Pharmacological Treatment for Symptomatic Adenomyosis: A Systematic Review

► adenomyosis ► abnormal menstrual bleeding ► pelvic pain ► systematic review ► medical treatment

Objective To assess the ef ficacy of non-surgical treatment for adenomyosis. Data Sources A search was performed by two authors in the Pubmed, Scopus, and Scielo databases and in the grey literature from inception to March 2018, with no language restriction. Selection of Studies We have included prospective randomized studies for treating symptomatic women with adenomyosis (abnormal uterine bleeding and/or pelvic pain) diagnosed by ultrasound or magnetic resonance imaging. Data Collection Studies were primarily selected by title and abstract. The articles that were eligible for inclusion were evaluated in their entirety, and their data was extracted for further processing and analysis. Data Synthesis From 567 retrieved records only 5 remained for analysis. The intervention groups were: levonorgestrel intrauterine system (LNG-IUS)(n ¼ 2), dienogest (n ¼ 2), and letrozole (n ¼ 1). Levonorgestrel intrauterine system was effective to control bleeding when compared to hysterectomy or combined oral contraceptives (COCs). One study assessed chronic pelvic pain and reported that LNG-IUS was superior to COC to reduce symptoms. Regarding dienogest, it was efficient to reduce pelvic pain when compared to placebo or goserelin, but less effective to control bleeding than gonadotropin-releasing hormone (GnRH) analog. Letrozole was as efficient as GnRH analog to relieve dysmenorrhea and dyspareunia, but not for chronic pelvic pain. Reduction of uterine volume was seen with aromatase inhibitors, GnRH analog, and LGN-IUD.

Levonorgestrel intrauterine system and dienogest have signi ficantly improved the control of bleeding and pelvic pain, respectively, in women with adenomyosis. However, there is insuf ficient data from the retrieved studies to endorse each medication for this disease. Further randomized control tests (RCTs) are needed to address pharmacological treatment of adenomyosis. Cristina Laguna Benetti-Pinto ’s ORCID is https://orcid.org/0000- 0001-6198-5593. received March 12, 2019 accepted June 24, 2019 DOI https://doi.org/ 10.1055/s-0039-1695737. ISSN 0100-7203. Copyright © 2019 by Thieme Revinter Publicações Ltda, Rio de Janeiro, Brazil Review Article THIEME 564 Published online: 2019-09-23

Adenomyosis is a benign disorder in which basal endometrial glands and stroma are found in the myometrium with reactive hyperplasia of the surrounding smooth muscle myometrial cells.1–5 It is a complex, gynecological condition with unknown incidence and etiology. Clinical symptoms are related to pain and bleeding, and they include dysmenorrhea, abnormal uterine bleeding, chronic pelvic pain (CPP), dyspareunia, and infertility; however, a third of women can be asymptomatic. 6 Symptoms typically are reported to develop between the ages of 40 and 50 years; however, this may reflect the fact that the usual moment for diagnosing adenomyosis has been after performing a hysterectomy because of preoperative difficulty to establish the diagnosis. With improvement of diagnostic methods, like magnetic resonance imaging (MRI) and high- quality transvaginal ultrasound (TVUS), early diagnosis can be made with an accuracy of 80 to 90%. 7–11 Non-surgical treatment can be necessary or desirable for women who want to maintain the uterus for a future pregnan- cy, those with other comorbidities that pose a higher risk for surgery, or even those who are close to menopause and would not like to undergo a surgical procedure. There are systematic reviews about uterine artery embolization, fertility-sparing treatment in patients with infertility, and local excision of adenomyosis. 12–14 A published review about medical treat- ment for adenomyosis has presented their data narratively, since the author states that the aim was to discuss the medical approach to the management of adenomyosis symptoms, with no analysis of the risk of bias and methodological quality. 15 Given the need for systematic reviews and quality assessment for analyzing these data, we sought to perform a systematic review of the effectiveness of non-surgical treatment for adenomyosis on uterine volume, pelvic pain, and menstrual bleeding, when compared with other surgical and non-surgical interventions.

