{"paper_id":"608ecb25-519e-4655-8fdd-888e94bd6284","body_text":"Based on accumulating and consistent epidemiological data, the slope of the incidence curve\nof endometriosis rises rapidly and sharply around the age of 25 years ( Parazzini  et al. , 2020 ). However, while the\ndelay in diagnosis is generally reported to be between 5 and 8 years in adult women ( Chapron  et al. , 2019 ;  Horne and Missmer, 2022 ;  Allaire  et al. , 2023 ), it appears to be\n>10 years in young women ( Becker  et\nal. , 2022 ;  Pino  et\nal. , 2023 ). This is probably due to multiple reasons, including lack of\nawareness of the condition during adolescence among clinicians; normalization of pain by\nfamily, friends and schoolmates; erroneous exclusion of endometriosis after a negative\nultrasound (US) scan; hesitancy of gynaecologists to order pelvic MRI; and reluctance of\nyoung women to undergo laparoscopy for visual diagnosis ( Brosens  et al. , 2013 ;  Youngster  et al. , 2013 ;  ACOG, 2018b ;  Myszko\n et al. , 2020 ).\nIrrespective of the underlying reasons, if the reported data on diagnostic delay are\nreliable, the actual onset of endometriosis in many young women would occur in the early\npostmenarchal years (beginning of peak incidence, 25 years, minus diagnostic delay in young\nwomen, >10 years, equals endometriosis onset at ∼15 years of age). In fact, data on the\nincidence of endometriosis are mostly based on surgical diagnosis, so the actual slope\nshould be shifted to the left anyway, as lesions reasonably develop and causes pain symptoms\nsome years before definitive identification ( Kvaskoff  et al. , 2013 ). Therefore, attention should be focused on\nseverely symptomatic adolescents, who are at increased risk of being affected by the disease\nand could benefit most from secondary preventive interventions to reduce suffering and limit\nlesion progression.\nIn the first part of this opinion piece, we described the remarkable epidemiological\nchanges in reproductive patterns that have occurred over the last two centuries, leading to\nan extraordinary increase in the number of ovulatory menstrual cycles throughout the\nreproductive period and, perhaps most importantly, in the early postmenarchal years.\nRegardless of any additional contributing cause, if ovulatory menstruations play a role in\nthe development of endometriosis and adenomyosis, the 10-fold increase in their number\nbetween menarche and first full-term pregnancy since pre-industrial times ( Eaton  et al. , 2002 ) must be\ncarefully considered as a risk factor that can be modified.\nIn recent decades, several medical interventions have been proposed to control the\nmanifestations of endometriosis and adenomyosis, including short-term hormonal suppression\nof ovulation and surgical removal of the anatomical consequences of both diseases. However,\nthese treatments are not curative and may have limited efficacy. Therefore, the best\ntherapeutic strategy to achieve the above goals is still under debate.\nIn the second part of this opinion piece, we (i) propose the concept ‘ suspect\nendometriosis and adenomyosis in severely symptomatic young women, even when physical and\nUS findings are negative, until proven otherwise ’, with the aim of limiting the\ndiagnostic delay and the associated consequences; (ii) support a secondary prevention\nstrategy by long-term menstrual suppression commencing hormonal therapies promptly after\nclinical suspicion or imaging evidence of one or both diseases, with the aim of relieving\nsymptoms, avoiding lesion progression, and preserving future reproductive potential.\nThe literature search strategy and selection criteria for the evidence reviewed in this\narticle are described in part I ( Vercellini\n et al. , 2023 ).\n\nThe main criticism of the implantation theory is the discrepancy between the almost\nuniversal phenomenon of retrograde menstruation during the reproductive years and the\nrelatively low endometriosis prevalence in the general population ( Parazzini  et al. , 2020 ). This gap has been\ninterpreted both quantitatively, i.e. endometriosis only develops when the amount of\nrefluxed erythrocytes and endometrial fragments exceeds the scavenging capacity of the\nperitoneal macrophages and iron-transport proteins ( D'Hooghe and Debrock, 2002 ;  Wyatt\n et al. , 2023 ), and qualitatively, that is, endometriosis only\ndevelops when pathological endometrium, carrying specific abnormalities that confer to the\nshedding glands an increased capacity to implant on the peritoneum, to proliferate, and to\ninfiltrate tissues, reaches the pelvis ( Vinatier\n et al. , 2000 ).\nThe ‘quantitative’ hypothesis is supported by a large amount of data demonstrating\ndysregulated iron homeostasis in endometriosis patients, caused by (i) an excess of\nerythrocytes entering the pelvis during menses exceeding the degradation capacity of pelvic\nmacrophages ( Donnez  et al. ,\n2016 ), and/or (ii) aberrant expression of iron-transport proteins, i.e. a defective\nprotective iron-sequestration mechanism ( Wyatt\n et al. , 2023 ). The iron overload resulting from repeated\nbleeding episodes would trigger local oxidative stress, maintain a pro-inflammatory state,\ninduce anomalous resistance to ferroptosis (i.e. a form of iron-dependent, non-apoptotic\nprogrammed cell death, caused by toxic lipide peroxidation-mediated membrane damage) and\nprogesterone resistance, and favour ectopic endometrium proliferation ( Ng  et al. , 2020a ;  Li  et al. , 2023 ;  Ma  et al. , 2023 ;  Wyatt  et al. , 2023 ).\nThe ‘qualitative’ hypothesis is based on the differential expression of many molecules, at\nthe gene and/or protein level, observed in the eutopic endometrium of patients with\nendometriosis compared with that of individuals without the disease ( Vinatier  et al. , 2000 ;  Ulukus  et al. , 2006 ). Endometrial abnormalities\nhave been identified in several cellular processes, including, but not limited to,\nproteolysis, angiogenesis, oestrogen synthesis, response to progesterone, and apoptosis\n( Viganò  et al. , 2023 ). In\naddition, eutopic and ectopic endometrial oligoclones carrying somatic mutations in cancer\ndriver genes have been consistently detected in the epithelial cells of the mucosa of\nindividuals with endometriosis and adenomyosis.\nAnglesio  et al.  (2017) \nstudied non-ovarian, infiltrating endometriotic lesions from 39 patients and found somatic\nmutations in the majority of them, including cancer driver mutations in\n ARID1A ,  PIK3CA ,  KRAS , or\n PPP2R1A  in five cases ( Anglesio\n et al. , 2017 ). After sequencing epithelial cells from 107\nendometriomas as well as 82 samples of normal endometrium from control subjects with benign\ngynaecological conditions,  Suda  et\nal.  (2018)  suggested that ovarian endometriosis also develops from\nclonal expansion of endometrial epithelial cells carrying distinct somatic mutations within\ncancer-associated genes, in particular  KRAS .\nOrr  et al.  (2023a ) have\nrecently reported that the presence of  KRAS  mutations in endometriotic\nlesions excised in a series of 122 patients, was associated with greater anatomic disease\nseverity and increased surgical difficulty. Somatic  KRAS  mutations were\nmore frequently detected in subjects with infiltrating fibrotic lesion or endometrioma only\n(11/19; 58%) and mixed subtypes (40/66; 61%), than in those with superficial implants only\n(13/37; 35%).\nInstead of exploring the presence of somatic cancer-driver mutations in the same\nendometriotic lesion type in different patients,  Praetorius  et al.  (2022)  analysed mutations across different\nlesions types excised from the same patient. Alterations on cancer-associated genes were\ndetected in lesions from 13 of 27 study subjects, with more lesions again affected by\n KRAS  changes (15/53 lesions in 6 cases). In nine of these 13 patients\nmutations were identical across distinct lesions. These findings are consistent with\nindividual lesions being oligoclonal, with different lesions within the same patient sharing\na common cell lineage, and with a metastatic model of disease propagation ( Praetorius  et al. , 2022 ).\nRecurring  KRAS  mutations were found by  Inoue  et al.  (2019)  also in 37.1% (26/70) of\npatients with adenomyosis. Oligoclonality was demonstrated, with some mutations identified\nin co-occurring endometriosis.