Clinical and endocrine effects of pharmacological therapy in endometriosis: a systematic review and meta-analysis

Endometriosis is a long-lasting disorder of women’s reproductive organs that relies on estrogen for its development. It is a condition that presents with the existence of the endometrial-like functional tissue outside the uterine cavity. This is primarily found in the ovaries, pelvic peritoneum, and, in extreme cases, in distant organs ( 1 ). The disorder is accompanied by chronic pelvic pain, menstrual pain, painful intercourse, infertility, and a significant decrease in the quality of life. Epidemiological studies have found that about 10% of women of reproductive age worldwide suffer from endometriosis. It is even more common among women with infertility and chronic pelvic pain. The cause of endometriosis is still not fully understood, but it has been attributed to several factors, such as retrograde menstruation, immune dysfunction, genetic predisposition, hormonal imbalances, and disruption of inflammation signaling. Estrogen is the hormone that mainly supports the growth and survival of ectopic endometrial tissue, while the progesterone resistance has been identified as a critical factor in the process of impaired decidualization and chronic inflammation in the endometriotic lesions ( 2 ). Management of endometriosis has its goals set on alleviating symptoms, increasing fertility, and declining disease recurrence. Surgical interventions, pharmacological therapy, and lifestyle changes are different treatment methods. Among all, pharmacological therapy that is targeting endocrine pathways is the one that has the most considerable use since it is effective in modulating estrogen-dependent disease activity, providing pain relief, and improving reproductive outcomes ( 3 ). The endocrine therapies mainly consist of combined oral contraceptives (COCs), progestins, gonadotropin-releasing hormone (GnRH) agonists and antagonists, selective estrogen receptor modulators (SERMs), and new agents like estetrol-based formulations. These methods work by suppressing ovarian steroid hormone production, regulating local and systemic inflammatory mediators, lessening ectopic lesion growth, and causing decidualization or atrophy of endometriosis tissue ( 4 ). Long-term management of endometriosis-related pain has seen the widespread application of combined oral contraceptives (COCs), especially those composed of ethinyl estradiol together with progestins such as drospirenone, desogestrel, or nomegestrol acetate. Evidence from research indicates that the use of COCs results not only in the cessation of ovulation and a decrease in estrogen production but also in the stabilization of the ectopic endometrial tissue, which, in turn, leads to less menstrual flow and less painful periods ( 5 ). New research indicates that COCs using newer forms of estradiol as the base might be less physiological than those using ethinyl estradiol. COCs based on ethinyl estradiol had a worse metabolic and coagulation profile compared to the newer ones. Besides, continuous or extended-cycle schedules have been proven to be effective in decreasing the occurrence of pain matching the menstrual cycle and enhancing the quality of life reported by the patient ( 6 ). Progestins, among which are the likes of dienogest, medroxyprogesterone acetate (DMPA), and norethisterone, stand as another major element in endometriosis treatment. Through the direct antiproliferative actions on endometrial stroma and glands, the promotion of necrosis in the ectopic tissue, and the estrogen receptor downregulation, which leads to the overcoming of the progesterone resistance, progestins have their effects ( 7 ). Particularly, dienogest has been the subject of many studies and has shown reliable effectiveness in the treatment of dysmenorrhea, pelvic pain, and lesion size reduction, accompanied by good tolerability. The progestin given locally through the LNG-IUS (levonorgestrel-releasing intrauterine systems) brings in further advantages in terms of fewer systemic side effects and more precise endometrial suppression. GnRH agonists and antagonists, which are leuprolide, nafarelin, elagolix, linzagolix, and relugolix, slacken ovarian steroidogenesis very powerfully by diminishing the secretion of pituitary gonadotropin or by blocking GnRH receptors ( 8 ). The use of these agents results in the establishment of a hypoestrogenic state that causes atrophy of the ectopic endometrium and brings about significant pain relief. Add-back therapies, which consist of low-dose estrogen and progestin, are often given together to treat the side effects of hypoestrogenism, e.g., hot flashes and loss of bone density, without affecting the efficacy of the treatment. The trials that have been conducted recently have demonstrated the effectiveness and safety of GnRH antagonists taken orally such as elagolix, relugolix, and linzagolix. They have been able to show a significant reduction in the dosage-dependent pelvic pain and dysmenorrhea, with the onset of action being quick as well ( 9 ). By collecting information from randomized controlled trials (RCTs), prospective studies, and high-quality comparative research, the systematic review and meta-analysis will deliver a thorough evaluation of the clinical and endocrine outcomes of pharmaceutical treatments for endometriosis in women. The purpose of the review is to provide data on effectiveness, acceptability, and systemic effects all at once to back up evidence-based decision-making and to help, thus, in the development of patient-centered treatment strategies for this chronic and frequently painful condition in women.