Study Design, Data Search, Inclusion/Exclusion Criteria The present review was recorded in the International Pro- spective Register of Systematic Reviews (PROSPERO)16 under the number CRD42017057896 and was developed according to Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA). 17 Searches in the databases included articles from the following sources: Pubmed, Scopus, Scielo, and the grey Resumo Objetivo: Avaliar a e ficácia de tratamento não cirúrgico para adenomiose. Fontes de dados: U m ap e s q u i s af o ir e a l i z a d ap o rd o i sa u t o r e sn a sb a s e sd ed a d o s Pubmed, Scopus, Scielo e na literatura cinzenta desde o início de cada base de dados até março de 2018, sem restrição de idioma. Seleção de estudos: Incluímos estudos prospectivos randomizados para tratamento de mulheres sintomáticas com adenomiose (sangramento uterino anormal e/ou dor pélvica) diagnosticadas por ultrassonogra fia ou ressonância magnética. Coleta de dados: Os estudos foram selecionados principalmente por título e resumo. Os artigos que preencheram os critérios de inclusão foram avaliados na íntegra, e seus dados foram extraídos para posterior processamento e análise. Síntese dos dados: De 567 registros recuperados, somente 5 permaneceram para análise. Os grupos de intervenção foram: sistema intrauterino de levonorgestrel (SIU-LNG) (n ¼ 2), dienogest (n ¼ 2), e letrozol (n ¼ 1). O SIU-LNG foi efetivo no controle do sangramento quando comparado à histerectomia ou aos contraceptivos orais combinados (COCs). Um estudo avaliou a dor pélvica crônica e relatou que o SIU-LNG foi superior ao COC para reduzir os sintomas. Em relação ao dienogest, este foi eficiente em reduzir a dor pélvica quando comparado ao placebo ou à goserelina, mas foi menos e ficaz no controle do sangramento do que o análogo do hormônio liberador de gonadotropina (GnRH). O letrozol foi tão e ficiente quanto o análogo do GnRH para aliviar a dismenorreia e a dispareunia, mas não para a dor pélvica crônica. Redução do volume uterino foi observada com inibidores de aromatase, análogo de GnRH, e SIU-LNG. Conclusão: O SIU-LNG e dienogest apresentaram bons resultados para o controle de sangramento e dor pélvica, respectivamente, em mulheres com adenomiose. No entanto, não há dados su ficientes para endossar cada medicação para tratar essa doença. Futuros estudos randomizados s ão necessários para avaliar o tratamento farmacológico da adenomiose. Palavras-chave ► adenomiose ► sangramento menstrual anormal ► dor pélvica ► revisão sistemática ► tratamento médico Rev Bras Ginecol Obstet Vol. 41 No. 9/2019 Pharmacological Treatment for Symptomatic Adenomyosis Benetti-Pinto et al. 565 literature. Two authors (L. G. O. B. and T. A. A. M.) performed a distinct search using the following strategic combination of

( “medical treatment ” OR “clinical treatment ” OR “hormone treatment ” OR goserelin OR leuprolide OR GnRH OR “GnRH analog ” OR “GnRH antagonist ” OR progesterone OR Dienogest OR desogestrel OR COC OR “oral contracept ” OR “non-surgical treatment ” OR levonorgestrel OR drug OR medroxyprogesterone OR mifepristone OR sprm OR ulipristal OR progestin OR “combined oral contraceptive ” OR aromatase OR letrozole OR anastrozole (adenomyosis)) NOT (animals OR children). All articles inserted in these databases were in- cluded up to March 2018. We have included in this review prospective randomized and non-randomized studies with symptomatic women with image diagnosis (ultrasound [US] or MRI) suggestive of adeno- myosis who were submitted to medical treatment versus any other comparator group. The main symptoms were: abnormal uterine bleeding, pain and/or changes in uterine volume. There were no restrictions regarding the language. We have excluded studies with no control group, with cross-sectional or case-control designs, case series, or retrospective studies. Primary and Secondary Outcomes Our primary outcomes were: menstrual bleeding through any kind of measurements, like hemoglobin (Hb) by labora- tory test or number of pads/days by the number of protector changes per day according to women ’s report; pelvic pain through the visual analogue scale (VAS); and reduction of uterine volume measured in milliliters or cubic centimeters by TVUS or MRI. Quality of life was the secondary outcome, which was measured by questionnaires such as the World Health Orga- nization Quality of Life (WHOQOL) short version or the Medical Outcomes Study (MOS). Studies were primarily selected by title and abstract by the same authors that conducted the searches. Thus, articles that presented the eligibility criteria were evaluated in their entirety, and their data was extracted for further processing and analysis. Possible disagreements were discussed with a third author (C. L. B. P.) to obtain a consensus. The reviewers sought data that were not possible to obtain after reading the manuscript after an e-mail sent to the authors. Statistical analysis We tabulated mean difference (MD) and their standard deviations (SDs) between pre and posttreatments and their confidence intervals from continuous variables. In order to build forest plots, a mathematical calculation of error prop- agation was used, according to the author, 18 to identify the SDs not described in some of the studies included in the present review after unanswered contact with the authors, once the publication did not mention these data. It was not possible to conduct a meta-analysis by the differences of the studies heterogeneity of the studies regarding the proposed treatments, because each study assessed a different treat- ment with or without a different comparator, and a single paired comparison was not present in more than one study. As the number of studies was scant, an indirect meta- analysis was not considered either. Funnel plots (publication bias) were not elaborated due to the scant number of retrieved studies. The risk of bias in the studies was assessed using the Cochrane bias risk assessment tool, 19 which clas- sifies studies at risk of low, high, or unclear bias. The Grading of Recommendations Assessment, Development and Evalua- tion (GRADE) criteria 20 were used to build a summary of findings (SOF) table to evaluate the quality of the evidence.

