Endometrial ablation for treatment of heavy menstrual bleeding: For whom it may not work?

Heavy Menstrual Bleeding (HMB) is the most prevalent form of abnormal uterine bleeding, a complaint that affects approximately 14% of reproductive aged women and presents both a health and fi nancial burden [1]. For decades, hysterectomy had been the only treatment approach when medical treatment is contraindicated or ineffective. Endometrial ablation (EA) has been introduced as an alternative to radical treatment since late 80’s and early 90’s and has gained substantial popularity as a convenient procedure of less cost and high safety profi le. Thereafter, evidence from the literature has supported EA as a second line treatment if medical treatment fails or is contraindicated [2]. Second generation EA devices (also known as global EA) have been well-established in gynecologic care and are offered to all eligible women. Because second generation devices allow offi ce-based treatment under local anaesthesia, EA has become widely acceptable by both patients and health care providers and has become the most common treatment of HMB in the United States [3].


Introduction
Heavy Menstrual Bleeding (HMB) is the most prevalent form of abnormal uterine bleeding, a complaint that affects approximately 14% of reproductive aged women and presents both a health and fi nancial burden [1]. For decades, hysterectomy had been the only treatment approach when medical treatment is contraindicated or ineffective. Endometrial ablation (EA) has been introduced as an alternative to radical treatment since late 80's and early 90's and has gained substantial popularity as a convenient procedure of less cost and high safety profi le.
Thereafter, evidence from the literature has supported EA as a second line treatment if medical treatment fails or is contraindicated [2]. Second generation EA devices (also known as global EA) have been well-established in gynecologic care and are offered to all eligible women. Because second generation devices allow offi ce-based treatment under local anaesthesia, EA has become widely acceptable by both patients and health care providers and has become the most common treatment of HMB in the United States [3].
Although EA provides fewer complications and signifi cantly costs less than hysterectomy [4], treatment failure, defi ned as recurrence of symptoms that necessitates reintervention, presents a major concern. Treatment failure is time-dependent and accordingly, the incidence of reintervention has been increasingly recognizable while 2-3 decades passed since EA was introduced [5]. Failure of EA reached up to 26% after 8 years of treatment [3]. Because failure of EA doubles treatment burden in terms of morbidity and costs and because EA failure seems to vary widely among reports, identifi cation of determinants of treatment success had become imperative to improve patient selection and counselling [6].
In this review, our objective is to present available evidence on predictors of long-term failure of EA. We also aim to explain the mechanisms by which these predictors may contribute to recurrence and to recommend strategies that could improve treatment outcomes and minimize the need for intervention.

Endometrial ablation: An overview
EA refers to a set of procedures that aim to destroy the endometrium beyond its natural capacity to regenerate.
Although the term seems to be relatively recent, trials on endometrial destruction were initiated many decades ago.
One of the oldest trials was conducted by Droegemueller et al., who treated 11 women with dysfunctional uterine bleeding using cryosurgery of the endometrium based on a Frigitronics nitrous oxide cryosurgical system [7]. One decade later, ND: Twenty one out of 22 responded to this approach and no major events were reported [8]. In 1987, hysteroscopic resection of the endometrium was introduced by DeCherney et al., using modifi ed urologic hysteroscope. Destruction of endometrial lining was achieved using coagulating current. The procedure was highly successful and no complications were reported [9].
These techniques were gradually established as treatment Mamdouh M Shaaban, Ali M El Saman*, Ahmed I Hasanein and Sherif A Shazly options and in the 90's, evidence from the literature was reviewed and these modalities were considered in clinical guidelines as an alternative to hysterectomy [10]. Nineteen ninety six and afterwards, new modalities of endometrial ablation, that do not necessitate direct visualization of the uterine cavity, have emerged and have been known as secondgeneration EA or global EA [11]. Because older methods (fi rst generation) require special training and are associated with risk of uterine perforation and fl uid overload, new EA modalities have provided obvious privilege in terms of safety and feasibility [12].

Failure of endometrial ablation: Incidence and time factor
Many studies provide information on probability of failure of EA particularly on long-term followup. However, data were predominantly inconsistent. Some of these studies are summarized in Table 1 [13][14][15][16][17][18][19][20][21]. The rate of EA failure, which denotes subsequent interventions, ranges widely between 8% and 29%. Although it has been perceived that failure is primarily infl uenced by followup length, these studies did not convey this perception while the lowest rate (8%) was reported after 84 months and the highest rate (29%) was reported after 24 months of followup [16,18]. However, this information does not reject the signifi cance of treatment-followup interval.
Among studies that included larger sample sizes (816-3,681), failure rates had tighter ranges (13.4% -21%) and were parallel increasing with the length of followup [18][19][20]. Apparently, older studies tended to present higher failure rates than more recent evidence. This may suggest a higher probability of failure among women treated with fi rst generation EA particularly endometrial resection; failure rates in studies including endometrial resection consistently exceeded 20% [13,15,16,18]. Furthermore, microwave EA yielded a signifi cantly lower failure rate compared to studies conducted at that time [14]. Second generation EA devices were generally associated with less than 20% failure rates. However, there is no evidence that second generation EA provides more favorable outcomes than fi rst generation devices [22]. Accordingly, this discrepancy may refl ect a change in attitude of patient selection that copes with emerging evidence that particular characteristics may increase the chance of failure. Many gynecologists believe that an older age group may respond better to EA that younger women even before this has become evident in the literature. Therefore, it is diffi cult to assume a single failure rate while many confounders could play a role. Identifi cation of predictors of treatment failure is prioritized to achieve the lowest of this failure range which seems to be principally acceptable.