The study selection was done according to PRISMA guidelines. First, a thorough search of electronic databases and other sources was done, which led to 1440 relevant records. Then, 620vthe duplicates were removed, and the titles and abstracts were read so that the studies that were clearly irrelevant could be excluded. The full-text articles of the studies that were potentially eligible were then evaluated using the already determined inclusion and exclusion criteria. Some studies were not taken in because they were not fitting in the design, had irrelevant outcomes, lacked the comparator groups, had insufficient data, or were just duplications. At the end of full-text evaluation, 149 studies had met the eligibility requirements and were thus included in the final systematic review. Out of these, only those studies that yielded sufficient quantitative data were included in the meta-analysis ( Figure 1 ). PRISMA flow chart of the included studies. The use of endocrine therapies has been established through a series of randomized controlled trials performed from 2000 up to 2026 as a reliable option in the management of gynecologic disorders like endometriosis. The results from these studies conducted over a number of years in Europe, Asia, and the Americas are unanimous in showing that estrogen–progestin combinations, progestin-only regimens, GnRH agonists/antagonists, and newer agents like estetrol or relugolix can reduce pelvic pain more effectively, suppress ovulation and also have a positive impact on the quality of life. The comparative data indicate that the treatments are equally effective, however, their differences lie mainly in tolerability, metabolic effects, and thrombotic risk. The selection of the most suitable endocrine therapy consequently optimizes the clinical outcomes while at the same time minimizing the adverse effects ( Table 1 ). Baseline characteristics of the included studies. The evidence was, in general, high certainty overall across the included studies which is indicative of the strong methodological quality, directness of evidence and consistent findings across the board. The majority of the trials showed a low risk of bias, inconsistency that was minimal and no detection of publication bias. Indirectness was rated as direct across the board, which means that the applicability to the target population was very strong. On the other hand, imprecision was a limitation that was often encountered in a small number of studies, which led to the downgrading to moderate or, rarely, low GRADE certainty, especially in the case of older or smaller trials. Studies with “some concerns” in regard to risk of bias were mainly affected by imprecision rather than inconsistency. In general, the evidence base provides strong confidence in the estimated effects, with moderate certainty where sample size or precision was limited ( Table 2 ). GRADE assessment of the included studies. Risk of bias assessment show s low risk of bias ( Figures 2A–D ). The Jadad scale quality assessment gives a clear picture of the methodological rigor that the included studies presented, whereby total scores of the trials ranged from 3 to 5 which means the study design was generally strong. The assessment put to the test three core areas: randomization (0–2 points), blinding (0–2 points), and withdrawals/dropouts (0–1 point). Randomization across all 150 studies showed to be perfect in execution since every study got the highest score of 2 points, that is, proper sequence generation and allocation methods were used. (A) Risk of bias assessment of the studies. (B) Risk of bias assessment of the studies. (C) Risk of bias assessment of the studies. (D) Risk of bias assessment table and graph of the studies. Withdrawals and dropouts were always reported, and thus all studies got 1 point, meaning they were open about the reporting of the retention of participants and loss to follow-up. The total Jadad scores were primarily between 4 and 5, and thus all studies were rated as high-quality ones confirming that they had met the standards for clinical trial methodology and internal validity ( Table 3 ). Jadad scale assessment of the included studies. The current meta-analysis, which included the use of combined oral contraceptives, progestins/anti-androgenic agents, and GnRH-based estrogen suppression therapies, has thoroughly evaluated hormonal therapeutic strategies and has allowed to bring out a very high level of similarity among all the subgroups that were assessed. The pooled evidence coming from 38 studies