Study Selection and Characteristics ►Figure 1 describes the flowchart regarding the studies that comprise the present review. From 567 records that were retrieved in this search, 5 were removed due to duplication, 562 were screened, and 11 were fully assessed for eligibility, but only 5 remained in the final model. ►Table 1 displays all selected studies that comprised 288 women; a total of 267 women completed the treatment and were included in the final results. No studies have mentioned whether their

were interpreted by intention-to-treat or per proto- col analysis. In summary, three from thefive studies were held in Egypt, four presented a randomized controlled design,21–25 and one was prospective, non-randomized.24 One study was placebo- controlled,25 and the others were pharmacological treatments versus surgery or other drugs for adenomyosis. Transvaginal ultrasound diagnosis was present in all studies, and MRI was added as an option toTVUS in two studies.21–24 The duration of treatment varied from 4 weeks to 12 months. The intervention x comparator groups were: levonorgestrel intrauterine system (LNG-IUS) versus hysterectomy21; LNG-IUS versus combined oral contraceptive (COC) 23; letrozole versus goserelin 22; dienogest versus triptorelin 24; dienogest versus placebo. 25 Some side effects of using pharmacological treatment were mentioned in all studies except one. 24 ►Figure 2 condenses all forest plots from the analyzed outcomes. From our planned primary outcomes, almost all of them (4 of 5) were present. Treatment with LNG-IUS Two studies have assessed the use of LNG-IUS 21,23 versus hysterectomy or combined oral contraceptive (COC), respec - tively. In the first study, the LNG-IUS was effective to control bleeding, with an improvement of hemoglobin levels, and reduction in the number of days with bleeding. In the second study, a reduction in the number of days with bleeding was observed. Compared to the other medical treatment using COC, the bleeding pattern was improved in both arms; in the case of LNG-IUS, the mean number of bleeding days per month decreased from 9.81 /C6 1.82 days before recruitment to 2.63 /C6 2.13 days after the 6 th m o n t ho fi n s e r t i o n(p < 0.001). In the COCs group, the number of bleeding days per month reduced from 9.97 /C6 1.52 days to 5.52 /C6 1.00 days (p < 0.001). Pelvic pain was assessed in one study, 23 and the LNG-IUS was more efficient in the improvement of chronic pelvic pain than COC (6.23 /C6 0.67–1.68 /C6 1.25 - p < 0 . 0 0 1 ) ,a sw e l la st h e reduction in uterine volume (10.23/C6 1.06 mL–7.63 /C6 0.49 mL, p < 0.001). Rev Bras Ginecol Obstet Vol. 41 No. 9/2019 Pharmacological Treatment for Symptomatic Adenomyosis Benetti-Pinto et al.566 Only one study assessed improvement of quality of life, 21 with superior effects on psychological and social life with LNG-IUS when compared to hysterectomy. Women that used LNG-IUS presented adverse effects: headache (11.9%), breast tenderness (7.1%), acne (4.8%), and transient depressive episode (2.4%). Treatment with an Aromatase Inhibitor (Letrozole) Only one study 22 evaluated letrozole in the treatment of adenomyosis compared to the GnRH analog goserelin. Letro- zole was as ef ficient as goserelin to relieve dysmenorrhea (p ¼ 0.48) and dyspareunia ( p ¼ 0.70), but the CPP control was statistically higher with goserelin ( p ¼ 0.04). Regarding the control of bleeding and the reduction of uterine volume, both medications presented similar response, but more side effects (hot flushes) were reported with goserelin (81.3%). Treatment with Gonadotropin-releasing Hormone Agonist (GnRH Analog: Goserelin or Triptorelin Acetate) Two studies evaluated the use of GnRH analog: one compared to an aromatase inhibitor (goserelin X letrozole), 22 and the other compared to dienogest (triptorelin X dienogest). 24 The GnRH analog was more ef ficient than the aromatase inhibitor in controlling CPP ( p ¼ 0.04), but they were equally efficient in the control of dysmenorrhea and dyspareunia. When compared to dienogest, the GnRH analog was more efficient in controlling dysmenorrhea at 16 weeks (30.6/C6 18.4 versus 0.0, p < 0.0001) but equally ef ficient at reducing dys- pareunia and CPP . Regarding bleeding control, the GnRH analog did not present a statistical difference when compared to letrozole; conversely, it was superior to dienogest. Finally, regarding the reduction of uterine volume, the GnRH analog and aromatase inhibitor were equivalents; 18 however, when compared with dienogest, the GnRH analog was more ef fi- cient.21 One study evaluated side effects, having reported hot flushes in 81.3% of women treated with GnRH analog. 22 Treatment with dienogest Two studies evaluated dienogest; one compared to GnRH analog (triptorelin) 24 and the other compared to placebo. 25 Dienogest was ef ficient in both studies to reduce pain complaints (dysmenorrhea, dyspareunia, and CPP). When dienogest was compared to the GnRH analog (triptorelin), both were similar to control dyspareunia (20.7 /C6 16.5 versus 25.8 /C6 19.1, p ¼ 0.3899) and CPP (21.7 /C6 11.6 versus 24.5 /C6 13.8, p-value ¼ 0.5076). There was a signi ficant difference in the posttreatment dysmenorrhea between dienogest and triptorelin at 16 weeks, when the GnRH analog presented a better result (30.6 /C6 18.4 versus 0.0, p < 0.0001). Fig. 1 Flow diagram describing the search and study inclusion processes. Rev Bras Ginecol Obstet Vol. 41 No. 9/2019 Pharmacological Treatment for Symptomatic Adenomyosis Benetti-Pinto et al. 567 Table 1 Characteristics of the included prospective studies Author, Year Country Study aim Diagnosis Sample Study design Treatment Follow-up Evaluation