\nAlthough somatic mutations in cancer-associated genes are per se insufficient for malignant\nderailment, several lines of evidence support the hypothesis that intrauterine,\ndeep-invaginating endometrial crypts harbouring these mutations may be selectively\nadvantaged ( Bulun, 2022 ). In particular, the\nemergence of distinct  KRAS -mutated clonal epithelial cell populations\ncharacterized by epigenetically downregulated progesterone receptors, enhanced survival and\nproliferative capacity, and invasiveness, may be considered a key component of the molecular\npathogenesis of both adenomyosis and endometriosis ( Anglesio  et al. , 2017 ;  Suda  et al. , 2018 ;  Inoue\n et al. , 2019 ;  Praetorius  et al. , 2022 ). Identical  KRAS \nmutations have been detected in epithelial cells from the basalis layer of eutopic\nendometrium, adjacent adenomyosis foci and coexistent endometriotic lesions, supporting a\ncommon pathogenic process for both adenomyosis and endometriosis ( Bulun  et al. , 2021 ).\nHowever, both positions have been questioned. On the one hand, the ‘universality’ of\nretrograde menstruation has not been definitively demonstrated, if it must entail the\nperimenstrual presence of endometrial fragments in addition to blood in the peritoneal fluid\n(see Part I of this article –  Vercellini\n et al. , 2023 ). On the other hand, the various abnormalities\nfound in the endometrium of patients with endometriosis have been considered as a potential\nepiphenomenon of the disease itself ( Guo  et\nal. , 2023 ;  Viganò  et\nal. , 2023 ).\nDespite the above criticisms, a now vast body of evidence supports the notion that\nrepetitious episodes of ovulatory menses, by favouring both excessive bleeding and\nintramyometrial entrapment of mutated cell populations as well as their extra-uterine\ndissemination via transtubal retrograde flow, may constitute the early steps in the\nestablishment and persistence of most adenomyosis and endometriosis cases ( Donnez  et al. , 2016 ;  Ng  et al. , 2020a ;  Bulun  et al. , 2021 ,  2023 ;  Bulun, 2022 ;  Kobayashi, 2023 ;  Wyatt  et al. , 2023 ).\nImportantly,  KRAS  mutations have been suggested to confer resistance to\nferroptosis in lung cancer ( Bartolacci  et\nal. , 2022 ) and pancreatic ductal adenocarcinoma ( Li  et al. , 2022 ). The relationship between\n KRAS  mutations and resistance to ferroptosis should also be investigated\nin endometriosis and adenomyosis. A sequence of events starting with iron overload that\ncould lead to  KRAS  mutations and hence resistance to ferroptosis, would\ncombine the quantitative and qualitative theories. In addition, this would further emphasize\nthe importance of pharmacologically reducing the exposure of the uterine wall and the pelvis\nto excessive amounts of blood as a potential source of free iron to reduce the risk of the\nemergence of oxidative stress-generated mutated endometrial oligoclones.\nTherefore, whether the development of endometriosis from retrograde menstruation is a\nquantitative or a qualitative problem or a synergistic effect between the two, transtubal\nreflux of sloughed endometrium should be limited as much as possible in patients with even a\nsuspicion of the disease. Of relevance here, the endometrium can reach a thickness of\n12–16 mm during the late secretory phase ( Nalaboff\n et al. , 2001 ;  D’Arpe\n et al. , 2016 ), whereas the average endometrial thickness is\ngenerally 3–5 mm during protracted use of combined oral contraceptives (COCs) ( ESHRE Capri Workshop Group, 2001 ;  D’Arpe  et al. , 2016 ) and even\nlower during progestogen monotherapy ( ESHRE Capri\nWorkshop Group, 2001 ;  Laganà  et\nal. , 2017 ). Given the state of endometrial atrophy achieved with\nprolonged use of COCs or progestogens, it is tempting to speculate that the menstrual\neffluent associated with bleeding when using these drugs may not predispose to the\ndevelopment of endometriosis to the same extent as that associated with physiological\nmenses. Indeed, the remarkable reduction in the amount of menstrual flow in COC users ( Mansour  et al. , 2017 ) and the\ngenerally scanty irregular bleeding in progestogen users ( Vercellini  et al. , 2016a , b ) should also result in a proportional reduction in the amount\nof transtubal retrograde bleeding.\n\nThe current reproductive pattern is the result of a profound and inalienable social\nevolution in favour of women’s professional advancement, economic and psychological\nindependence, self-determination of their personal future, and an increasingly participatory\nand decision-making role in all aspects of life. There is no going back.\nTo counteract the rising incidence of diseases associated with decades of uninterrupted\novulatory menstruation and excessive oestrogen exposure (i.e. ovarian, endometrial and\nbreast cancer and endometriosis),  Eaton  et\nal.  (2002)  proposed early endocrinological interventions to reduce\naverage serum oestrogen levels and simulate the ancestral hormonal milieu by inducing\npseudopregnancy with COCs.\nHowever, there is currently no evidence to suggest that pharmacological suppression of\novulatory menstruation in all adolescent women at average risk as a primary preventive\nmeasure would significantly reduce the incidence of endometriosis later in life ( Vercellini  et al. , 2011 ).\nFurthermore, considering that the prevalence of endometriosis in the general female\npopulation of reproductive age is ∼5% ( Parazzini\n et al. , 2020 ), several young women would be treated\nunnecessarily to potentially prevent or delay a single case of endometriosis. In addition,\nthe financial implications of such an approach would be burdensome for public health\nsystems, with improper opportunity costs, especially given the uncertain benefits. Finally,\nthe acceptability of and adherence to systematic induction of amenorrhoea immediately after\nmenarche even in asymptomatic girls would likely be very limited, undermining the\neffectiveness of this approach.\nA completely different strategy would be to focus on a selected population subgroup of\nadolescents with strong clinical indicators of early-onset endometriosis and adenomyosis\n( Table 1 ). The likelihood of severe\nmenstrual pain increases significantly with lower menarchal age ( Hoppenbrouwers  et al. , 2016 ), and younger\npatients have higher levels of dysmenorrhoea, dyspareunia and non-cyclic pelvic pain than\nolder ones ( Treloar  et al. ,\n2010 ;  DiVasta  et al. ,\n2018 ;  Wüest  et al. ,\n2023 ). Along this line,  Lund  et\nal.  (2022)  recently observed a robust inverse relationship between age\nat menarche and chronic pain outcomes in adult women. Each additional year (increase) in\nage at menarche was associated with a ≥5-year reduction in the risk of future chronic pain\nsymptoms. The higher oestrogen levels found in menstruators with early menarche persist\nfor years after puberty and may promote the development of chronic pain. Thus, early\nmenarche could be seen as a proxy measure of elevated and pro-inflammatory oestrogen\nexposure during development ( Lund  et\nal. , 2022 ). Furthermore, acute dysmenorrhoea episodes early after\nmenarche, if repeated unremittingly, may not only be an indicator of ensuing endometriosis\nand adenomyosis, but may also promote the transition from acute to chronic pelvic pain\nthrough central sensitization mechanisms and the onset of chronic overlapping pain\nconditions ( Jarrell and Arendt-Nielsen,\n2016a , b ;  de Arruda  et al. , 2022 ).\nIndicators that should raise the suspicion of early-onset endometriosis in\nadolescents despite normal physical examination and pelvic ultrasound findings.*\nACOG (2018b ),  DiVasta  et al.  (2018 ,  2021 ),  Geysenbergh  et al.  (2017) ,  Martire  et al.  (2020 ,\n 2023 ),  Wüest  et al.  (2023) , and  Zannoni  et al. \n(2014) .\nThe notion of subclassifying early dysmenorrhoea into primary (painful menses in the\nabsence of pelvic pathology;  ACOG, 2018b )\nand secondary (painful menses due to pelvic pathology or a recognized medical condition;\n ACOG, 2018b ) may be irrelevant, once the\npresence of specific obstructive genital anomalies that do not completely impede menstrual\noutflow (i.e. a rudimentary, cavitated, non-communicating rudimentary horn, or a didelphic\nuterus with imperforate hemi-vagina) is ruled out ( Fig. 1 ). Indeed, a non-invasive diagnosis of endometriosis is currently proposed\nas the standard of care ( Taylor  et\nal. , 2018 ;  Agarwal  et\nal. , 2019 ;  Chapron  et\nal. , 2019 ;  Becker  et\nal. , 2022 ). Given that neither physical examination nor US and MR\nimaging can reliably exclude the presence of the most common lesion type detected in\npostmenarchal years, i.e. endometriotic superficial peritoneal implants ( Rasp  et al. , 2022 ), it seems\nunclear on what basis painful menstruation can be defined as ‘primary’.