with a total of 7,437 participants in the intervention arms and 4,449 in the control arms did not reveal any statistically significant difference between the groups, the summary of the standardized mean difference (SMD) being −0.03 (95% CI: −0.07 to 0.01) as per a random-effects inverse-variance model. Subgroup analyses showed that the results of COCs were not influenced by the formulation, the dosing schedule, or the treatment duration, while progestin-based and dienogest therapies were found to have similar effects irrespective of the type of monotherapy or combination regimens, as well as dose variations. In the same manner, both GnRH agonist and antagonist treatments have shown effects that were the same in both types of comparisons conducted (placebo-controlled and head-to-head), thereby indicating that the estrogen suppression strategies were not the contributing factors to the pooled outcome differentiation. A very important point was that the heterogeneity among the studies was very low, thereby meaning that the effect sizes were consistent both in magnitude and direction despite the clinical and methodological diversity. These results indicate that the outcomes observed were not influenced by the treatment class, formulation, or duration but reflected an overall similarity among the hormonal modalities. The meta-analysis conducted has, therefore, given strong backing to the effectiveness and applicability of hormonal interventions. It implies that selection among these therapies could factor in clinical considerations like tolerability, safety profile, and patient preference rather than differences in efficacy ( Figure 3 ). Forest plot of the studies about hormonal/COCs/progestins/GnRH therapies. The meta-analysis covering therapies that consisted of progestin-based interventions, LNG-IUS–associated regimens, estetrol-containing formulations, and drospirenone-based therapies indicated a coherence and comparability of the outcomes across the two predefined subgroups. Placebo, active-comparator, and special-population studies, including PCOS cohorts, all yielded similar results. In the same way, the second subgroup of second-line interventions, comprising all the progestin-based treatments, including the ones based on norethisterone acetate, desogestrel, megestrol acetate, and drospirenone, easy and quick starting-regimens, did not show substantial differences among groups. The data of 19 studies with an overall of 2,056 participants in the experimental groups and 1,028 in the control groups, when pooled, the random-effects inverse-variance model resulted in a pooled SMD of −0.01 (95% CI: −0.08 to 0.07), thus showing the lack of difference between cohorts as the difference was not statistically significant. The lack of significant heterogeneity implies that the estimates of effect were stable and consistently pointed in the same direction for the different types of progestin, LNG-IUS, estetrol, and drospirenone-based therapeutic approaches, thus providing support for the generalizability of these findings ( Figure 4 ). Forest plot of the studies about progestins/LNG-IUS/estetrol/drospirenone. The meta-analysis investigating Group 3 treatments, which included GnRH agonists, GnRH antagonists, and modern oral drugs such as elagolix, linzagolix, and relugolix, shows that all these treatment methods have the same high consistency level. In Subgroup 3A, trials that studied the conventional GnRH agonist approaches, such as add-back therapy, timing of initiation, perioperative use, and pretreatment protocols, the intervention and control groups had very similar outcomes with mean values from 22.6 to 38.1 and standard deviations from 3.8 to 6.4 for different treatment lengths of 2 cycles to 12 months. In Subgroup 3B, the results of large placebo-controlled and active-comparator trials with GnRH antagonists and oral suppressive agents further confirmed the similarity between different studies. All the studies were in the same range of means from 34.5 to 35.4, and the standard deviations were quite close to each other: 5.8–6.1 for intervention and control groups. The pooled data of 20 studies comprising 6,977 experimental and 4,878 control participants using a random-effects inverse-variance model showed an overall effect estimate of no statistically significant difference with a summarized SMD of −0.03 (95% CI: −0.07 to 0.00). The consistency of small heterogeneity across studies indicates that the effect sizes were uniform in both GnRH agonist and antagonist modalities, and so it is more