Pain -

Bleeding loss -

Uterine volume -

Quality of life Other

Side effects Ozdegirmenci et al (2011) 21 Turkey To compare LNG-IUS versus hys- terectomy US and MRI criteria N ¼ 86; 43 in each group Prospective randomized clinical trial (do not register) LNG- IUS (final partici- pants, n ¼ 43); Hysterecto- my ( final participants n ¼ 32) 6a n d 12 months 1. Menstrual bleeding (number of pads/day and hemo- globin levels) 2. Quality of life (WHOQOL- brief) Did not assess LNG-IUS increased the Hb levels to compara- ble levels with hyster- ectomy Did not assess Both treatments have improved health- related quality of life, but LNG-IUS had as u p e r i o r effect on psychologi- cal and social life LNG-IUS: headache (11.9%), breast tenderness (7.1%), acne (4.8%), a transient depressive episode (2.4%). Hysterecto- mies: 1 (3.1%) post- operative infection Badawy et al (2012) 22 Egypt To compare aromatase inhibitor (Letrozole 2.5 mg/d) versus GnRH analog (goserelin 3.6 mg) US criteria N ¼ 32 Prospective randomized, non-blind controlled clinical trial (retrospec- tively registered) Letrozole 2.5mg/d (final partici- pants n ¼ 15) Goserelin 3.6mg/ month ( final participants n ¼ 16) 4, 8, and 12 weeks 1. Pain Pelvic pain, dys- menor- rhea, dys- pareunia (VAS) 2. Menstrual bleeding and sub- fertility (question- naire, but it is not clear) 3. Uterine volume (TVUS) Goserelin was more effective in relieving chronic pelvic pain (p ¼ 0.04) Goserelin was so efficient than letrozole in relieving menorrhagia and metror- rhagia Reduction in uterine volumes, without difference between the two groups Hot flashes (0% with letrozole, 81,3% with goserelin), Pregnancy (n ¼ 2w i t h letrozole) Shaaban et al (2015) 23 Egypt To compare LNG-IUS versus COC US criteria N ¼ 62; 31 in each group Prospective randomized clinical trial, registered LNG-IUS (final partici- pants n ¼ 29) COC (35mcg of gesto- dene þ 30 mcg of EE, 21/7) (final partici- pants n ¼ 28) 6 months 1. Pain (VAS) 2. Menstrual bleeding (menstrual diary, pads/day), 3. Uterine volume (TVUS and Doppler) Both treat- ments reduced pains, however, the reduction was greater in the LNG-IUS group Both treat- ments decreased the number of bleeding days, number of sanitary pads per day, but the reduction was greater with LNG-IUS Both treat- ments de- creased the uterine volume, but the reduc- tion was greater with LNG-IUS Uterine arteries and intramyome- trial Doppler indices in- crease in both groups, but the increases were greater with LNG-IUS LNG-IUS: expulsion (n ¼ 1) Rev Bras Ginecol Obstet Vol. 41 No. 9/2019 Pharmacological Treatment for Symptomatic Adenomyosis Benetti-Pinto et al.568 Table 1 (Continued ) Author, Year Country Study aim Diagnosis Sample Study design Treatment Follow-up Evaluation