\nProposal for a diagnostic and therapeutic algorithm, including self-reported\noutcome measures, for the young menstruator with severe dysmenorrhoea, chronic,\nacyclic pelvic pain symptoms, and a clinical suspicion of early-onset endometriosis\nwho prefers medical suppression of menses to surgery, and accepts, tolerates, and\nhas no contraindications to long-term hormonal treatment.  US,\nultrasonographic scan; MRI, magnetic resonance imaging; LNG-IUS,\nlevonorgestrel-releasing intra-uterine system; PGIC, patient global impression of\nchange 7-point scale ( Guy, 1976 ;  Dworkin  et al. , 2005 );\nCPP, chronic pelvic pain. *After at least 3-month treatment.\n † Based on data from  Vercellini  et al.  (2023a ).  ‡ Central\nSensitization Inventory 0–100 score ( Mayer\n et al. , 2012 ;  Orr\n et al. , 2022 ,  2023b ;  Cetera  et\nal. , 2023a ).  § Based on data from  Neblett  et al.  (2017)  and  Orr  et al.  (2020 ,  2023b ).\nHowever, with proper history taking and accurate US imaging criteria application, the\nprevalence of ovarian and infiltrating fibrotic endometriotic lesions in symptomatic young\nindividuals appears to be higher than previously thought ( Martire  et al. , 2023 ;  Millischer  et al. , 2023 ). Moreover, contrary\nto prior assumptions, mild to moderate adenomyosis is also common in the adolescent\npopulation complaining of heavy menstrual bleeding and dysmenorrhoea, and its frequent\nco-existence with endometriosis ( Exacoustos\n et al. , 2022 ) suggests a common pathogenesis.\nIn a series of 371 young women with severe dysmenorrhoea and heavy menstrual bleeding,\n Martire  et al.  (2023) \nfound US evidence of endometriosis in over one-third of them. In addition to posterior\ninfiltrating fibrotic endometriosis, mainly focal thickening of the uterosacral ligament\n(53%) and small ovarian endometriomas (41%), also adenomyosis, mostly in a mild form, was\nidentified in more than half of the women with endometriotic lesions (67/131, 51%).\nUsing MRI in another large series of 308 adolescents reporting severe dysmenorrhoea\nunresponsive to non-steroidal anti-inflammatory drugs (NSAIDs),  Millischer  et al.  (2023)  confirmed a high\nprevalence of endometriomas, infiltrating and fibrotic endometriosis, and adenomyosis, and\nobserved a linear increase in frequency over time. In the 18–20 year age group, the\nmajority of young menstruators with severe dysmenorrhoea had MRI evidence of endometriosis\nand/or adenomyosis (endometriomas, 21.5%; posterior infiltrating fibrotic endometriosis,\n89.8%; adenomyosis, 21.5%), further supporting the hypothesis that both conditions\nprogress during the early postmenarchal years.\nEventually, two-thirds of adolescents undergoing laparoscopy for severe dysmenorrhoea and\nchronic pelvic pain symptoms have endometriosis ( ACOG, 2018b ;  Hirsch  et\nal. , 2020 ). Thus, endometriosis must always be suspected and treated\npromptly when dysmenorrhoea and chronic acyclic abdominopelvic pain do not respond to\nNSAIDs, interfere with daily and academic activities, and worsen health-related quality of\nlife ( Wüest  et al. , 2023 ;\n Table 1 ).\nAs a secondary prevention measure, ovulatory menstruation could be suppressed in severely\nsymptomatic adolescents from the onset of pelvic pain symptoms or US identification of\nendometriotic and adenomyotic lesions until conception seeking ( ACOG, 2018a ). The primary goal of this neo-evolutionary strategy\nwould be to restore a more physiological menstrual pattern during the currently\nsubstantially prolonged ‘nubility interval’, establishing a pseudopregnancy hormonal\nmilieu in young menstruators with a clinical diagnosis of endometriosis or adenomyosis\n( Eaton  et al. ,\n2002 ).\nMedically induced amenorrhoea may relieve pain, improve quality of life, and limit\ndisease progression ( Unger and Laufer,\n2011 ), tipping the therapeutic balance in favour of this clinico-epidemiological\napproach ( Table 2 ). However, prompt\nhormonal treatment of symptomatic young patients does not seem to constitute the standard\nof care. In a multicentre cross-sectional study,  Pino  et al.  (2023)  found that only one in five adolescents\nreporting symptoms highly suggestive of endometriosis were using COCs or progestogens.\nPathogenic and clinical goals of menstrual suppression in the period from symptom\nonset to conception seeking.\nJarrell and Arendt-Nielsen\n(2016a , b ),  Clemenza  et al.  (2021) ,\nand  de Arruda  et al. \n(2022) .\nAlthough surgery retains a fundamental therapeutic role in the adolescent population\ncomplaining of chronic pelvic pain, whether it should be the first-line approach is a\nmatter of debate ( Brosens  et al. ,\n2013 ;  Gordts  et al. ,\n2015 ;  Laufer and Einarsson, 2019 ),\nespecially considering that (i) endometriosis would not be found in one in three severely\nsymptomatic adolescents undergoing laparoscopy ( Hirsch  et al. , 2020 ;  Becker  et al. , 2022 ); (ii) surgery alone is not always\neffective, or is only partially or temporarily effective in relieving pain ( Youngster  et al. , 2013 );\n(iii) there is no definitive evidence that early removal of endometriotic lesions as a\nstand-alone measure has a major impact on the natural history of the disease and on\noutcomes that matter to patients, i.e. likelihood of pain recurrence and future fertility\n( Evers, 2013 ); (iv) surgery removes\nlesions, not individual disease predisposition, and symptom and lesion recurrence are\nparticularly common in adolescents ( Tandoi\n et al. , 2011 ;  Audebert\n et al. , 2015 ); (v) if long-term postoperative medical therapy\nshould be used anyway ( ACOG, 2018a ;  Zakhari  et al. , 2021 ),\nwhether surgery is indispensable within a strategy of medically induced amenorrhoea for\nyears seems unclear, especially in young menstruators who respond to ovulation suppression\nbefore laparoscopy. In other words, the actual choice would not be between medical\ntreatment or surgery, but indeed between medical treatment or surgery plus medical\ntreatment; (vi) systematic laparoscopy in all adolescents reporting symptoms suggestive of\nearly endometriosis implies proven, albeit limited, harms. Moreover, it would be costly\n( Becker  et al. , 2022 ),\nunlikely to be cost-effective, and would consume large amounts of health care resources\ndespite the unknown number needed to treat; (vii) acceptance of surgery as a first-line\napproach by symptomatic adolescents and their families would likely be limited, reducing\nthe effectiveness of this intervention.\nUnfortunately, there is no convincing evidence that early excision of limited\nendometriotic lesions alone, performed at a specific time during adolescence, is an\neffective secondary prevention measure. Therefore, surgery remains an invaluable treatment\noption in selected young patients who do not respond to, cannot tolerate, have\ncontraindications to, or refuse the use of hormonal medications to suppress repetitive\novulatory menstruation (ROM), but may not be considered as an alternative to medically\ninduced amenorrhoea in a secondary prevention strategy setting. Moreover, surgery removes\nendometriosis but not adenomyosis of the inner myometrium, the form most commonly observed\nin adolescent women ( Martire  et\nal. , 2023 ;  Millischer\n et al. , 2023 ), whereas combined oestrogen–progestogen therapy\nand progestogen monotherapy suppress both diseases.\nFinally, the impact of surgery on ovarian reserve in the presence of endometriomas should\nalso be considered in the context of possible future pregnancy seeking. In addition to the\nfollicular damage caused by the presence of an endometrioma per se, cyst removal inflicts\nfurther injury to the gonadal parenchyma. This has been demonstrated by prospectively\nevaluating the variation in serum anti-Müllerian hormone (AMH) levels before and after\nsurgery and in patients with unilateral or bilateral endometriomas ( Yılmaz  et al. , 2019 ;  Younis  et al. , 2019 ). Overall, a sustained\nreduction in AMH values was observed following endometrioma excision, with an average drop\nof ∼40% when a unilateral cyst was removed and 57% when bilateral cysts were present\n( Younis  et al. ,\n2019 ).