Rhizome

FDA and EMA Endpoints and Pivotal Trial Designs for Endometriosis-Associated Pain Drugs

Chetan Mishra
Chetan Mishra
Sep 27, 2026

Endometriosis-associated pain is a crowded, high-precedent indication. Sponsors developing new hormonal or non-hormonal therapies must choose primary endpoints, rescue-analgesia rules, trial duration, and target populations that both FDA and EMA will accept. Differences between the agencies in labeled indication scope and long-term safety controls can reshape a global development plan, so regulatory and clinical teams need a clear view of what each agency has actually approved.

The analysis below reviews the pivotal programs behind FDA-approved therapies for endometriosis-associated pain, including trial design, co-primary pain endpoints, responder definitions, and duration limits tied to bone safety. It then sets those programs against EMA authorizations for the same indication, comparing endpoint structure, trial length, indication wording, and restrictions on the treated population.

Want Rhizome's help on your own question? Try it for free.

Endometriosis-associated pain: FDA-accepted endpoints and pivotal trial designs compared with EMA decisions

Key takeaways

  • The current FDA standard for endometriosis-associated pain comes from two oral GnRH antagonist programs. Orilissa (elagolix) was approved in July 2018 49 and Myfembree (relugolix/estradiol/norethindrone acetate) received its endometriosis efficacy supplement in August 2022 46193. Each rested on two replicate, randomized, double-blind, placebo-controlled 6-month trials. Each used co-primary responder endpoints for dysmenorrhea (DYS) and non-menstrual pelvic pain (NMPP), and a responder had to reduce pain without increasing analgesic use 82108115.
  • EMA has authorized two products for this use. Ryeqo (relugolix combination) gained the indication in October 2023 161 and Yselty (linzagolix) in November 2024 254. Both used the same endpoint structure: DYS and NMPP responders, corrected for rescue analgesia, in 24-week or 6-month placebo-controlled trials 140169.
  • The main differences between the agencies are the indication wording and the treatment-duration controls, not the endpoints. FDA labels cover "management of moderate to severe pain associated with endometriosis" and set explicit maximum treatment durations driven by bone mineral density (BMD) 229193. EMA restricted both products to women with a history of previous medical or surgical treatment for endometriosis, placing them second line 156167197.
  • EMA accepted one pivotal trial (EDELWEISS 3) for linzagolix, supported by an extension. It approved only the 200 mg plus add-back regimen because the 75 mg regimen failed the NMPP co-primary endpoint 169202204.

FDA: the GnRH antagonist-era evidence standard

Pivotal trial designs

FeatureOrilissa (elagolix)Myfembree (relugolix/E2/NETA)
Pivotal trialsEM-1 (M12-665, N=871) and EM-2 (M12-671, N=815) 6487SPIRIT 1/S1 (NCT03204318) and SPIRIT 2/S2 (NCT03204331) 115116
DesignMultinational, randomized, double-blind, placebo-controlled, parallel-group, 6 months 6487Multinational, randomized, double-blind, placebo-controlled, 24 weeks 115116
Arms3:2:2 placebo : elagolix 150 mg QD : elagolix 200 mg BID 6484871:1:1 combination tablet for 24 weeks : placebo : relugolix 40 mg alone for 12 weeks, then combination for 12 weeks (delayed add-back) 115116
Pain instrumentDaily eDiary Endometriosis Daily Pain Impact Scale (0 to 3, severity and impact on daily activities), averaged over 35 days 8262Daily 11-point NRS (0 to 10) 115116
Primary timepointMonth 3 8263Week 24, using the final 35 days of treatment 108115
Entry pain criteriaComposite Pelvic Signs and Symptoms Score of 6 or more, DYS and NMPP each 2 or more, NMPP on at least 4 of the prior 35 days 87DYS NRS of 4.0 or more on at least 2 days, plus a mean NMPP NRS threshold during run-in 115116

Co-primary endpoint architecture

Both programs used two co-primary responder endpoints, DYS and NMPP. A responder had to meet both of the following 8263108:

  1. A prespecified reduction from baseline in the pain score.
  2. No increase in analgesic use (NSAID or opioid) for endometriosis-associated pain.