Pain -

Bleeding loss -

Uterine volume -

Quality of life Other

Side effects Fawzy and Mesbah (2015) 24 Egypt To compare dienogest versus triptorelin acetate US criteria N ¼ 41; dienogest (n ¼ 22); triptorelin acetate injection (n ¼ 19) Prospective non- ran- domized clinical trial (did not register) Dienogest 2mg/day (final partici- pants n ¼ 19) Triptorelin acetate ( final participants n ¼ 18) 16 weeks 1. Pain, dysme- norrhea, dyspareu- nia and chronic pelvic pain (VAS) 2. Menstrual bleeding (3 levels of satisfied, blood cell count, ferritin) 3. Uterine volume (TVUS) Both treat- ments reduced chronic pelvic pain and dyspar- eunia. Triptorelin was more effective in relieving dysmenor- rhea Triptorelin was more effective than dienogest (100 and 73.7% respectively) in control- ling men- strual bleed- ing. Both treatments improved Hb levels and ferritin Triptorelin was more effective in the reduc- tion of uterine volume Did not assess Osuga et al (2017) 25 Japan To compare dienogest versus placebo US and MRI criteria N ¼ 67 Randomized, double-blind, multicenter, placebo-con- trolled phase III study Dienogest 2mg/day ( fi- nal partici- pants n ¼ 34 Placebo ( fi- nal partici- pants n ¼ 33) 16 weeks 1. Pain Pelvic pain (pain severity score in order to access work and analgesics use and VAS) 2. Menstrual bleeding Patient diary form, classified by the number of days and severity of bleeding. 3. Uterine volume (US and MRI) 4. Quality of life (MOS 36-item short- form health survey) Dienogest reduced pel- vic pain and other pain parameters Dienogest presented a lower num- ber of days with bleed- ing, and the most was spotting or break- through, but the numbers were not statistically compared between the groups Uterine volu- mereduction in both groups, but no differ- ence be- tween them Bodily pain reduction (the item of quality of life) Reduction of analgesics u s es c o r ei n the Dieno- gest group Anemia and menstrual bleeding (placebo) and hot flash (dienogest) Abbreviations: COC, combined oral contraceptive; GnRH, gonadotropin-releasing hormone; LVG-IUS, levonorgestrel intrauterine system; MOS, Med ical Outcomes Study; MRI, magnetic resonance imaging; TVUS, transvaginal ultrasound; US, ultrasound; VAS, visual analogue scale; WHOQOL, World Health Organization Quality of Life. Rev Bras Ginecol Obstet Vol. 41 No. 9/2019 Pharmacological Treatment for Symptomatic Adenomyosis Benetti-Pinto et al. 569 Although bleeding control was reported by many women, dienogest maintained bleeding in 26.3% of women versus none from the GnRH group. 24 Similarly, uterine volume was reduced according to 2 studies in women who used dieno- gest, but this reduction was lower than that obtained with the study that used GnRH analog (278 /C6 162–151 /C6 117 ml – p ¼ 0.01).24 One of these studies 25 reported hot flushes (5.3%) as a side effect of the dienogest. Treatment with Combined Oral Contraceptives Only one study included one COC to treat adenomyosis, 23 containing 75 mcg of gestodene þ 30 mcg of ethynylestradiol, that was taken for 21 days with 7 days without the pills (21/7), compared to LNG-IUS. The results showed a reduction of pain (6.55 /C6 0.68–3.90 /C6 0.54 - p < 0.001), decreased bleeding (numbers of days), and reduction of the uterine volume, but it was still less efficient that LNG-IUs for all evaluated param- eters (pain - 6.23 /C6 0.67–1.68 /C6 1.25 - p < 0.001). Risk of Bias and Methodological Quality ►Figure 3 discusses the risk of bias from the retrieved studies. Osuga et al 25 presented the lowest risk of bias when analyzing all criteria from this table. Almost all studies (except Osuga et al) 25 presented an unclear risk of bias for allocation concealment and selective reporting. Blinding was only possible in two studies (Osuga et al,25 and Ozdergimenci et al 21). About the GRADE criteria ( ►Table 2 ), all variables presented a moderate certainty assessment, except menstru- al bleeding (number of pads/day), comparing LNG-IUS versus COC for 6 months, that presented low certainty assessment. All studies presented serious imprecision due to the small number of events. Despite being sponsored by the pharma- ceutical industry, the side effects of the Osuga et al study were reported, and, therefore, we do not consider that publication bias was low. Moreover, it was not possible to perform GRADE criteria for the study from Ozdergimenci et al due to inconsistencies, inaccuracy, and poor data description (it would not fit all the criteria for analyzing it).