\nHowever, the use of hormonal contraceptives also causes a reduction in AMH values, which\ncan vary between 19% and 24% for COCs ( Birch\nPetersen  et al. , 2015 ;  Bernardi  et al. , 2021 ;  Hariton  et al. , 2021 ), 22–65% for vaginal rings, and 23–27% for\ndepot medroxyprogesterone acetate and progestogen implants ( Bernardi  et al. , 2021 ;  Hariton  et al. , 2021 ), while the hormonal IUD\nhas little or no effect ( Bernardi  et\nal. , 2021 ;  Hariton  et\nal. , 2021 ;  Nelson  et\nal. , 2023 ). The decline in AMH with hormonal contraception does not\nappear to be related to cumulative duration of use ( Bernardi  et al. , 2021 ). Importantly, and in\ncontrast to surgery, the suppressive effect of hormonal contraception on AMH levels is\nreversible ( Bernardi  et al. ,\n2021 ), and a return to normal levels has been observed within a few months of COC\ndiscontinuation ( Landersoe  et\nal. , 2020 ).\nThe above data provide further evidence in favour of a medical rather than a surgical\nfirst-line approach, with the aim of preventing both the formation of ovarian\nendometriomas and the potential need for their removal.\n\nMore than 25 years ago,  Brosens (1997) \nargued that suppression of recurrent menstrual bleeding alone should be effective in the\ntreatment of symptomatic endometriosis, as physical elimination of all ectopic endometrial\ncells would be unattainable. The therapeutic goal should therefore be to prevent or suppress\nthe recurrent bleeding associated with endometriotic lesions, with greater emphasis placed\non achieving amenorrhoea than on the degree of hypo-oestrogenism induced by hormonal\ndrugs.\nThis concept seems particularly relevant when choosing hormones to suppress menstruation,\nespecially in adolescents who have not yet reached their peak bone mass. Importantly, we are\nhere dealing with the treatment of oestrogen-dependent diseases and not just contraception.\nThis has implications both for the hormonal content of the drugs used to induce amenorrhoea\nand for the trade-offs between potential benefits and potential harms to be considered in\nthe two different clinical conditions.\nCasper (2017)  warns that the amount of\nethinyl oestradiol (EE) even in low-dose COCs, i.e. those containing 20–30 µg of EE, is\nsupraphysiological. This would impede adequate endometriosis control due to an excessive EE\nstimulatory effect that may not be effectively counteracted by the progestogens included in\navailable COCs. Of relevance here, he argues that 5 µg of EE is equivalent to 1 mg of\nmicronized E2 or 0.625 mg of conjugated equine oestrogens. The dose of oral conjugated\noestrogen required to most closely mimic physiological mean serum oestradiol levels during\nthe reproductive period is between 0.9 and 1.25 mg/day ( Kaunitz, 2000 ). Based on these estimates, even the currently\ndefined ‘low dose’ 20 µg COCs would at least double the average physiological oestrogen\nexposure. For this reason,  Casper (2017) \nsupports the systematic use of progestogen monotherapies rather than COCs as the first-line\ntreatment for endometriosis.\nAlthough this seems sensible when treating adult women who have already reached their peak\nbone mass, it is unclear whether such an approach is safe for adolescents, given that\npharmacological secondary prevention of endometriosis and adenomyosis from clinical\ndiagnosis until conception seeking may imply several years of ovulation suppression. In\nfact, it is well known that available medications can control but not eliminate lesions, and\nsymptoms very often recur soon after drug discontinuation ( Vercellini  et al. , 2016a ).\nAlmost half of adult women’s bone mass is achieved in the first few years after menarche,\nand bone mineral density (BMD) continues to increase after 20 years of age. Although the\nabsolute value attained during adolescence has not yet been shown to be a reliable\npredictor of future pathological fracture risk ( Golden, 2020 ;  Lahoti  et\nal. , 2021 ), limiting the optimal peak BMD by excessively and stably\nreducing serum oestrogen concentrations for several years in the decade between 15 and\n25 years of age could reasonably be considered a risk factor for the development of\npostmenopausal osteoporosis ( Golden,\n2020 ).\nDienogest is currently the reference progestogen monotherapy for the treatment of\nendometriosis and adenomyosis ( Andres  et\nal. , 2015 ;  Murji  et\nal. , 2020 ;  Kobayashi,\n2023 ), with demonstrated antiproliferative and anti-inflammatory effects on\nlesions in addition to pain relief ( Vannuccini\n et al. , 2018 ). However, the use of dienogest for 1 year in\nadolescents was associated with a decrease in lumbar BMD of more than 1% ( Ebert  et al. , 2017 ). Data on\nBMD changes and potential recovery after long-term use of dienogest in young women are not\ncurrently available. Therefore,  Sarıdoğan\n(2015 ,  2017 ) warns that\nprogestogen monotherapies may not be the optimal choice in adolescents with\nendometriosis.\nIn terms of bone loss, a very low oral dose of norethisterone acetate (NETA, 2.5 mg/day)\nmay be a safer alternative to suppress ovulation because of the demonstrated bone-sparing\neffect associated with its androgenic properties and partial conversion to oestrogens\n( Huvinen  et al. , 2021 ;\n American College of Obstetricians and\nGynecologists’ Committee on Clinical Consensus–Gynecology, 2022 ;  Roden, 2023 ). In particular, the use of NETA,\n2.5 mg/day corresponds to the intake of ∼2–5 µg EE/day ( Huvinen  et al. , 2021 ;  American College of Obstetricians and Gynecologists’ Committee on\nClinical Consensus–Gynecology, 2022 ). However, NETA appears to be somewhat less\nwell tolerated than dienogest, mainly due to androgenic-type cutaneous side effects and\nweight gain, and may induce serum lipid changes with unknown long-term effects on\ncardiovascular risk ( Vercellini  et\nal. , 2016b ).\nDepot medroxyprogesterone acetate should not be used for prolonged periods in\nadolescents, also because of consistently demonstrated significant BMD loss and increased\nfracture risk, which led the Food and Drug Administration to issue a black box warning\nalmost 20 years ago. None of the available levonorgestrel-releasing intra-uterine system\n(LNG-IUS) (52, 19.5, and 13.5 mg) inhibit ovulation. This means that there is no adverse\neffect on BMD during their use, as mean serum oestrogen levels are unaffected. Data on the\netonorgestrel subdermal implant in adolescents are very limited and inconclusive.\nThe evidence on the effect of COC use during adolescence on future bone health is\ninconsistent. According to a meta-analysis of nine prospective studies, the impact of COCs\nuse in young individuals is small and the minor BMD reduction observed during medium-term\nfollow-up seems to be of questionable clinical importance (−0.2 g/cm 2  after\n1–2 years of use) ( Goshtasebi  et\nal. , 2019 ). Of relevance, only cyclical, but not continuous, COC use\nseems to limit BMD gains compared to untreated adolescents ( Gersten  et al. , 2016 ). However, it cannot be\nexcluded that prolonged use of COCs with EE content <20 µg, especially if a cyclical\nregimen is adopted, may have a negative impact on bone trophism, although definitive data\non the ultimate risk of fragility fractures are not available ( Golden, 2020 ).\nIt has been suggested that the transdermal delivery of oestrogens, by circumventing the\nliver first-pass metabolism, may prevent the adverse effects on bone health caused by the\nreduction in hepatic insulin-like growth factor-1 (IGF-1) synthesis induced by oral\noestrogens. In fact, IGF-1 stimulates osteoblast differentiation and bone formation (for\nreview, see  Lahoti  et al. ,\n2021 ). Thus, oestrogen-progestogen transdermal patches, and possibly vaginal\nrings, might be considered safe for bone health in young women. Unfortunately, there is a\npaucity of data on the use of these systems in adolescent populations ( Di Meglio  et al. , 2018 ;  Lahoti  et al. , 2021 ).\nThe long-term use of GnRH analogues, agonists and antagonists, in young girls raises\nconcerns even when combined with add-back therapy, so that the trade-offs between such\nmedical treatments and laparoscopy should be carefully considered in adolescents who do\nnot respond to or cannot tolerate first-line medications ( Sarıdoğan, 2015 ,  2017 ;  Becker  et al. ,\n2022 ). Indeed, one of the main adverse effects of GnRH analogues is precisely the\nreduction in BMD during prolonged treatment. As data on the long-term use of GnRH\nanalogues in teenagers are insufficient, caution is warranted, and these therapies should\nonly be chosen if they are clearly successful in patients who have failed other treatment\noptions.