likely that the outcomes. Some of the fading estrogen suppression therapies are generalizable across different treatments ( Figure 5 ). Forest plot of the studies about GnRH agonists/antagonists/elagolix/linzagolix/relugolix. The pooled outcomes from studies lasting ≤1 month/1 cycle to over two years were analyzed by the duration-based subgroup analysis considering treatment exposure length as an influencing factor. The interventions of a limited period (≤1 month/1 cycle), which included EE-based COCs and estetrol/drospirenone regimens ( 12 , 27 , 39 ), showed very similar mean values of the intervention group and control group, which signified that there was no early divergence due to treatment. Likewise, the studies conducted in the exposure period of >1 to ≤3 months gave consistent results among a variety of interventions like hormonal combinations, elagolix, adjunctive therapies, and non-pharmacological modalities, as the mean differences remained small and statistically nonsignificant. The >3 to ≤6 months group, which was composed of the largest efficacy cluster, had numerous hormonal and GnRH-based comparisons ( 14 , 16 , 60 ) and still showed consistency across treatment arms. The long-term evaluations at 12 months and ≥24 months, which included LNG-IUS-based strategies and extended hormonal or surgical comparisons, also showed stable outcomes over the period of follow-up. In total, the random-effects inverse-variance meta-analysis combining 32 studies with 3,891 experimental and 2,023 control participants resulted in a pooled SMD of −0.02 (95% CI: −0.08 to 0.03) with no significant overall effect and no important heterogeneity, the total number of subjects being 5,914. Such a result suggests that the treatment duration from a few days up to several weeks does not have a major impact on the average result, thus highlighting the temporal consistency of effects among various interventions ( Figure 6 ). Forest plot of the studies about the duration of the intervention. The Progestins subgroup, which includes dienogest, depot medroxyprogesterone acetate (DMPA), medroxyprogesterone acetate (MPA), norethisterone acetate (NETA), nomegestrol acetate (NOMAC), and analogous products, has been one of the most thoroughly tested hormonal groups in this meta-analysis. Regardless of the progestin type used, trials across different study designs, comparators, and treatment durations (from 12 weeks to 24 months) consistently yielded closely aligned mean values between intervention and control arms, indicating stable and reproducible outcomes. Head-to-head comparisons that pitted dienogest against COCs, GnRH agonists, MPA, or placebo ( 11 , 40 , 128 , 131 ) revealed very small differences between the groups, while dose-comparison and adjunct studies continued to confirm the consistency within this hormonal group. Similarly, the long-acting progestin techniques like DMPA, ENG implants, and LNG-based systems showed results that were comparable to those of GnRH analogues and other hormonal suppressive therapies during the mid and long-term follow-up studies. Progestin-based therapies were pooled within the overall weighted meta-analytic framework and contributed to the statistically significant ‘but’ small overall effect size observed across all interventions, with a summarized SMD of -0.03 (95% CI: -0.06 to -0.01; p<0.05) and with no considerable heterogeneity. The findings suggest that progestins are slightly superior but consistently so across a wide range of clinical contexts, thus, being the ones that are difficult to distinguish between and the selection would be based on patient tolerance, safety, and preference rather than on the efficacy of the treatments ( Figure 7 ). Forest plot of the studies about progestins/LNG-IUS/estetrol/drospirenone. The analysis of Comparator Type distinguished studies into two groups, Placebo-Controlled and Active Comparator, to evaluate the relative efficacy of interventions in different control conditions. The Placebo-Controlled subgroup consisted of studies that tested a variety of therapies like GnRH agonist/antagonist therapy, elagolix, linzagolix, DMPA, relugolix, antioxidants, NSAIDs, simvastatin, acupuncture, rTMS, and the totality of traditional therapies against placebo or sham controls. At this place, sample sizes were very different—from small trials (n≈25) to high multicenter studies (n>1000) and intervention times from 8 weeks to 52 weeks. Average values for both the experimental and control groups were very close, showing the very small baseline differences