Endometrial ablation: Baseline predictors of failure
Identifi cation of baseline characteristics that would increase the probability of treatment failure poses an appropriate strategy to reduce the rate of surgical reintervention and improve patient counselling. However, baseline predictors of failure seem to widely vary among studies. Some of these predictors are summarized in Table 2 [5,[18][19][20][23][24][25][26]. A younger age at EA has been the most consistent predictor of failure among most of these studies [18][19][20][23][24][25][26]. The impact of age may be ascribed to procedure-menopause interval; failure of EA is time dependent and accordingly, older women may get to menopause within few years and avert failure.
While approaching menopause, bleeding pattern also tends to be irregular rather than heavy. A younger age also predicted 2), respectively [18]. Accordingly, CIs are overlapping and the differences do not seem to convey a strong clinical predilection. Age less than 35 years was associated with future bleeding but hysterectomy rate was not higher than other age groups [26]. Accordingly, establishing patient selection of women between 40 and 45 years may be infl uenced by other clinical circumstances.
Tubal ligation has been also recognized as a potential predictor of treatment failure [19,26]. Although it has not been defi ned as a predictor in all studies, tubal ligation was recognized as a risk factor for postablation cornual and tubal hematometra (postablation syndrome) which necessitates intervention [28][29][30]. The syndrome was not restricted to a particular ablation device [19,29,30]. It is, therefore, justifi able that women with prior tubal ligation be counselled for the risk of reintervention particularly for postablation pain. Women were also at risk of signifi cant postoperative pain if they were less than 40 years old or if they experienced preablation dysmenorrhea [27]. Dysmenorrhea is another risk factor that has been considerably recognized both in practice and in the literature. Preablation dysmenorrhea was defi ned as a predictor of EA failure and is commonly addressed during counselling in current practice [19,25,27]. The presence of dysmenorrhea may signify underlying endometriosis or adenomyosis which cannot be defi nitely excluded prior to EA. Therefore, EA will not largely relieve patient complaint.
The role of tobacco use was addressed in 2 previous studies specially as predictors of postablation pain [23,27]. However, this association was not explained given the anticipated relation between smoking and hypoestrogenism [31]. Although the relation between smoking and pain perception is debatable, there is evidence that smoking may increase pain intensity in women with chronic pain owing to central mechanisms [32]. Nevertheless, evidence on smoking contribution in EA failure or its mechanism is lacking and interpretation of this information in practice is still limited. Other predictors of failure were described in individual studies.

Endometrial ablation: Perioperative predictors of failure
While most studies primarily assessed how baseline characteristic would impact treatment success, few studies address the role of perioperative measurements including premedications, preprocedure sonographic fi ndings, uterine measurements, and procedure parameters. However, it is generally accepted that a particular uterine length would not allow EA; one of the earliest guidelines for EA, published in 1995, stated that EA should be performed for uterine length <12cm [10]. Although this cutoff point was considered in many studies, A greater success was recognized when a lower cutoff is used [33,34]. A cutoff point of 10cm is now widely adopted [35]. This explains why uterine depth was not frequently evaluated as a predictor of failure in most studies. According to El-Nashar et al., a uterine length ≥9cm was not connected to failure. However, women with uterine lengths <9cm were more prone to amenorrhea [19]. Shaamash and Sayed found that uterine depth would predict EA failure in women treated with thermal balloon EA. The range of uterine depth was 7.5-12cm but no particular cutoff was defi ned [24]. A recent study found that uterine length >10.5cm was associated with REA failure particularly those caused by post-ablation bleeding [5]. On the other side, there is no much information in the literature on the role of uterine width. Uterine length and width could be measured sonographically prior to the procedure and automatically at the time of REA procedure. Shazly et al., found that failure of REA is higher if uterine width exceeded 4.5cm [5]. This is particularly predictive of postablation pain and may indicate an undiagnosed uterine adenomyosis or a uterine width that extends beyond ablation zone leaving active peripheral endometrial tissue behind a central obliterated cavity. A radiofreqency ablation (RFA) index was thus established as: REA procedure duration divided by uterine surface area (uterine length multiplied by uterine width). This equation seemed to be more predictive than uterine length, width, or duration alone [5]. Other perioperative predictors include uterine retroversion and endometrial thickness which were found signifi cant by one study on thermal balloon EA and were not signifi cant by another larger study that included a mixed a cohort of thermal balloon EA and REA [19,25]. Premedications were not given in the latter study.

Endometrial ablation failure: Conclusion
Although the probability of further intervention following EA is considerable, EA remains an appealing alternative to hysterectomy. It is anticipated that proper patient selection would substantially lower the rate of reintervention.
Unfortunately, potential predictors have not been metaanalyzed to data owing to the wide heterogeneity of available studies. However, it seems that age is the most acceptable predictor to date. Using either 40 or 45 years as a cutoff is not defi nite and it should be made according to other clinical circumstances. Tubal ligation and dysmenorrhea should also be considered. Perioperatively, women should be advised that EA success would be lower if uterine length is greater than 10 to 10.5cm. RFA index is a novel index that considers uterine surface area rather than uterine length alone. However, further studies are needed to validate its use. There is currently no evidence that uterine position, endometrial thickness, and premedication change long-term outcomes. Although adoption of a selective strategy is justifi able, further research on the impact of its implementation on long-term outcome is recommended.