\nThe use of COCs has been consistently shown to be associated with a 3- to 5-fold increase\nin the risk of venous thromboembolism (VTE). However, the relative risk information should\nbe translated into absolute risk changes to allow people understand the practical\nindividual implications of COC use. In this regard, the baseline absolute incidence of\nspontaneous VTE in an average-risk adolescent population is between 4 and 11 per 100 000\nwomen per year ( Di Meglio  et al. ,\n2018 ). The risk in young COC users varies between 10 and 30 events per 100 000\nwomen per year ( Powell, 2017 ). Given that\nthe mortality rate from VTE in women aged 20–44 years is <1% ( Manzoli  et al. , 2012 ), in the worst-case\nscenario of COC use in healthy adolescents, approximately one additional death would occur\nper 4–500 000 young women treated annually. The baseline risk of stroke in the 15–19 age\ngroup is even lower, between 3 and 6 per 100 000 per year.\nIt seems fairly clear that, when considering the absolute attributable risk in\nadolescents without known risk factors, the excess of VTE events caused by COC use is not\nsufficient to offset the benefits of ovulation suppression when early-onset endometriosis\nand adenomyosis are diagnosed. Furthermore, patients with severe endometriosis do not\nappear to be at increased risk of VTE compared with the general female population of the\nsame age ( Wiegers  et al. ,\n2022 ).\nIn general, the thromboembolic risk of the available COCs is determined not only by the\ntype of progestogen (third- and fourth-generation progestogens confer a significantly\nhigher risk than second-generation ones), but also by the type and dose of oestrogen they\ncontain. In fact, COCs with ≥30 µg EE are associated with a higher risk of VTE than COCs\nwith ≤20 µg ( Lidegaard  et al. ,\n2011 ;  Stegeman  et al. ,\n2013 ). In addition, the use of micronized 17β-estradiol (E2), or E2 valerate, or\nestetrol instead of EE appears to limit the likelihood of COCs’ side effects, including\nblood pressure increase, adverse serum lipid changes, and thromboembolic events ( Klipping  et al. , 2021 ; Chen\n et al. , 2022 ;  Heikinheimo  et al. , 2022 ;\n Morimont  et al. ,\n2022 ).\nThe transdermal patch significantly increases EE exposure to a greater extent than a\n30 µg EE containing COC ( Di Meglio  et\nal. , 2018 ), and is associated with the highest risk of VTE among\navailable oestrogen-progestogen contraceptive combinations ( Lidegaard  et al. , 2012 ; Galzote  et al. , 2017 ;  Tepper  et al. , 2017 ;  Heikinheimo  et al. , 2022 ). The\nuse of the vaginal ring is also associated with an increased risk of VTE compared with the\nuse of COCs ( Lidegaard  et al. ,\n2012 ).\nOral progestogens, subdermal implant progestogens, and the LNG-IUS are not associated\nwith an increased risk of VTE ( Lidegaard\n et al. , 2012 ;  Heikinheimo  et al. , 2022 ).\nThe relative risk increase for breast cancer in long-term users of COCs is ∼20–30%.\nHowever, the effect is transient and disappears a few years after COC discontinuation\n( Mørch  et al. , 2017 ).\nMoreover, the average relative risk increase in the overall study population translates\ninto very diverse absolute risk increase when COC use in different age groups is\nconsidered ( Hunter, 2017 ). Only two excess\nbreast cancer cases per 100 000 women younger than 35 years were observed in a large\npopulation study in Denmark ( Mørch  et\nal. , 2017 ). In a recent population-based nested case–control study\n( Fitzpatrick  et al. ,\n2023 ), the 15-year absolute excess risk associated with use of COCs or\nprogestogen-only contraceptives from the age of 16 to 20 years was 8 per 100 000 users.\nWhether the type and amount of oestrogen in COCs affects risk is not fully understood\n( Lovett  et al. ,\n2017 ).\nOn the other hand, the use of COCs dramatically reduces the incidence of ovarian cancer,\nwith a clear time-dependent response gradient. In a nationwide cohort study conducted in\nDenmark, the relative risk of ovarian cancer decreased from 0.82 after ≤1 year of use to\n0.26 after >10 years of use. In the study population, use of hormonal contraception\nprevented about one in five ovarian cancers ( Iversen  et al. , 2018 ).\nAccording to the very long-term results of the Royal College of General Practitioners’\nOral Contraception Study, ever-use of COCs was associated with a reduced risk of\ncolorectal (incidence rate ratio (IRR), 0.81), endometrial (IRR, 0.66), ovarian (IRR,\n0.67), and lymphoid and haematopoietic (IRR, 0.74) cancers. The increased risk of breast\nand cervical cancer in current and recent users disappeared after ∼5 years since COC\ndiscontinuation. In this large cohort, approximately one-third of ovarian and endometrial\ncancers and one-fifth of colorectal cancers were prevented by COC use. Indeed, the\nfavourable slowdown in ovarian cancer mortality observed in Europe over the last three\ndecades is largely due to the widespread use of COCs ( Malvezzi  et al. , 2016 ;  Dalmartello  et al. , 2022 ).\nThe effect of COCs on ovarian cancer risk is particularly relevant in women with\nendometriosis because, based on the findings of a recent meta-analysis ( Kvaskoff  et al. , 2021 ), their\nrisk of this malignancy is doubled (summary relative risk, 1.93; 95% CI, 1.68–2.22).\nImportantly, the direction of the association between endometriosis and ovarian cancer\nrisk can be reversed by inhibiting ovulation with COCs for several years. In fact, when\n Modugno  et al.  (2004) \npooled information from four population-based case–control studies of incident epithelial\novarian cancer, they observed an almost 80% risk reduction in patients with a history of\nendometriosis who used COCs for >10 years (odds ratio, 0.21; 95% CI, 0.08–0.58).\nOverall, menstruators should be informed that the net effect of long-term COC use is a\nsmall reduction in overall cancer risk ( Hunter,\n2017 ;  American College of Obstetricians\nand Gynecologists’ Committee on Clinical Consensus–Gynecology, 2022 ).\n\nWhen considering the use of hormones for menstrual suppression, a detailed personal and\nfamily history should be obtained and the World Health Organization’s medical eligibility\ncriteria for contraceptive use should be applied ( Altshuler  et al. , 2015 ). In particular, Categories 3 (‘theoretic\nor proven risks usually outweigh advantages of contraceptive methods’) and 4 (‘unacceptable\nhealth risk if contraceptive method used’) preclude the use of combined\noestrogen-progestogen methods. Guidelines on contraception from the Royal College of\nObstetricians and Gynaecologists (RCOG), the National Institute for Health and Care\nExcellence (NICE), and the Faculty of Sexual and Reproductive Healthcare (FSRH) are\navailable at  https://elearning.rcgp.org.uk/mod/page/view.php?id=6961  (accessed on 24 April\n2023). The U.S. Medical Eligibility Criteria for Contraceptive Use ( https://www.cdc.gov/reproductivehealth/contraception/mmwr/mec/summary.html ;\naccessed on April 24, 2023), published by the Centers for Disease Control and Prevention,\nmay also be consulted to obtain useful information, particularly regarding relative\n(Category 3) or absolute (Category 4) contraindications to the use of COCs.\nIn young menstruators without major contraindications to oestrogen–progestogen\ncombinations, several alternative options are available for ovulation suppression, and\nvarious factors should be considered: (i) overall, oestrogens have a prevalent\npro-inflammatory effect ( Straub, 2007 ;  Cutolo  et al. , 2014 ;  Bulun  et al. , 2019 ,  2021 ), whereas progestogens have an\nanti-inflammatory effect ( Fedotcheva  et\nal. , 2022 ); (ii) oestrogens stimulate endometriotic and adenomyotic\nmetabolism and mitotic activity, whereas progestogens inhibit them; (iii) progestogen\nmonotherapies are theoretically better than COCs in suppressing endometriosis, but exert a\nlarger adverse effect on BMD compared with low-dose COCs; (iv) breakthrough bleeding and\nspotting during menstrual suppression are generally easier to manage with COCs than with\nprogestogens alone; (v) the adolescent’s preference for any of the available treatment\noptions must be given the highest priority, including the choice of surgery and the refusal\nof hormone therapy ( Yong  et al. ,\n2020 ).