Background
Endometriosis is a chronic estrogen dependent gynaecological condition characterized by the presence of ectopic glands and stroma outside the uterine cavity. It affects at least 3.6% of women [
1] and it often causes infertility and/or pain symptoms (dysmenorrhea, deep dyspareunia, chronic pelvic pain and dyschezia). In some patients, pain symptoms are extremely severe and negatively affect quality of life, work efficiency and sexual life [
2‐
4]. Several hormonal therapies have been proposed for the treatment of endometriosis related pain, including oral contraceptive pill and other estroprogestin formulations (such as the vaginal ring and the transdermal patch), progestins (such as medroxyprogesterone acetate, norethisterone acetate and the levonorgestrel-releasing intrauterine device), gonadotrophin releasing hormone analogues and danazol [
5]. These traditional endocrine therapies for endometriosis inhibit estrogens production in the ovary. However, in some patients, pain symptoms may persist despite the use of endocrine therapies.
Since the late 1990s, several independent studies based either on polymerase chain reaction or immunohistochemistry have demonstrated that aromatase P450 is over-expressed in both eutopic and ectopic endometrium of patients with endometriosis, while this enzyme is not detectable in eutopic endometrium obtained from healthy women and in endometriosis free peritoneal tissue [
6‐
13]. Although the aromatase P450 can been detected in both epithelial and stromal cells obtained from ectopic and eutopic endometrium of women with endometriosis, its expression is higher in epithelial than in stromal cells. In contrast with majority of the literature, some recent studies questioned the aberrant expression of the aromatase P450 in endometriosis [
14‐
16]. Delvoux
et al. reported the absence of aromatase P450 in eutopic and ectopic tissue obtained from women with endometriosis [
14]. In addition, it was demonstrated that endometriotic lesions could create a hyperestrogenic environment increasing the reduction of estrone into 17-β estradiol and decreasing the oxidation of 17-β estradiol into estrone [
14]. In agreement with these observations, Colette
et al. found that aromatase P450 is undetectable by immunohistochemistry in the glandular and stromal compartments of ectopic endometrial tissue [
16]. Furthermore, the authors showed that the expression of the aromatase gene, measured by quantitative polymerase chain reaction using three different protocols, is low in endometriomas and barely detectable in only a small percentage of eutopic endometrial samples, peritoneal lesions and rectovaginal nodules [
16]. The authors suggested that what was believed to be aromatase protein was mainly endogenous biotic labeling or iron deposits [
15].
Based on the molecular observations of increased expression of aromatase P450 in endometriotic tissue published over the last ten years, some authors used aromatase inhibitors (AIs) to treat pain symptoms in premenopausal women with endometriosis [
17,
18]. This systematic review was performed to assess the efficacy of AIs in treating pain symptoms caused by endometriosis.
Methods
This systematic review was carried out according to the MOOSE guidelines [
19]. No institutional review board approval was required because only published, de-identified data were analysed.
Identification of the literature
A systematic literature search was performed to identify all the published studies evaluating the efficacy of AIs in treating pain symptoms associated with endometriosis. The search included the combination of the following medical subjects heading terms: "endometriosis", "deep endometriosis", "deeply infiltrating endometriosis", "rectovaginal endometriosis", "ovarian endometriosis", "endometriomas", "bowel endometriosis", "bladder endometriosis", "dysmenorrhea", "pelvic pain", "deep dyspareunia", "medical therapy/treatment", "aromatase inhibitors", "letrozole" and "anastrozole".
The following electronic databases were searched MEDLINE, EMBASE, PubMed, SCOPUS and the Cochrane System Reviews from inception until October 2010.
All pertinent articles were examined and their reference lists were systematically reviewed in order to identify other studies for potential inclusion in this systematic review. Review articles, books and monographs published on endometriosis were consulted and their reference lists were systematically searched to identify further studies that could be reviewed.
No language restriction was applied. The search was run every month between January 2009 and October 2010 to identify new articles.