In the present systematic review, we have shown that the studied treatments (LNG-IUS, aromatase inhibitor [letrozole], GnRH agonist [goserelin and triptorelin], dienogest, and COC [75mcg of gestodene þ 30 mcg of ethinyl estradiol]), were efficient for the control of the two most common symptoms of adenomyosis: heavy menstrual bleeding and dysmenorrhea/ pelvic pain. Equally, regarding enlarged uterus, the treatments Fig. 2 Forest plot for the comparisons of the non-surgical treatments in the reduction of uterine volume, chronic pelvic pain, and menstrual bleeding. Rev Bras Ginecol Obstet Vol. 41 No. 9/2019 Pharmacological Treatment for Symptomatic Adenomyosis Benetti-Pinto et al.570 promoted a reduction in the uterine volume. However, the number of retrieved studies is low and, according to the literature available up to this point, still not enough to endorse any of the analyzed treatments; the follow-up period of these studies was not long enough to permit a conclusion on how effective these treatments would be in the long term (only the LNG-IUS study presented a 12-month follow-up, which can be considered short to medium term). Unfortunately, we also noticed a lack of studies investigating the quality of life as a primary outcome; since most of the studies were concerned with objective outcomes (reduction of uterine volume, and number of days with abnormal bleeding); however, subjective improvement or patient satisfaction should also be consid- ered. 26,27 Moreover, the differences regarding intervention and comparator groups did not allow us to perform metanalysis or subgroup analysis. The most frequent symptoms of adenomyosis are pelvic pain and abnormal uterine bleeding. Dysmenorrhea is present in 50 to 93% of women, while abnormal bleeding is present in 27 to 65%. 28 Despite the different comparators, it seems that the different pharmacological treatments evaluated, dienogest, COC, GnRH analog (triptorelin and goserelin), letrozole, and LNG-IUS, were effective to reduce pelvic pain complaints and to reduce bleeding in women with adenomyosis. These results suggest that hormonal treatment improves the symptoms. The strengths of this review are: the inclusion of randomized controlled trials (RCTs) in a systematic review, and the quality assessment of these studies by the GRADE criteria. However, it is important to mention that our findings were limited by differences in the inclusion criteria of the studies, length of follow-up periods, different comparators, different scales used to measure blood volume loss or pelvic pain severity, which do not allow conclusions about delaying or avoiding surgical procedures. We have found another review in which the authors presented their data in a narrative format, citing the different available treatments. Also, in this publication, it is possible to visualize the dif ficulty of comparing treatments, limiting conclusions about pharmacological treatment. 15 Therefore, the present study is the first systematic review evaluating the results from the pharmacological treatment in adenomyosis, with the intention to promote standardization of the methods and to fulfill gaps in the next prospective studies. With the intention of standardizing future studies, in the present systematic review , we included below some Fig. 3 Risk of bias summary ⊖ ¼ high risk of bias; ? ¼ uncertain risk of bias; ⊕ ¼ low risk of bias. Rev Bras Ginecol Obstet Vol. 41 No. 9/2019 Pharmacological Treatment for Symptomatic Adenomyosis Benetti-Pinto et al. 571 Table 2 Summary of findings table according to Grading of Recommendations Assessment, Development and Evaluation (GRADE) criteria Certainty assessment - Reduction in uterine volume (ml or cc) Patients Effects Certainty Importance Study Study design Bias risk Inconsistency Indirectness Imprecision Other considerations Group 1 Group 2 Relative (95% CI) Absolute(95% CI) Dienogest x Placebo (16 weeks) 1 Randomized trial Not serious Not serious Not serious Serious a Undetected 34 33 /C0 MD 10.4 higher (2.06 lower to 22.86 higher) ⊕⊕⊕/C14 MODERATE Letrozole x GnRH analog (12 weeks) 1 Randomized trial Not serious Not serious Not serious Serious a Undetected 15 16 /C0 MD 20.6 higher (0.09 lower to 41.29 higher) ⊕⊕⊕/C14 MODERATE Dienogest X GnRH analog (16 weeks) 1 Randomized trial Not serious Not serious Not serious Serious a Undetected 19 18 /C0 MD 94 higher (35.68 lower to 223.68 higher) ⊕⊕⊕/C14 MODERATE Certainty assessment – Chronic Pelvic Pain (VAS) Patients Effects Certainty Importance Study Study design Bias risk Inconsistency Indirectness Imprecision Other considerations Group 1 Group 2 Relative (95% CI) Absolute(95% CI) Dienogest x Placebo (16 weeks) 1 Randomized trial Not serious Not serious Not serious Serious a Undetected 34 33 /C0 MD 37.8 lower (49.1 lower to 26.5 lower) ⊕⊕⊕/C14 MODERATE Dienogest x GnRH analog (16 weeks) 1 Randomized trial Not serious Not serious Not serious Serious a Undetected 19 18 /C0 MD 5.3 lower (19.43 lower to 8.83 higher) ⊕⊕⊕/C14 MODERATE Certainty assessment – Hemoglobin level (g/dL) Patients Effects Certainty Importance Study Study design Bias risk Inconsistency Indirectness Imprecision Other considerations Group 1 Group 2 Relative (95% CI) Absolute(95% CI) Dienogest x GnRH analog (16 weeks) 1 Randomized trial Not serious Not serious Not serious Serious a Undetected 19 18 /C0 MD 1.2 lower (2.46 lower to 0.06 higher) ⊕⊕⊕/C14 MODERATE Dienogest x Placebo (16 weeks) 1 Randomized trial Not serious Not serious Not serious Serious a Undetected 34 33 /C0 MD 0.4 higher (0.13 lower to 0.93 higher) ⊕⊕⊕/C14 MODERATE Certainty assessment – Menstrual bleeding (pad/day) Patients Effects Certainty Importance Study Study design Bias risk Inconsistency Indirectness Imprecision Other considerations Group 1 Group 2 Relative (95% CI) Absolute(95% CI) LNG-IUS x COC (6 months) 1 Randomized trial Not serious Not serious Not serious Very serious a Undetected 31 31 /C0 MD 1.74 lower (2.42 lower to 1.06 lower) ⊕⊕/C14/C14 LOW Abbreviations: CI, con fidence interval; COC, combined oral contraceptive; GnRH, gonadotropin-releasing hormone; LVG-IUS, levonorgestrel intrauterine system; MD, mean difference/ a; VAS, visual analogue scale. Total sample size of fewer than 400 patients. Rev Bras Ginecol Obstet Vol. 41 No. 9/2019 Pharmacological Treatment for Symptomatic Adenomyosis Benetti-Pinto et al.572 suggestions that could be considered important to be includ- ed during any checklist for preparing a prospective study for women with symptomatic adenomyosis. More studies are needed to allow comparisons, conclusions on long-term efficacy, and side effects that limit its use. Diagnosis To use pelvic US (preferably transvaginal probe —TVUS and 3DTVUS—when available) or MRI for the diagnosis of adeno- myosis,28,29 describing the presence or absence of at least the following criteria:  globular uterus with regular contours (US or MRI);  asymmetrical thickening of the myometrial walls (US or MRI);  thickening of the junctional zone (JZ) /C21 12 mm (MRI or, eventually, by US);  greatest JZ thickness to total myometrium ratio > 40 to 50% (US or MRI);  foci of high signal intensity running alongside the endo- metrium on T2 and sometimes also T1-weighted;  images that persist on Fat-Sat (FS) (MRI);  anechoic sub endometrial microcysts in the myometrium (around 2 –4 mm in diameter) (US);  description of association or not with leiomyoma. Uterine Volume  To perform the same imaging technique used for diagno- sis, preferably at the same time that clinical complaints are re-evaluated, to correlate the results. Symptoms (Pain and Bleeding)  There is no speci fic questionnaire for adenomyosis  To evaluate pain-related symptoms: pelvic pain, dysme- norrhea, and deep dyspareunia. The assessment criteria should include the VAS. When possible, make daily con- trol diaries for each type of pain and register the frequen- cy of pain.  To evaluate bleeding symptoms through a scale with the number of days of bleeding and number of pads per 30- day (interval). We suggest using the Pictorial Blood As- sessment Chart (PBAC) and serum levels of hemoglobin and ferritin.  Both symptoms (pain and bleeding) should be evaluated at the initial time before treatment, and every 4 months (120 days). Side Effects  To describe in detail all possible and unpredictable side effects, especially when these were indicative of discon- tinuation of treatment, including the number of losses.  To report cases of non-response to pharmacological treatment. Quality of Life  To use the 36-item short-form survey (SF-36) or the WHOQOL questionnaire.