\nIf oestrogen-progestogen combinations are ultimately chosen, only monophasic COCs should be\nused for continuous, tailored regimens ( Nash\n et al. , 2020 ). To minimize the risk of VTE, COCs containing E2\nvalerate or oestradiol should be preferred, and those with >20 µg EE should be avoided,\nalso to prevent undue activation of endometriotic lesions. Moreover, E2 valerate is\nassociated with a less pro-inflammatory effect compared with EE ( Kangasniemi  et al. , 2020 ).\nThe discontinuation rate of oestrogen-progestogen transdermal patches is particularly high\nin young women, probably because of detachment frequency ( Powell, 2017 ;  Lahoti\n et al. , 2021 ). Vaginal rings are associated with frequent\nspotting and breakthrough bleeding when used continuously ( Vercellini  et al. , 2010 ), and cannot be\nprescribed in adolescents before their sexual debut.\nIf a progestogen is preferred, oral NETA 2.5 mg/day could be the first choice, based on\ngood efficacy, satisfactory bleeding control, bone-sparing activity, and limited cost ( Kaser  et al. , 2012 ;  Vercellini  et al. , 2016a , b ,  2018 ). If NETA is not tolerated due to androgenic-type side effects, switching to\noral dienogest 2 mg/day is indicated. However, due to a reduction in BMD after prolonged\ntreatment ( Ebert  et al. ,\n2017 ;  Kim  et al. ,\n2021 ), the concomitant use of transdermal oestradiol gel, 1 mg/day, is suggested.\nAlternatively, an EP combination licenced in Europe for postmenopausal HRT that contains\ndienogest 2 mg and oestradiol valerate 1 mg, can be used with a continuous, tailored\nregimen.\nProgestogen-only pills and the etonorgestrel 68 mg subdermal implant may not be considered\na valid alternative for menstrual suppression, as their use is associated with frequent\nbreakthrough bleeding and spotting, and amenorrhoea is achieved in only one in five women\n( American College of Obstetricians and\nGynecologists’ Committee on Clinical Consensus–Gynecology, 2022 ;  Edelman  et al. , 2023 ).\nSeveral authoritative international gynaecological scientific societies consistently\nsupport the use of IUDs in adolescents ( AAP,\n2014 ;  Ott  et al. ,\n2014 ;  Black  et al. ,\n2016 ;  Di Meglio  et al. ,\n2018 ;  ACOG, 2018a ;  Margaritis  et al. , 2023 ), but\nwhether LNG-IUSs are appropriate alternatives for menstrual suppression in sexually active\nadolescents with endometriosis is controversial. Some experts suggest that an LNG-IUS should\npreferably be placed at the end of a laparoscopy to avoid the pain and discomfort likely to\nbe experienced when inserting these devices in a young nulligravida ( Sarıdoğan, 2015 ,  2017 ;  Becker  et al. ,\n2022 ).\nAccording to the  American College of Obstetricians\nand Gynecologists’ Committee on Clinical Consensus–Gynecology (2022) ,  ‘For\npatients who may benefit from suppression of ovulation with their method of menstrual\nsuppression, consideration should be given to the unpredictable suppression of ovulation\nwith the LNG-IUD’.  Actually, the results of one RCT showed that an LNG-IUS was\nineffective in preventing the recurrence of ovarian endometrioma after surgery ( Chen  et al. , 2017 ).\nNevertheless, a high proportion of menstruators experience amenorrhoea a few months after\ninsertion of the LNG-IUS, and uterine blood loss can be reduced by more than 90% due to a\ndirect effect on the endometrium ( Abbas  et\nal. , 2020 ). Therefore, if the main complaints are dysmenorrhoea and\nheavy menstrual bleeding and no ovarian endometriomas or infiltrating, fibrotic lesions are\nfound, insertion of the 52 or 19.5 mg LNG-IUS should be discussed appropriately with the\nyoung patient and her parents, taking into account the very long period of efficacy of both\ndevices (8 and 5 years, respectively). Importantly, the LNG-IUS does not have a detrimental\neffect on BMD, and this is relevant when treating symptomatic adolescents. In particular,\nthe LNG-IUS should be included among the first-line options for sexually active young women\nwith early-onset adenomyosis detected at second-level US ( Fig. 2 ).\nProposal for a diagnostic and therapeutic algorithm, including self-reported\noutcome measures, for the young menstruator with heavy menstrual bleeding, severe\ndysmenorrhoea, and a clinical suspicion of early-onset adenomyosis who accepts,\ntolerates, and has no contraindications to long-term hormonal menstrual\nsuppression.  US, ultrasonographic scan; MRI, magnetic resonance imaging;\nLNG-IUS, levonorgestrel-releasing intra-uterine system; PGIC, patient global impression\nof change seven-point scale ( Guy, 1976 ;\n Dworkin  et al. , 2005 ).\n*After at least 3-month treatment.\nGnRH analogues should be proposed as the last medical option for the shortest possible\ntime, when all other pharmacological alternatives have failed and the young patient and her\nparents refuse laparoscopy. Triptorelin 3.75 mg i.m. depot preparations can be injected\nevery 6 weeks instead of every 4 weeks ( Vercellini\n et al. , 2023a ), with tibolone 2.5 mg as adjunctive therapy to\nprevent vasomotor symptoms and BMD decline without risking reactivation of endometriotic\nlesions ( Lindsay  et al. ,\n1996 ;  Taskin  et al. ,\n1997 ;  Castrejón-Delgado  et\nal. , 2021 ). Vitamin D3 and calcium supplementation could be considered\nwhen prolonged treatment is planned.\nGnRH antagonists are gaining momentum for the treatment of endometriosis-associated pain,\nand the results of several phase III RCTs have been published, showing efficacy on pain\nsimilar to that of GnRH agonists, and a safety and tolerability profile directly correlated\nwith the degree of ovarian inhibition achieved ( Yan\n et al. , 2022 ;  Xin\n et al. , 2023 ). In particular, two non-peptide and orally active\nsmall molecules, i.e. elagolix and relugolix, are already marketed for the treatment of\nendometriosis. Combination therapy with oestradiol 1 mg and norethisterone acetate 0.5 mg,\nalso in the same tablet, limits hypoestrogenic side effects and bone resorption and may be\nparticularly convenient to use, thus potentially increasing adherence ( Giudice  et al. , 2022 ).\nIt is now well established that the withdrawal bleeding associated with traditional COCs\nregimens is not physiologically necessary, was originally favoured solely for social,\ncultural and religious reasons, and is currently only offered with the marketing objective\nof increasing acceptance, compliance, and continuation by perceiving cyclical use as\n‘natural’ ( Kaunitz, 2000 ;  Thomas and Ellertson, 2000 ;  Renfree, 2012 ;  Benson and Micks, 2015 ;  American College of Obstetricians and Gynecologists’ Committee on Clinical\nConsensus–Gynecology, 2022 ). Furthermore, at an individual level, continued use\nof COCs does not result in a clinically relevant absolute increase in thromboembolic or\nother adverse events ( Benson and Micks,\n2015 ;  MacGregor and Guillebaud,\n2018 ).\nDespite accumulating data showing that continuous use of COCs is as safe as cyclical use\n( Nash  et al. , 2020 ),\npeople and many doctors still hold the unfounded belief that monthly bleeding is\nphysiological and contributes to female health, ignoring evolutionary evidence to the\ncontrary. As a result, menstruators continue to be prescribed COCs on a cyclical basis,\neven though the potential health benefits of induced monthly uterine bleeding are\ncompletely unknown ( Kaunitz, 2000 ). Indeed,\nthere appear to be good reasons, from evolutionary, pathogenic, and clinical viewpoints,\nto avoid such bleeding episodes in young individuals with a clinical suspicion of\nendometriosis or adenomyosis ( Table 2 ).\nAnticipatory counselling regarding unscheduled bleeding for patients initiating continuous\nCOC or progestogen use is important to limit discomfort and anxiety deriving from\nassociated pain and to provide instructions on how to manage these relatively common\nepisodes ( Zigler and McNicholas, 2017 ;\n American College of Obstetricians and\nGynecologists’ Committee on Clinical Consensus–Gynecology, 2022 ). Women should be\nadvised that these events are generally more frequent in the early months of treatment and\ntend to resolve over time.