The responder thresholds differed by instrument:

ProductDYS thresholdNMPP threshold
Orilissa EM-10.81-point reduction or more on the 0 to 3 scale0.36 or more
Orilissa EM-20.85 or more0.43 or more
Myfembree S1/S22.8-point reduction or more on the 0 to 10 NRS2.1 or more

Orilissa thresholds are from 82. Myfembree thresholds are from 108115.

The elagolix thresholds were trial-specific. They were derived by receiver operating characteristic (ROC) analysis, anchored to the Patient Global Impression of Change (PGIC) 8263.

Results that supported approval

EndpointEM-1 150 mg QDEM-1 200 mg BIDEM-1 placeboEM-2 150 mg QDEM-2 200 mg BIDEM-2 placebo
DYS responders, Month 346%76%20%43%72%23%
NMPP responders, Month 350%55%36%50%58%37%

All elagolix comparisons with placebo were statistically significant 82. For Myfembree, both co-primary endpoints were met in both trials 108:

EndpointS1 MyfembreeS1 placeboS2 MyfembreeS2 placebo
DYS responders, Week 2474.5%26.9%75.1%30.5%
NMPP responders, Week 2458.5%39.6%65.9%42.5%

The Myfembree key secondary endpoints were 108111114:

  • change in DYS NRS and NMPP NRS
  • the EHP-30 pain domain
  • dyspareunia NRS
  • opioid use

What FDA reviewers said about the endpoints

The elagolix review record is the clearest statement of FDA's thinking on endpoints for this indication.

  • Co-primary endpoints and scales. The Division found the co-primary DYS and NMPP endpoints acceptable. It considered the modified daily pain scales appropriate and found the cognitive-debriefing data supportive 86.
    • The Clinical Outcome Assessment reviewer judged the eDiary scales fit for purpose on face validity.
    • The same reviewer noted that the submission had no evidence dossier on reliability, construct validity, ability to detect change, or cultural adaptation 77.
  • Responder thresholds. FDA accepted a responder framework but required the threshold to reflect a pain reduction that is clinically meaningful to patients 8681.
    • FDA treated ROC-derived thresholds as supportive, not preferred. It recommended anchor-based methods as the primary approach, supplemented by cumulative distribution function analyses 60.
    • Earlier reviewers stated that PGIC was "not a well-defined and reliable measure of endometriosis-associated pain in the proposed context of use" 75.
  • Rescue analgesia. FDA did not specifically endorse the sponsor's rule of a 15% or greater increase in rescue analgesic use. It stated that any change in rescue use counted against response should be clinically relevant 75.
    • In the Type A/SPA discussion, FDA said the primary framework should not use a mandatory analgesic threshold, although a threshold-based analysis could serve as a sensitivity analysis 75.
    • FDA still considered it appropriate to build rescue use into the responder definition and required labeling to say so 75.
  • Dyspareunia. FDA found dyspareunia efficacy only for elagolix 200 mg BID, and only at Month 3 899279.
    • FDA had encouraged the sponsor to rank dyspareunia higher in the testing hierarchy and to assess it through Month 6 to show persistence. The sponsor did not include Month 6 dyspareunia in the ranked testing 79.
  • Opioid-sparing claims. The 200 mg BID arm reduced opioid use at Month 6 with statistical significance, but FDA concluded that the mean reduction in opioid pills had not been shown to be clinically relevant 8992.

Drugs@FDA review text for the 2022 Myfembree endometriosis supplement was not retrievable. FDA's specific evaluation of the SPIRIT thresholds, and of any dyspareunia or opioid claims, therefore cannot be confirmed from the review record 193.

Treatment-duration limits driven by BMD

FDA set dose-specific maximum durations based on bone safety:

  • Orilissa: up to 24 months at 150 mg QD and up to 6 months at 200 mg BID 229209.
    • At 6 months, the decline in lumbar spine BMD compared with placebo was about 1% at 150 mg QD and about 3% at 200 mg BID 213.
    • Bone loss did not plateau. FDA expected 24 months at 150 mg QD to cause bone loss no greater than 6 months at 200 mg BID 213.
  • Myfembree: use is limited to 24 months because continued bone loss may not be reversible. The label recommends a baseline BMD assessment and annual BMD assessment for women treated for endometriosis 193.
    • At Month 6 in S1/S2, the least-squares mean change in lumbar spine BMD was -0.72% with Myfembree and +0.12% with placebo 182.
    • The label states that BMD changes beyond 24 months "have not been elucidated" 182.