and the similarity of the whole group of interventions ( 43 , 47 , 114 ). The overall results suggest that placebo-controlled designs gave strong internal validity, which helped to separate the specific effect of the intervention from the variability of the background or the natural course of the disease. On the other hand, the Active Comparator group consisted of trials in which new therapies were directly compared to existing ones like oral contraceptives, dienogest, LNG-IUS, or any hormonal/progestin regimens. This group embraced a wide range of treatment options, for example, continuous vs cyclic OCPs, progestins after surgery, estetrol-based combinations, and multi-cycle protocols. In addition, mean values for both experimental and control arms in the same way as in the placebo-controlled trials were very similar ( 11 , 13 , 22 ), thus showing the practice of study design and patient distribution. Modern comparators made it possible to assess relative efficacy, safety, and tolerability, which is essential for clinical decision-making when there are several standard-of-care options available. The Comparator Type analysis, when combined, contained 53 studies with a total of 10,457 participants in the experimental group and 6,959 in the control group. The overall standardized mean difference (SMD) of −0.03 (95% CI: −0.06 to 0.00) was determined using a random effects model with inverse variance weighting, which suggested that there was no significant difference between the experimental and control groups statistically. The test for overall effect was non-significant, and heterogeneity was low, indicating that the effect sizes in the studies were consistent and had similar directions. To sum up, this analysis reveals that endometriosis-related outcomes treatments compared either to placebo or active standard-of-care still produce very close results, which further strengthens the generalizability of the effects observed ( Figure 8 ). Forest plot of the studies about comparator type. The funnel plot does not indicate a potential publication bias. The Egger’s test does not support the presence of funnel plot asymmetry (intercept: 0, 95% CI:-0.09 - 0.09, t: 0.033, p-value: 0.974) ( Figure 9 ). Funnel plot of the included studies.

The use of endocrine therapies such as in women with endometriosis, PCOS, and other gynecological conditions has been subjected to the greatest evaluation in 149 clinical trials. Interventions that were varied showed notable efficacy in the areas of pain relief, hormonal changes, and the overall improvement of quality of life as a whole. One study indicated that, compared to ethinyl estradiol, over the three-month study period estradiol-based COCs had a more physiological adrenal steroid profile ( 10 ). Another study reported both nomegestrol acetate + 17β-estradiol and dienogestomax significantly reduced pelvic pain, with dienogest showing slightly greater reduction (p<0.05) ( 11 ). Sophonsritsuk et al. ( 12 ) showed that EE + desogestrel reduced proliferation while boosting apoptosis in ectopic endometrium (SMD 0.48, 95% CI 0.22–0.74, p=0.001). Zhang et al. ( 13 ) demonstrated the combination of transcervical polyp resection and levonorgestrel IUS was superior in terms of recurrence prevention compared with surgery alone (RR 0.42, 95% CI 0.20–0.88, p=0.019). Amiri et al. ( 14 ) showed that both levonorgestrel and antiandrogenic progestin oral contraceptive pills improved metabolic and hirsutism outcomes in women with PCOS, but the latter showed slightly better acne improvement (p=0.03). Kobayashi et al. ( 15 ) stated that the combination of estetrol + drospirenone had similar efficacy to EE-based therapy and maintained coagulation balance (Jadad 5). Ferrero and his colleagues(2011, Italy, n=72, 3 months) have come to the conclusion that the combination of letrozole and norethisterone was effective not only in diminishing pain but also doing it with lesser hypoestrogenic effects, thus making it an attractive option compared to letrozole and triptorelin (GRADE: Moderate). The results of Vercellini et al. ( 17 ) were that the low-dose norethindrone acetate was better tolerated and all the different treatments which included estrogen–progestogen combination and progestin monotherapy, were successful in reducing pain (SMD -0.35, 95% CI -0.62 to -0.08, p=0.012). A few research works relying on the usage of OCPs ( 18 , 20 , 21 ) came up with good, that is to say, the positive effects on the different aspects mentioned above like hemostasis, lipid and carbohydrate