\nWhen COCs are taken in an uninterrupted fashion, breakthrough bleeding and prolonged\nspotting are frequent. Therefore, the best choice for suppression of ROM in subjects with\nsymptomatic endometriosis is the so-called tailored extended regimen ( Benson and Micks, 2015 ) or continuous flexible\nregimen ( MacGregor and Guillebaud, 2018 ),\nwith 4- to 7-day hormone-free intervals triggered by breakthrough bleeding or prolonged\nspotting of ≥5 days, and followed by resumption of continuous oral contraceptive use until\nthe next bleeding episode ( Sulak  et\nal. , 2006 ;  Jensen  et\nal. , 2012 ;  Hee  et\nal. , 2013 ;  Zorbas  et\nal. , 2015 ;  Vercellini\n et al. , 2016b ;  MacGregor and Guillebaud, 2018 ;  Nash\n et al. , 2020 ).\n\nWith the purpose of immediately suppressing ROM as soon as endometriosis or adenomyosis is\nsuspected or diagnosed and until pregnancy is attempted, the algorithms shown in  Figs 1  and  2  are proposed based on the following assumptions: (i) from an evolutionary\nviewpoint, ROM for decades may not be considered the physiological norm; (ii) ovulation and\nmenstruation are inflammatory events that, if repeated unremittingly, may favour the\nearly-onset of endometriosis and adenomyosis; (iii) the years from menarche to the first\nconsiderable rise of the incidence curve of both diseases appear crucial for lesion\ndevelopment; (iv) endometriosis has been observed in two-thirds of young women undergoing\nlaparoscopy for CPP symptoms, and adenomyosis in one-fifth of adolescents complaining of\nheavy menstrual flow and dysmenorrhoea; (v) a non-surgical diagnosis of both conditions\nbased on history, physical findings and US or MR imaging is valid, reliable, and has been\nconsistently recommended ( Taylor  et\nal. , 2018 ;  Agarwal  et\nal. , 2019 ;  Chapron  et\nal. , 2019 ;  Becker  et\nal. , 2022 ).\nTherefore, suppression of ROM is here intended not only as a treatment for a symptomatic\ncondition, but primarily as a long-term, secondary-prevention interventional\nendocrinological measure aimed at interrupting the oestrogen-based inflammatory and\nfibrogenic mechanisms and, ultimately, at limiting the consequences of the mismatch between\nthe very slow Darwinian genetic adaptation and the currently very rapid environmental (i.e.\nsocial) evolution ( Table 2 ).\nThe suggested clinical algorithms are proposed only for adolescents who have no absolute\ncontraindications to hormone therapy and who accept and tolerate pharmacologically induced\namenorrhoea. The 5 cm diameter cut-off to define an ovarian endometrioma as ‘small’, for\nwhich surgery can be withheld in the absence of suspicious US features, is arbitrary and\nbased on the proposal of  Muzii  et\nal.  (2017) . In adolescents with clinical suspicion of early-onset\nendometriosis ( Fig. 1 ) who do not accept or\ntolerate ( Yong  et al. ,\n2020 ), or have contraindications to, hormonal therapy, laparoscopy should be\nperformed without delay ( Becker  et\nal. , 2022 ).\nThe algorithms take into account the individual patient’s judgement of the outcome of\nmedical interventions for symptom relief. As recommended by the Initiative on Methods,\nMeasurement, and Pain Assessment in Clinical Trials (IMMPACT), we included the Patient\nGlobal Impression of Change (PGIC;  Guy, 1976 )\nas the preferred patient-reported measure to capture women’s assessment of global\nimprovement and satisfaction with treatments for their condition ( Dworkin  et al. , 2005 ). The PGIC is a\nsingle-item, seven-point scale (1 = very much improved; 2 = much improved; 3 = minimally\nimproved; 4 = no change; 5 = minimally worse; 6 = much worse; 7 = very much worse) that\npatients are asked to use to rate their overall condition since commencing treatment ( Dworkin  et al. , 2005 ). For the\npurposes of the proposed algorithms, we have defined the one- and two-point ratings as\n‘response’ and the three- to seven-point ratings as ‘non-response’. The definition of\n‘non-response’ applies after at least 3 months of unsuccessful treatment. We included the\nthree-point rating in the ‘non-response’ category because, in our opinion and experience, a\n‘minimally improved’ status may not be considered satisfactory enough to suggest continuing\ntreatment in an adolescent complaining of severe symptoms and for whom alternative options\nexist.\nIn the algorithm for young patients with CPP symptoms and suspected early-onset\nendometriosis ( Fig. 1 ), we have also included a\nsecond patient-reported outcome measure, the Central Sensitization Inventory (CSI), a\nquestionnaire designed to identify those patients whose pain is complicated by CNS\nsensitization ( Mayer  et al. ,\n2012 ) and who may not respond optimally to conventional treatments ( Cuesta-Vargas  et al. , 2020 ;\n Cetera  et al. , 2023b ). The\nCSI has been validated in both the general chronic pain population ( Cuesta-Vargas  et al. , 2018 ;  Scerbo  et al. , 2018 ) and in\nwomen with endometriosis ( Orr  et\nal. , 2020 ,  2022 ;  Raimondo  et al. , 2023 ;  Cetera  et al. , 2023a ). The CSI\npart A consists of 25 questions, and patients are asked to rate each question from 0 to 4\n(0 = never; 1 = rarely; 2 = sometimes; 3 = often; 4 = always) for a total maximum score of\n100.  Neblett  et al.  (2017) \nestablished the following CSI severity levels: 0–29 points = subclinical; 30–39 points =\nmild; 40–49 = moderate; 50–59 = severe; 60–100 = extreme. A CSI part A cut-off score of 40\npoints has a sensitivity of 78% and a specificity of 80% for detecting women with\nendometriosis and ≥3 co-existing central sensitization syndromes ( Orr  et al. , 2022 ).\nA ≥40 CSI score can be used to screen those endometriosis patients in whom central\nsensitization mechanisms are likely to be involved in determining the overall pain\nexperience ( Orr  et al. ,\n2020 ,  2022 ). In addition, it was\nobserved that individuals with higher baseline CSI scores had, as expected, worse follow-up\noutcomes after surgery for symptomatic endometriosis ( Orr  et al. , 2023b ). Given the very high\nprevalence (∼50%) of a CSI score ≥40 in women with endometriosis ( Orr  et al. , 2022 ;  Raimondo  et al. , 2023 ), we indicated a cut-off\nscore range between ≥40 and ≥60 in the algorithm for early-onset endometriosis ( Fig. 1 ). Thus, patients with moderate, severe, or\nonly extreme central sensitization levels may be considered according to local protocols\n( Cetera  et al. , 2023a ).\nTranslated and psychometrically validated CSI versions in different languages can be found\nat  https://www.pridedallas.com/questionnaires .\nThe addition of a GnRH agonist test solely in non-responders to first-line therapy and in\nthe absence of USA and MRI evidence of ovarian and infiltrating endometriosis, is based on\nthe assumption that lack of improvement after 3 months of profound hypo-oestrogenism\nsubstantially reduces the likelihood that superficial peritoneal disease is the cause of\npain symptoms. Although a response to the GnRH agonist test is not definitive proof of the\npresence of endometriosis, it increases the likelihood enough to consider that laparoscopy\nmay be reasonably indicated ( Practice Committee of\nAmerican Society for Reproductive Medicine, 2008 ;  Howard, 2009 ;  Vercellini\n et al. , 2014 ). Laparoscopy has been included in the final part\nof the algorithm only, unless patients specifically request it early in the diagnostic\nwork-up, also because the therapeutic value of this surgical procedure in the case of\notherwise unidentifiable superficial peritoneal implants has recently been questioned ( Chapron  et al. , 2019 ;  Horne  et al. , 2019 ;  Becker  et al. , 2022 ;  Tucker  et al. , 2023 ).\nThe recommendation for the use of postoperative medical therapy is based on the results of\na recent systematic review and meta-analysis, which showed an impressive reduction in the\nrisk of pain and lesion recurrence after surgery for symptomatic endometriosis in patients\non long-term suppressive therapy compared with those on expectant management alone ( Zakhari  et al. , 2021 ).\nThe above algorithms are amenable to even substantial changes as soon as relevant\nscientific information regarding the pathogenesis, diagnosis, and treatment of endometriosis\nand adenomyosis becomes available.\n\nIn young women, the normalization and dismissal of menstrual pain and heavy menstrual flow,\nand the minimization of non-menstrual, acyclic abdominal pain by parents, relatives,\nteachers, and peers, combined with limited medical awareness of the possible organic causes\nof the reported symptoms, can be extremely detrimental to health-related quality of life in\nall its aspects ( Ng  et al. ,\n2020b ;  Cetera  et al. ,\n2023b ), with possible consequences for the progression of endometriosis and\nadenomyosis and impairment of reproductive potential.