Study selection
Randomised controlled trials (RCTs), patient preference trials and observational studies were included in this review, whereas case reports, abstracts and proceedings of scientific meetings were excluded.
The studies included in the current review were selected accordingly to the following criteria:
Population. Premenopausal women with diagnosis of endometriosis (either as a primary or recurrent disease) based at least on vaginal and/or rectal examination; ideally on previous surgery and histological examination of the lesions. Patients complaining of pain symptoms such as dysmenorrhea, deep dyspareunia, chronic pelvic pain and dyschezia.
Interventions. Treatment of pain symptoms with third generation nonsteroidal (type II) AIs either alone or in combination with other hormonal therapies.
Design. The minimum number of individuals in each trial was 10. The minimum size of follow-up was 80% or more, though there was no minimum length of follow-up. Assessment of the intensity of pain symptoms by using standardised scales such as the visual analogue scale (VAS).
Outcome. Changes in the intensity of endometriosis-related pain symptoms during treatment with AIs either alone or combined with other hormonal therapies but not combined with surgery (primary outcome). Efficacy of AIs either alone or combined with other hormonal therapies in preventing the recurrence of pain after surgical treatment of endometriosis (secondary outcome).
The abstract of studies retrieved in the search were reviewed by two authors (S.F. and D.J.G.) to exclude citations deemed irrelevant. The reviewers worked independently and in duplicate. Any discrepancy between the two reviewers was resolved by consensus or arbitration by a third reviewer (V.R.). The reviewers were not blinded to the names of investigators or sources of publication.
Data were extracted from each article and collected in standardised forms in duplicate by two reviewers (S.F. and D.J.G.). A final abstraction form was compiled from the two evaluation forms, with correction and resolution of any discrepancy between reviewers by consensus reached after discussion or arbitration by a fourth reviewer (P.L.V.).
Discussion
All the observational studies included in this review [
38,
40‐
43], one patient-preference trial [
44] and one RCT [
47] demonstrated that AIs combined with either progestogens, oral contraceptive pill or gonadotropin releasing hormone analogue reduce the intensity of pain symptoms caused by endometriosis (Table
2). In addition, two observational studies showed that the administration of AIs improves quality of life [
41,
43]. The effect of AIs on the volume of rectovaginal endometriotic nodules remains unclear. In one observational non-comparative study [
41], the administration of vaginal anastrozole (0.25 mg/day) for 6 months did not cause a decrease in the volume of rectovaginal endometriotic nodules. In contrast, a RCT [
47] showed that the oral administration of letrozole (2.5 mg/day) combined with either gonadotropin releasing hormone analogue or progestogen significantly reduces the size of rectovaginal endometriotic nodules.
Endometriosis is a chronic disease [
48] and pain typically recurs to a degree similar to that at baseline after discontinuation of endocrine therapies [
49,
50]. Most of the published observational non-comparative studies did not report a follow-up after discontinuation of treatment with AIs [
38,
40,
41,
47]. Two observational non-comparative studies [
42,
43] and one patients-preference trial [
44] observed a quick recurrence of pain symptoms after discontinuation of therapy with AIs. In particular, after 6 months from the completion of the treatment, the intensity of pain symptoms was similar to pre-treatment values [
42‐
44]. Furthermore, patients with persistent pain symptoms during the administration of AIs have been found to have rectovaginal nodules with preserved glandular epithelium and abundant stromal cells having proliferative activity [
43].
The adverse effects caused by AIs in premenopausal women (such as arthralgia and myalgia) may be more severe than those caused by progestogens and oral contraceptive pill thus reducing the compliance of the patients in the long-term treatment. Recently, a patient preference study demonstrated that letrozole and norethisterone acetate are more effective in reducing pain and deep dyspareunia than norethisterone acetate; although the AI resulted in a higher incidence of adverse effects [
44]. As a consequence patients' satisfaction was not improved by the administration of letrozole combined with norethisterone acetate when compared with norethisterone acetate alone [
44]. In this perspective, it is preferable to combine AIs with progestogens rather than with gonadotropin releasing hormone analogues. In fact a RCT showed that combining the AI with norethisterone acetate causes a lower incidence of adverse effects and lower discontinuation rate than combining letrozole with triptorelin [
47].