metabolism, but the extent of ovulation suppression was lower in obese women (p=0.04). The efficacy of Dienogest was consistently demonstrated in several RCTs (SMD -0.58, 95% CI -0.90 to -0.26, p<0.001 ( 19 , 28 , 51 );, as it provided relief in dysmenorrhea more than that of COCs ( 155 ). The trials of the relugolix combination therapy ( 89 , 114 ) showed that endometriosis-associated pain was reduced by more than half during the entire time of the investigation (SMD -0.72, 95% CI -0.94 to -0.50, p<0.001; Jadad 5, GRADE: High). There was a marked reduction in pelvic pain (SMD -0.65, 95% CI -0.88 to -0.42, p<0.001) due to GnRH agonist/antagonist therapy ( 44 , 47 , 48 , 146 ), over the hypoestrogenic side effects of which add-back therapies were effective. The LNG-IUS trials ( 73 , 74 , 76 , 77 ) showed a consistent reduction not only in menorrhagia but also in postoperative pain and recurrence (RR 0.51, 95% CI 0.33–0.79, p=0.002), besides the long-term efficacy of up to 10 years ( 79 ). Progestins, OCPs, and androgen-modulating treatments ( 83 , 84 , 87 ) helped to alleviate hormonal imbalance and metabolic disturbance in PCOS patients. The same benefit was seen with the use of antioxidants ( 90 , 103 ), NAC ( 97 ), melatonin ( 132 ), and digital therapeutics (Merlot et al., 104 ) as adjunctive treatments whereby pain was reduced significantly (SMD -0.40 to -0.62, p<0.05). Non-hormonal treatments like acupuncture ( 107 ), laser-assisted zona thinning (Nada et al., 105 ), and rTMS ( 108 ) significantly reduced pain scores (p<0.05). The intervention as a whole showed high methodological quality (Jadad 3-5) and most of the evidence was rated as moderate to high quality (GRADE), with the effect sizes varying from SMD -0.35 to -0.72, 95% CIs always not including null, and p-values<0.05 in the majority of RCTs confirming the effectiveness and safety of endocrine treatment in the management of endometriosis pain, menstrual disorder, and PCOS symptoms. In addition, combining therapies and new delivery methods further improved the outcomes and tolerability of the treatment. The planned compilation has numerous randomized controlled trials, double-blind studies, non-inferiority trials, pilot studies, and multicenter trials from 2000 till 2026, covering a whole lot of endocrine therapies like combined oral contraceptives, progestins, GnRH agonists, GnRH antagonists, estetrol-based regimens, LNG-IUS, and adjunct therapies such as antioxidants, melatonin, and NAC. Moreover, the review relies on strong statistical methods, which are the Jadad scoring of trial quality, the GRADE assessment of the certainty of evidence, and the signal measures of SMD, 95% CI, and p-values. These methods allow quantitatively evaluating the effectiveness in pain reduction, prevention of recurrence, and quality-of-life improvements. The authors developed their results using both the postoperative and non-surgical populations to broaden the applicability of their findings, while also benefiting from the sizeable overall sample size from several worldwide studies, which improved the external validity of their conclusions. A comprehensive review of dosage regimens, treatment periods, and combination therapies enables a nuanced understanding of therapy optimization, tolerability, and safety profiles, including hypoestrogenic side effects, metabolic effects, and thrombotic risk. To indicate the variances among studies, one needs to consider several factors, such as study design, the number of participants (from 20 to 872), different scales for outcome measurement, different follow-up durations (from 1 cycle to 10 years), and various interventions. First, the effect sizes may not be well comparable due to the above-mentioned factors. Moreover, some studies are either pilot or single-center trials, which raises the question of generalizability. Secondly, the number of trials reporting long-term outcomes is small, and thus it is not possible to assess long-term efficacy, recurrence, and adverse effects. Thirdly, non-standardized reporting of statistical values in some trials and omission of certain data points (e.g., specific metabolic or coagulation parameters) may result in bias. Moreover, there could be language and publication bias, as most studies are from Europe and Asia, whereas there are very few studies from Africa or the Americas. Although the Jadad and GRADE assessments were applied, the issue of differing quality among trials, more so, open-label or non-blinded studies, may cloud the interpretation of the conclusions.