Legacy GnRH agonist programs

Older FDA-approved endometriosis products used a different evidence model.

  • Synarel (nafarelin): compared with danazol 800 mg/day for 6 months. Efficacy was assessed as relief of pelvic pain, dysmenorrhea and dyspareunia, plus laparoscopic reduction of implants 239.
  • Zoladex (goserelin): two 6-month trials against danazol. Efficacy was assessed from symptoms, pelvic tenderness and induration, and laparoscopic lesion extent. The label noted that the clinical significance of lesion reduction was unknown 231.
  • Lupron Depot add-back studies (M92-878, M97-777): designed mainly around BMD. Efficacy was a secondary objective, measured with investigator- or patient-rated 4-point severity scores for five signs and symptoms 126.

The move from laparoscopic lesion scores and investigator-rated signs to patient-reported daily DYS and NMPP responder endpoints is the main change between the GnRH agonist era and the GnRH antagonist era.

EMA: Ryeqo and Yselty

Ryeqo (relugolix/estradiol/norethisterone acetate)

CHMP adopted a positive opinion on 14 September 2023 (procedure II/0013/G) 248, and the extension was authorized on 30 October 2023 161. The evidence was the same SPIRIT 1 and SPIRIT 2 program that supported the US approval. Both were 24-week, double-blind, placebo-controlled trials with a delayed add-back arm 137147141.

  • Endpoints. The co-primary endpoints were DYS responders (NRS reduction of 2.8 or more with no increase in analgesic use) and NMPP responders (reduction of 2.1 or more with no increase in analgesic use) at Week 24 140. These are the same thresholds used in the US.
  • Results. DYS responders were 74.5% compared with 26.9% on placebo in SPIRIT 1 and 75.2% compared with 30.4% in SPIRIT 2. NMPP responders were 58.5% compared with 39.6% in SPIRIT 1 and 66.0% compared with 42.6% in SPIRIT 2 134143138133.
    • Dyspareunia NRS improved significantly in both trials 139145.
    • In SPIRIT 2, the analgesic pill-count endpoint was not met 145.
  • CHMP view of endpoints. CHMP considered the following acceptable 148:
    • an endpoint package focused on pain
    • NRS as a validated pain measure
    • separate DYS and NMPP measures, justified because treatment can cause amenorrhea
    • a placebo comparator, even though dienogest is authorized in the EU, because placebo also allowed assessment of BMD effects
  • Long-term data. Benefit was maintained through 104 weeks in an open-label extension 153160.
    • Mean change in lumbar spine BMD at Week 104 was -0.45%. Some women lost more than 5% 153157.
    • The risk management plan lists "long-term use beyond 24 months" as a safety concern 157.
    • The product information recommends DXA scanning after one year of treatment 159.
    • The assessment material reviewed does not state a fixed maximum duration for endometriosis 153.

Yselty (linzagolix)

CHMP adopted its opinion on 17 October 2024 (II/0013), and the European Commission decision followed on 22 November 2024 254.

  • Pivotal design. EDELWEISS 3 was the single pivotal trial: randomized, double-blind, double-dummy and placebo-controlled, with up to 6 months of treatment 169.
    • The three arms were placebo (N=162), linzagolix 75 mg without add-back (N=160), and linzagolix 200 mg with add-back (E2 1 mg/NETA 0.5 mg) (N=162) 197.
    • EDELWEISS 6 was a supportive extension to 12 months 169.
    • EDELWEISS 2 ended early (N=84), so it did not provide replication 204.
  • Endpoints. The co-primary endpoints were DYS and NMPP responders at Month 3, based on a 4-point verbal rating scale averaged over the last 28 days 169202.
    • The thresholds were a reduction of 1.10 or more for DYS and 0.80 or more for NMPP, with stable or decreased analgesic use 202.
  • Dose outcome.
    • The 200 mg plus add-back arm met both co-primary endpoints: DYS 72.9% compared with 23.5% on placebo, and NMPP 47.3% compared with 30.9% 202.
    • The 75 mg arm met the DYS endpoint (44.0% compared with 23.5%) but failed NMPP (38.9% compared with 30.9%, p=0.279). It was therefore not accepted for endometriosis 202201.
  • CHMP comments.
    • CHMP noted that the composite responses were driven mainly by the pain-score components, because most participants had stable or reduced analgesic use 207.
    • CHMP accepted the Month 3 primary timepoint because onset was early and the effect was stable through Month 6 174.
    • Dyspareunia results were described as more modest (52.9% compared with 46.2% at Month 6) 206.
    • The sponsor did not seek CHMP scientific advice. It obtained national scientific advice from Sweden 177.