Levonorgestrel intrauterine system and dienogest presented good results for controling bleeding and pelvic pain, respec - tively, versus their comparators. However, there is insufficient data from the retrieved studies to endorse each medication for symptomatic adenomyosis. Future RCTs comparing pharma- cological treatments for adenomyosis are needed to bolster the available data. Contributions All the authors participated actively in the study, as follows: Yela D. A., Benetti-Pinto C. L., and Brito L. G. O. were responsible for writing the protocol and the final manuscript. Teatin-Juliato C. R. and Mira T. A. A. collected the data and conducted a review of the literature. Conflicts of Interest The authors have no con flicts of interest to declare.

1 Levy G, Dehaene A, Laurent N, et al. An update on adenomyosis. Diagn Interv Imaging 2013;94(01):3 –25. Doi: 10.1016/j.diii.2012. 10.012 2 Bazot M, Daraï E. Role of transvaginal sonography and magnetic resonance imaging in the diagnosis of uterine adenomyosis. Fertil Steril 2018;109(03):389–397. Doi: 10.1016/j.fertnstert.2018.01.024 3 Azziz R. Adenomyosis: current perspectives. Obstet Gynecol Clin North Am 1989;16(01):221 –235 4 Brosens JJ, Barker FG, de Souza NM. Myometrial zonal differenti- ation and uterine junctional zone hyperplasia in the non-preg- nant uterus. Hum Reprod Update 1998;4(05):496 –502. Doi: 10.1093/humupd/4.5.496 5 Leyendecker G, Herbertz M, Kunz G, Mall G. Endometriosis results from the dislocation of basal endometrium. Hum Reprod 2002;17 (10):2725–2736. Doi: 10.1093/humrep/17.10.2725 6 Peric H, Fraser IS. The symptomatology of adenomyosis. Best Pract Res Clin Obstet Gynaecol 2006;20(04):547 –555. Doi: 10.1016/j. bpobgyn.2006.01.006 7 Exacoustos C, Manganaro L, Zupi E. Imaging for the evaluation of endometriosis and adenomyosis. Best Pract Res Clin Obstet Gynae- col 2014;28(05):655–681. Doi: 10.1016/j.bpobgyn.2014.04.010 8 Dueholm M, Lundorf E. Transvaginal ultrasound or MRI for diagnosis of adenomyosis. Curr Opin Obstet Gynecol 2007;19 (06):505–512. Doi: 10.1097/GCO.0b013e3282f1bf00 9 Bergeron C, Amant F, Ferenczy A. Pathology and physiopathology of adenomyosis. Best Pract Res Clin Obstet Gynaecol 2006;20(04): 511–521. Doi: 10.1016/j.bpobgyn.2006.01.016 10 Valentini AL, Speca S, Gui B, Soglia G, Miccò M, Bonomo L. Adeno- myosis: from the sign to the diagnosis. Imaging, diagnostic pitfalls and differential diagnosis: a pictorial review. Radiol Med (T orino) 2011;116(08):1267–1287. Doi: 10.1007/s11547-011-0714-5 11 Andres MP, Borrelli GM, Ribeiro J, Baracat EC, Abrão MS, Kho RM. Transvaginal ultrasound for the diagnosis of adenomyosis: sys- tematic review and meta-analysis. J Minim Invasive Gynecol 2018;25(02):257–264. Doi: 10.1016/j.jmig.2017.08.653 12 Younes G, Tulandi T. Conservative surgery for adenomyosis and