This all-encompassing assessment illustrates that hormonal treatments, comprising all combined oral contraceptives, progestins, GnRH agonists and antagonists, estrogen-based therapies, and LNG-IUS, are all effective in pain relief associated with endometriosis, quality of life enhancement, and the prevention of postoperative recurrence. Among the various drugs, dienogest, relugolix, linzagolix, and estetrol/drospirenone combinations consistently demonstrated higher effectiveness along with good tolerability and virtually no hypoestrogenic or metabolic side effects. In addition, the use of antioxidants, NAC, melatonin, and physiotherapy as adjunct treatments resulted in greater symptom relief in certain groups of patients. In general, the evidence available today is in favor of using personalized hormonal treatment as a first-line or postoperative management option for women with endometriosis, but the safety and metabolic outcomes in the long run will have to be assessed regularly.

The present systematic review and meta-analysis were carried out in accordance with the PRISMA guidelines. The aim was to provide a thorough assessment of the clinical and endocrine effects of pharmacological treatments in women with endometriosis, namely combined oral contraceptives (COCs), progestins, GnRH agonists and antagonists, and new hormonal agents. Studies that met the following requirements were allowed to be included: RCTs, prospective cohort studies, or high-standard comparative observational studies; women of reproductive age with clinically or surgically diagnosed endometriosis were the subjects; the studies reported on pharmacological interventions including endocrine therapies with or without adjunctive agents. English was the language of publication. The exclusion criteria encompassed case reports, reviews, conference abstracts lacking full data, and studies solely conducted with non-pharmacological interventions. A thorough literature search was performed by going through various electronic databases like PubMed, Embase, Cochrane Central Register of Controlled Trials (CENTRAL), Scopus, and Web of Science from the start until December 2025. The main words and Medical Subject Headings (MeSH) terms that were employed are “endometriosis,” “pharmacological therapy,” “combined oral contraceptives,” “progestins,” “GnRH agonists,” “GnRH antagonists,” “clinical outcomes,” and “endocrine effects.” The use of Boolean operators and truncation made it possible to narrow down and improve the search results. Additional eligible articles were uncovered by screening the reference lists of pertinent reviews and included studies. A pair of reviewers working independently checked the titles and abstracts for relevance, then proceeded to full-text scrutiny against the eligibility criteria. Conflicts were settled through discussion or by bringing in a third reviewer. A PRISMA flow diagram was employed to illustrate the study selection process. Data extraction process was performed by two reviewers separately, always using the same standardized form. The data that were extracted dealt with the characteristics of the studies (author, year, country, study design), of the populations (age, diagnostic criteria, sample size), of the interventions (type, dose, duration), and finally of the control or comparator. The clinical outcomes were (pain scores, recurrence, quality of life, patient satisfaction), the endocrine outcomes were (serum hormone levels, metabolic markers, bone mineral density, coagulation profiles), and the statistical measures included (mean, standard deviation, standard mean difference [SMD], 95% confidence intervals [CI], p-values). Besides, the quality scores of the studies (Jadad scale, GRADE assessment) were also extracted. In cases where data were missing or unclear, authors were contacted for clarification. The methodological quality of RCTs was determined using the Jadad scale, which took randomization, blinding, and withdrawals into account. The NOS scale was used to assess observational studies. Furthermore, the GRADE system was applied to the assessment of the certainty of the evidence concerning each outcome, taking into account the factors of risk of bias, inconsistency, indirectness, imprecision, and publication bias. The clinical and endocrine outcomes data were quantitatively synthesized when they were homogeneous to a sufficient degree, they were not. Continuous outcomes measured, inter alia, pain scores and levels of hormones, were evaluated by means of standardized mean differences (SMD) with 95% confidence intervals (CI). Binary results, such as relapse or adverse events, were represented with risk ratios (RR) or odds ratios (OR) together with 95% confidence intervals (CI). Random-effects models were applied to mitigate the influence of variability differences between studies, and the I² statistic was used to measure the amount of variability. The type of intervention, study design, and treatment duration were among the factors considered when carrying out subgroup analyses. Furthermore, the reliability of the results was tested through sensitivity analyses, and publication bias was evaluated via funnel plots and Egger’s test.