The second-line restriction

EMA's main departure from FDA is the treatment population named in the indication. For both products, CHMP limited the indication to women with a history of previous medical or surgical treatment for their endometriosis.

  • Ryeqo. In the phase 3 program, 98.6% of participants had used endometriosis medication before and 83.2% had had surgery. CHMP considered second-line use consistent with clinical guidelines 156158. The assessment cites the ESHRE 2022 guideline, which treats GnRH antagonists as generally second line because hypoestrogenic effects limit how long they can be used 168.
  • Yselty. CHMP noted that guidelines do not recommend starting treatment with a GnRH agonist or antagonist because of BMD effects. The proposed indication was revised during assessment to match the EDELWEISS 3 population, in which 94% had prior surgery 167174.

No standalone EMA scientific guideline on endometriosis turned up in EMA's guideline collection. The EPARs describe the accepted evidence package, a 6-month placebo-controlled trial with rescue-adjusted DYS and NMPP responder endpoints plus extension data for durability and BMD, as the established approach rather than a codified requirement 168169174. No product-specific FDA guidance on endometriosis turned up either. The only related FDA guidance found was the general guidance on patient-focused clinical outcome assessments 94.

FDA compared with EMA

DimensionFDAEMA
Indication wording"Management of moderate to severe pain associated with endometriosis" 229193"Symptomatic treatment of endometriosis in women with a history of previous medical or surgical treatment" 100101
Co-primary endpointsDYS and NMPP responders, no increase in analgesic use 82108Same structure: DYS and NMPP responders, rescue-adjusted 140202
Primary timepointMonth 3 (elagolix) or Week 24 (relugolix combination) 82108Week 24 (Ryeqo) or Month 3 in a 6-month trial (Yselty) 140169
Number of pivotal trialsTwo replicate trials per product 64115Two for Ryeqo; one pivotal trial plus extension accepted for Yselty 137169204
ComparatorPlacebo 64115Placebo, accepted despite authorized dienogest 148
Responder threshold methodROC with PGIC anchor, which FDA viewed as supportive; FDA prefers anchor-based methods with CDF support 6082Prespecified clinically meaningful change on NRS or VRS 140202
Duration controlHard caps: 24 months (Orilissa 150 mg QD, Myfembree) and 6 months (Orilissa 200 mg BID) 229209193No fixed cap stated in the Ryeqo assessment reviewed; DXA after 1 year; use beyond 24 months listed as a safety concern 153157159
DyspareuniaEfficacy limited to elagolix 200 mg BID at Month 3 7989Secondary endpoint; significant for Ryeqo, modest for Yselty 139145206

Implications for development programs

  • Endpoints. Co-primary DYS and NMPP responder endpoints with an analgesic-use condition are accepted by both agencies. Separating menstrual from non-menstrual pain is needed because hypoestrogenic therapies cause amenorrhea 148.
  • Threshold derivation. Plan an anchor-based derivation of meaningful change with CDF support, and do not rely on ROC alone. Be ready to justify the anchor, since FDA questioned PGIC for this indication 6075.
  • Rescue analgesia. Avoid arbitrary percentage-increase rules in the primary definition. FDA wanted changes in rescue use to be clinically relevant, with any threshold handled as a sensitivity analysis 75.
  • Replication and dose selection. Two replicate trials remain the US precedent. The linzagolix case shows that each dose must meet both co-primary endpoints to be approved, and that one pivotal trial was accepted in the EU 202169.
  • Bone safety. BMD data out to 24 months drive the US labeled duration. In the EU, the lack of longer data is handled through risk management and monitoring 213193157.
  • EU positioning. A population that is mostly previously treated, which is typical of GnRH antagonist trials, should be expected to lead to a second-line EU indication 156167.
Want Rhizome's help on your own question? Try it for free.