a systematic review. J Minim Invasive Gynecol 2018;25 (02):265–276. Doi: 10.1016/j.jmig.2017.07.014 13 Rocha TP , Andres MP , Borrelli GM, Abrão MS. Fertility-sparing treatment of adenomyosis in patients with infertility: a systematic review of current options. Reprod Sci 2018;25(04):480 –486. Doi: 10.1177/1933719118756754 Rev Bras Ginecol Obstet Vol. 41 No. 9/2019 Pharmacological Treatment for Symptomatic Adenomyosis Benetti-Pinto et al. 573 14 de Bruijn AM, Smink M, Lohle PNM, et al. Uterine artery emboli- zation for the treatment of adenomyosis: a systematic review and meta-analysis. J Vasc Interv Radiol 2017;28(12):1629 –1642.e1. Doi: 10.1016/j.jvir.2017.07.034 15 Pontis A, D’Alterio MN, Pirarba S, de Angelis C, Tinelli R, Angioni S. Adenomyosis: a systematic review of medical treatment. Gynecol Endocrinol 2016;32(09):696 –700. Doi: 10.1080/09513590.2016. 1197200 16 National Institute for Health Research. Prospero: International Prospective Register of Systematic Reviews. 2013. https://www. crd.york.ac.uk/prospero/. Accessed September 20, 2018. 17 Moher D, Liberati A, Tetzlaff J, Altman DG; PRISMA Group. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. BMJ 2009;339:b2535. Doi: 10.1136/bmj.b2535 18 Preston DW, Dietz ER. The Art of Experimental Physics. New York, NY: John Wiley & Sons; 1991 19 Higgins JPT, Altman DG. Assessing risk of bias in included studies. In: Higgins JPT, Green S, eds. Cochrane Handbook for Systematic Reviews of Interventions. New York, NY: John Wiley & Sons; 2008:187–241 20 Schünemann H, Brożek J, Guyatt G, Oxman A. GRADE Handbook for Grading Quality of Evidence and Strength of Recommendations. Hamilton: The GRADE Working Group; 2013https://gdt.gradepro. org/app/handbook/handbook.html. Accessed September 20, 2018. 21 Ozdegirmenci O, Kayikcioglu F, Akgul MA, et al. Comparison of levonorgestrel intrauterine system versus hysterectomy on ef fi- cacy and quality of life in patients with adenomyosis. Fertil Steril 2011;95(02):497–502. Doi: 10.1016/j.fertnstert.2010.10.009 22 Badawy AM, Elnashar AM, Mosbah AA. Aromatase inhibitors or gonadotropin-releasing hormone agonists for the management of uterine adenomyosis: a randomized controlled trial. Acta Obstet Gynecol Scand 2012;91(04):489 –495. Doi: 10.1111/j.1600- 0412.2012.01350.x 23 Shaaban OM, Ali MK, Sabra AM, Abd El Aal DE. Levonorgestrel- releasing intrauterine system versus a low-dose combined oral contraceptive for treatment of adenomyotic uteri: a randomized clinical trial. Contraception 2015;92(04):301 –307. Doi: 10.1016/ j.contraception.2015.05.015 24 Fawzy M, Mesbah Y. Comparison of dienogest versus triptorelin acetate in premenopausal women with adenomyosis: a prospec - tive clinical trial. Arch Gynecol Obstet 2015;292(06):1267 –1271. Doi: 10.1007/s00404-015-3755-5 25 Osuga Y, Fujimoto-Okabe H, Hagino A. Evaluation of the ef ficacy and safety of dienogest in the treatment of painful symptoms in patients with adenomyosis: a randomized, double-blind, multi- center, placebo-controlled study. Fertil Steril 2017;108(04): 673–678. Doi: 10.1016/j.fertnstert.2017.07.021 26 Elwyn G, O ’Connor A, Stacey D, et al; International Patient Decision Aids Standards (IPDAS) Collaboration. Developing a quality criteria framework for patient decision aids: online inter- national Delphi consensus process. BMJ 2006;333(7565):417. Doi: 10.1136/bmj.38926.629329.AE 27 Frank L, Basch E, Selby JV; Patient-Centered Outcomes Research Institute. The PCORI perspective on patient-centered outcomes research. JAMA 2014;312(15):1513 –1514. Doi: 10.1001/jama.20 14.11100 28 Gordts S, Grimbizis G, Campo R. Symptoms and classi fication of uterine adenomyosis, including the place of hysteroscopy in diagnosis. Fertil Steril 2018;109(03):380 –388.e1. Doi: 10.1016/ j.fertnstert.2018.01.006 29 Van den Bosch T, Dueholm M, Leone FP , et al. Terms, definitions and measurements to describe sonographic features of myometrium and uterine masses: a consensus opinion from the Morphological Uterus Sonographic Assessment (MUSA) group. Ultrasound Obstet Gynecol 2015;46(03):284–298. Doi: 10.1002/uog.14806 Rev Bras Ginecol Obstet Vol. 41 No. 9/2019 Pharmacological Treatment for Symptomatic Adenomyosis Benetti-Pinto et al.574