AIs may have a role in the prevention of recurrence after surgical treatment of endometriosis. Two RCTs [
39,
45] demonstrated that the administration of AIs (alone or combined with gonadotropin releasing hormone analogue) for 6 months reduces the risk of endometriosis recurrence when compared with postoperative treatment with gonadotropin releasing hormone analogue or danazol. In contrast, one RCT demonstrated that the postoperative administration of AIs for only 2 months does not decrease the risk of endometriosis recurrence when compared with gonadotropin releasing hormone analogue or placebo [
46].
The heterogeneity of the analyzed studies is an important limitation of the current systematic review. In fact, AIs were administered either alone or in combination with various hormonal therapies (progestin, oral contraceptive pill or gonadotropin releasing hormone analogues). In addition, the severity of endometriosis was established accordingly to the classification of the American Society for Reproductive Medicine (ASRM) only in four studies [
51]; furthermore, two of these studies included women with various stage of disease. Four studies included patients with rectovaginal endometriosis, but the severity of the disease was not established accordingly to the ASRM classification. A further limitation of this review consists in the fact that AIs were administered only for a relatively short period of time (from two to six months). Despite these important limitations, the findings of the analysed studies suggest the efficacy of AIs in the short-term treatment of endometriosis-related pain symptoms. However, women with endometriosis should receive a chronic treatment [
5,
48] and the optimal therapy should balance potential benefits and side affects. In this perspective, the role of AIs in the long-term treatment of pain symptoms caused by endometriosis remains uncertain because of the persistence of endometriotic lesions during treatment [
43] and of the high incidence of the side effects observed not only in women with endometriosis [
44,
47] but also in premenopausal breast cancer patients [
52]. In addition, there are several concerns on the safety of a long-term administration of AIs to premenopausal women. Third-generation AIs negatively affect bone health [
53,
54]. They suppress aromatase activity within the osteoblasts presumably favouring increased osteoclastic activity and net loss of bone mineral density [
53]. One RCT included in this review [
47] and studies performed in premenopausal breast cancer patients [
55] showed that a 6 month treatment with AIs and gonadotropin releasing hormone analogue significantly decrease bone mineral density. The hypoestrogenism caused by AIs may be associated with other adverse effects. Studies performed in breast cancer patients showed that AIs may lack the lipid lowering and cardioprotective effect of tamoxifen, but they do not seem to increase the risk of cardiovascular events [
56]. However, data on the effects of AIs on cardiovascular disease is limited, particularly in premenopausal women.
Given this background, it seems unlikely that AIs, particularly combined with gonadotropin releasing hormone analogues, may be used as a standard long-term treatment of premenopausal endometriosis. Combining AIs with other hormonal therapies (such as norethisterone acetate) may decrease the incidence of adverse affects and reduce the loss of bone mineral density. However, only women failing to improve after the administration of standard endocrine therapies and surgical excision of endometriosis may be candidate to receive AIs. These agents should be prescribed only to women with histologically proven endometriosis and normal mineral bone density. In addition, before starting the treatment, patients should be informed in details of the potential adverse effects caused by AIs, particularly musculoskeletal symptoms.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
All the authors contributed to the conception of the review. SF and DJG performed literature search, selected the abstracts and abstracted the data. VR resolved the dicrepancies between the two reviewers (SF and DJG) in the selection of the study of interest. PLV resolved the dicrepancies between the two reviewers (SF and DJG) in the abstraction of the data from the study of interest. SF prepared the first draft of the manuscript and performed subsequent amendments. PLV and VR reviewed the manuscript. All authors read and approved the final manuscript.