\nIn addition, when considering ovulation suppression in general and in adolescents in\nparticular, it is sometimes unclear whether it is the clinician or the patient and her\nfamily who find it more difficult to overcome false myths. Ideally, both parties should be\naware that, in the post-menarchal decade, ROM is not necessarily physiological.\nUnderstanding this concept is even more important when teenagers present with symptoms\nsuggestive of early-onset endometriosis and adenomyosis, as prejudices must not interfere\nwith a correct clinical diagnosis or delay prompt preventive interventional endocrinological\nmeasures. Young women (and their general practitioners, paediatricians, gynaecologists, and\nschool teachers also) should be reassured that hormonal suppression of ROM does not affect\novarian function, which resumes soon after drug discontinuation; does not impair future\nfertility; does not increase the overall risk of cancer; has clinically irrelevant effects\non the individual likelihood of VTE, provided that recognized major risk factors are\nexcluded; and does not in itself lead to significant weight gain.\nIt should be tactfully explained that ovulation suppression should not be considered an\nanomaly and that it can be useful not only for symptom relief but also for the prevention of\nfuture benign and malignant gynaecological conditions. It should also be explained, in\nsimple terms that lay people can understand, that interrupting repetitive cyclic acute\npainful events can reduce the risk of developing central sensitization and overlapping\nchronic pain conditions ( Jarrell and Arendt-Nielsen,\n2016a , b ;  de Arruda  et al. , 2022 ).\nIf ovulation suppression is ultimately chosen, the adolescent and parents should be\ninformed that amenorrhoea may not be achieved easily and immediately ( American College of Obstetricians and Gynecologists’ Committee on\nClinical Consensus–Gynecology, 2022 ). The occurrence of irregular bleeding should\nbe anticipated to avoid unpreparedness and undue anxiety, instructions given on how to\nmanage these events (e.g. tailored cycling), and reassurance given about the generally\ndecreasing frequency of bleeding episodes over time.\nIn adolescents with suspected severe (≥50 points) or extreme (≥60 points) central\nsensitization based on the CSI score and no clinical or imaging evidence of endometriotic\nlesions, the indication for laparoscopy should be considered with caution. Young patients\nand their parents need to be informed that, in these conditions, the likelihood of\nsuccessful and sustained pain relief may be reduced ( Tucker  et al. , 2023 ; Orr  et al. , 2023b ), and that in a small proportion\nof cases pain may even worsen ( Horne  et\nal. , 2019 ). Informed consent should be tailored to this particular\npatient profile, avoiding the use of standard forms. A multidisciplinary CPP assessment\nshould be proposed, and non-pharmacological interventions tried before a final decision is\nmade to proceed with surgery.\nHealthcare providers should be gentle and clear, offer complete and understandable\ninformation, be sensitive and empathetic, describe in detail all available treatment options\nwith their potential benefits and harms, always interact simultaneously with both the\nadolescent and her parents, give ample opportunity to express fears and doubts, and respect\nthe young woman’s preferences and priorities. This would promote a truly shared\ndecision-making process, confirming trust in the doctor and preventing the negative\nperception of the medical profession that can arise in endometriosis patients when they\nreceive suboptimal care ( Ng  et\nal. , 2020b ), and facilitate treatment acceptance and adherence, thus\noptimizing effectiveness.\n\nWe disclose our collective intellectual conflict of interest regarding the role of surgery\nin the management of endometriosis, which we believe arrives too late when lesions are\nalready established. With regard to adenomyosis, surgery is generally considered cumbersome,\noften not radical, and with unpredictable effects. Moreover, early-onset adenomyosis in\nyoung women usually affects the sub-endometrial part of the inner myometrium, making\nexcisional treatments impossible.\nThis opinion article has several limitations, which are listed in Part I ( Vercellini  et al. , 2023b ).\nBriefly, the relevant literature was reviewed comprehensively but not systematically,\ntherefore some important studies may have been overlooked, or we cannot exclude that our\nintellectual competing interest may have led to selective referencing. In addition, the\nquality of the included studies was not formally assessed.\nIn addition, there is currently no evidence that early and sustained suppression of ROM\nimproves long-term clinical outcomes, particularly fertility preservation, and it is\nquestionable to promote extended preventive hormonal intervention on the basis of data on\npathogenesis derived only from cross-sectional and case–control studies. Protracted cohort\nstudies would be needed to justify such a demanding approach. What may be desirable in\nendometriosis research are prolonged, prospective observational studies, that follow the\nnatural history of symptoms and lesions from adolescence to adulthood. Similar studies have\nbeen conducted in the field of reproductive endocrinology (e.g. the transition from\npremenopause to menopause; the Melbourne Women’s Midlife Health Project study) ( Guthrie  et al. , 2004 ) and\nobstetrics (e.g. the effect of stressful events during pregnancy on future events in the\noffspring; the Raine study) ( Straker  et\nal. , 2015 ).\nHowever, our aim was to raise awareness of the current unphysiological postmenarchal\nmenstrual pattern and to stimulate debate about the potentially related pathogenic\ndownstream consequences, not to provide definitive evidence of the validity of our\nconstruct, as this would require decades of future research, including not only prospective\nobservational studies but also intervention studies to assess treatment outcomes ( Shah and Missmer, 2011 ). Moreover, the potential\nbenefits of medically induced amenorrhoea, including pain relief and normalization of\nhealth-related quality of life, are important anyway, whereas the potential harms, including\nVTE and breast cancer, are extremely rare at such a young age. Achievement of peak bone mass\nmay be an issue, and monitoring BMD in young women during prolonged treatment may be\nconsidered. For those adolescents that cannot tolerate the side effects associated with\nfirst-line hormone therapy, such as intractable irregular bleeding, depressed mood, and\ndecreased libido, surgery remains an alternative option.\nJarrell and Arendt-Nielsen (2016a ) suggest\nthat a broader evolutionary perspective, including the notion of a maladaptive status\nbetween biological and cultural evolution leading to recurrent dysmenorrhoea, could modify\nthe concept of what constitutes a normal menstrual pattern, and potentially promote\nprevention and treatment studies that rely on menstrual suppression also.\nIndeed, the future health and reproductive potential of many young menstruators is at\nstake. This may tip the balance in favour of suppression of ROM, regardless of the actual\naetiology of both endometriosis and adenomyosis, as this seems the most prudent course of\naction. It is also urgent to verify whether timely prevention of lesion progression could\nhave an impact on some severe and increasingly common obstetrical complications associated\nwith advanced forms of endometriosis and adenomyosis ( Mandelbaum  et al. , 2023 ;  Park  et al. , 2023 ;  Vercellini  et al. , 2023c ).\nIn 1976, Roger Valentine Short wrote ‘ we should also try to recapture what\ncivilization has destroyed, the ability to keep the ovaries and the female reproductive\ntract in a state of quiescence when reproduction is not desired. Women may be\nphysiologically ill-adapted to spend the greater part of their reproductive lives having\nan endless succession of menstrual cycles ’ ( Short, 1976 ). After almost half a century, perhaps it is time to\nconsider whether his hypothesis is worth testing.","source_license":"CC0","license_restricted":false}