39% vs. 21% (p = 0,1)
Duration of RF on average 6.2 min shorter (p < 0.001)
|
|
Penninx et al. 2016 [34]
Randomised study
|
104
|
Comparison of bipolar radiofrequency ablation vs. Thermal balloon (Thermablate)
|
Amenorrhoea rate Patient satisfaction
Repeat intervention rate
|
12 (postop.)
|
Amenorrhoea rate 56% vs. 23%, RR 0.6, 95% CI 0.4 – 0.8
Patient satisfaction 87% vs. 69%, RR 0.44, 95% CI 0.2 – 0.97
Repeat intervention rate 10% vs. 12%, RR 1.02, 95% CI 0.9 – 1.2
|
|
Penninx et al. 2011 [35]
Randomised study
|
160
|
Radiofrequency ablation vs. Hydrothermal ablation
|
Amenorrhoea rate Repeat interventions
|
60 (postop.)
|
Amenorrhoea 55,4% vs. 35,3%, RR 1,5, 95% CI 1,05 – 2,3
Repeat interventions 17% vs. 48%, RR 0.43, 95% CI 0.23 – 0.80
|
|
Muller et al. 2015 [36]
Retrospective study
|
505
|
Radiofrequency ablation (289 pat.) vs. Thermal balloon (ThermaChoice) (216 pat)
|
Amenorrhoea rate
Hysterectomy rate
|
35 (Median)
|
Amenorrhoea 45% vs. 27% (p = 0.001)
Hysterectomy rate 13% vs. 19% (p = 0.066)
|
|
Ferguson et al. 2015 [37]
Retrospective study
|
1994
|
Hysterectomy rate following radiofrequency ablation
|
Hysterectomy rate
|
48 (Median)
|
Hysterectomy in 203 pat. (10%)
Indication: bleeding 117 (58%); pain 31 (15%), bleeding and pain 45 (22%), other 10
(5%)
|
|
Wyatt et al. 2016 [38]
Retrospective study
|
144
|
Dysmenorrhoea rate before and after bipolar radiofrequency ablation
|
Dysmenorrhoea rate
|
72 (Median)
|
Pretherapeutic 69%; post-therapeutic 38% (p < 0.001)
|
|
Shazly et al. 2016 [28]
Retrospective study
|
1178
|
Predictors for failure of the radiofrequency ablation
|
Failure: Hysterectomy or repeat ablation or drug-based ovarian suppression
|
52 (Median)
|
Hysterometer > 10.5 cm; HR 2.58 (p = 0.006)
Cavity length > 6 cm; HR 2.06 (p = 0.002)
Cavity width > 4.5 cm; HR 2.06 (p = 0.002)
Cavity surface > 25 cm2; HR 2.02 (p = 0.003)
Surgical time < 93 s; HR 2.61 (p = 0.01)
|
|
Present study
Prospective study
|
187
|
Predictors for failure of the radiofrequency ablation
|
Spotting, amenorrhoea rate, hysterectomy rate
|
17,5 (Median)
|
Spotting, amenorrhoea rate:
intramural mass, age ≤ 45 years:
HR 3.699 (p = 0.036), 95% CI 1.089 – 12.570
Hysterectomy rate:
intramural mass, age ≤ 45 years:
HR 7.873 (p = 0.033), 95% CI 1.176 – 52.701
|
The multivariate analysis of the risk factors for therapeutic failure or hysterectomy
showed that intramural masses significantly worsen the therapeutic success in patients
younger than age 45 and increase the risk of a hysterectomy 8-fold. This is consistent
with the results of Soini et al. [27]. In a Finnish population-based study, they found that the presence of fibroids,
a young age, status post Caesarean section and tubal sterilisation are associated
with an increased risk of hysterectomy following endometrial ablation. However, all
endometrial ablation methods were included in this study.
It is remarkable that the size of the intramural masses measured preoperatively was
moderate, since larger fibroids, DD: adenomyomas preoperatively led to exclusion in
our study. Nonetheless, intramural masses with a median diameter of only 15 mm demonstrated
significant worsening of the success of treatment. The fact that an increase in the
cavity surface leads to worse results was shown in a recently published retrospective
study [28]. This study identified large cavity dimensions and short ablation times as risk
factors for therapeutic failure of bipolar radiofrequency ablation. In our study,
the hysterometer alone was not a significant prognostic factor, since excessively
large dimensions of the uterine cavity were possibly already preoperatively excluded
in our study by means of ultrasound.
One strength of the study is that all patients were preoperatively evaluated via transvaginal
ultrasound by one of the authors. The evaluation of the success of the treatment as
well as the indication of a subsequent hysterectomy was performed by the gynaecologist
providing subsequent care and not primarily by the study team. We also consider this
to be a strength of the study, since the conditions of daily clinical practice are
reflected in a more realistic way as a result.
The weaknesses of our study are the lack of an objective measurement of the heaviness
of the menstrual flow pre- and postoperatively. However, the subjective assessment
of the amount of flow by the patient and the attending gynaecologist appears to us
to be sufficient since it reflects daily clinical practice in a practical way. The
analysis was divided according to age groups (< 45 years, > 45 years) to take the
effects of incipient menopause into account. As expected, the influence of intramural
masses was more pronounced in younger patients. Another weakness is the relatively
brief follow-up period of a median of 17.5 months. However, 12 months appear to be
an adequate follow-up period, since hardly any difference in the rate of amenorrhoea
after 12 or 60 months postoperatively can be observed ([Table 6], [29], [30]). Moreover, no evaluation of the quality of life was performed since the patients
were in part followed up on by private-practice colleagues. This also represents a
weakness of the study.
Twelve patients underwent bipolar radiofrequency ablation as well as the placement
of a progestogen IUD for contraception. Nine patients postoperatively demonstrated
amenorrhoea and 3 demonstrated spotting. However, the 2 : 1 paired sample test did
not demonstrate any significant differences between the patients with IUD with endometrial
ablation and the patients who only underwent endometrial ablation. On the other hand,
no reliable contraception is guaranteed by the bipolar radiofrequency ablation and
based on our analysis, there is nothing to argue against its placement. In one small
study [31], the progestogen IUD placement caused a single increase in the rate of amenorrhoea
in the case of a hysteroscopic endometrial resection.
Conclusion
The results of this study show that the treatment results of bipolar radiofrequency
ablation in young patients significantly worsen when small intramural masses are present
and lead to a hysterectomy rate which is eight times higher. We therefore recommend
performing an ultrasound examination preoperatively to exclude intramural masses,
in addition to ascertaining the cavity dimensions.
Note
Published translation. Original manuscript written in German.
