Reimbursement Review

Relugolix-Estradiol–Norethindrone Acetate (Myfembree)

Sponsor: Knight Therapeutics Inc.

Therapeutic area: Management of heavy menstrual bleeding associated with uterine fibroids in premenopausal women

This multi-part report includes:

Clinical Review

Pharmacoeconomic Review

Clinical Review

Abbreviations

Executive Summary

An overview of the submission details for the drug under review is provided in Table 1.

Table 1: Background Information on the Application Submitted for Review

NOC = Notice of Compliance.

Introduction

Uterine fibroids (UFs), also known as myomas or leiomyomas, are benign neoplasms of uterine smooth muscle and are the most common tumours in females.¹ UFs are typically diagnosed in people between menarche and menopause. The main risk factors are younger age at menarche, late onset of menopause, African ancestry, and a child-bearing history of few or no children.¹ A common symptom is abnormal uterine bleeding (AUB), particularly heavy menstrual bleeding (HMB), which is reported by 50% to 80% of patients with symptomatic UFs in surveys and trials.^2,3 Infertility, abdominal bloating, constipation, increased urinary frequency, incontinence, dyspareunia, anemia, and fatigue are among the other symptoms. These can affect health-related quality of life (HRQoL).¹ For research purposes, HMB is defined as more than 80 mL of blood loss during each menstrual cycle, based on the direct measurement of menstrual blood loss (MBL) using the alkaline hematin method.⁴ However, from a clinical practice point of view, it is defined as a volume that interferes with a patient's HRQoL, including physical, social, and emotional aspects. Prevalence estimates of UFs vary widely, but have been estimated to range from 4.1% to 5.5%, based on females aged 15 to 49 years in Canada who self-report having physician-diagnosed UFs.⁵

According to the literature and the clinical expert consulted on this review, the overall goals of treatment are to reduce menstrual bleeding, allow for activities of daily living, increase hemoglobin levels, address pain and bulk symptoms, and improve HRQoL.^6,7 The clinical expert also noted that reducing the size of UFs can allow for less invasive surgery, delay or reduce the need for surgery, manage symptoms until menopause (when symptoms tend to subside), and reduce the impact of UFs on fertility. Treatment can include surgical, procedural, and pharmacological options. The choice of treatment depends on the clinical presentation (i.e., symptoms, severity, size, and location of UFs), the patient’s age, and the patient’s goals (i.e., to preserve fertility and/or the uterus). First-line pharmacological options include combined estrogen-progestin contraceptives, progestin-releasing intrauterine systems, and tranexamic acid.⁶ However, the use of hormonal contraceptives for treating UFs is off-label, and HMB is often inadequately controlled with these drugs alone.⁸ For patients whose disease does not respond adequately to first-line treatments, gonadotropin-releasing hormone (GոRΗ) antagonists and agonists may be used with or without hormonal add-back therapy to mitigate hypoestrogenic side effects and protect the endometrium.⁹ GnRH agonists (e.g., leuprolide acetate [LA]) are often used as short-term preoperative therapies (i.e., for 3 months to 6 months) with concomitant iron therapy to reduce the size of UFs and improve anemia before surgery.¹⁰ A caveat with the use of GnRH agonists is the initial flare effect (i.e., symptom exacerbation), which is not seen with GոRΗ antagonists; in addition, once the GnRH agonist is stopped, UF regrowth often occurs within 3 months.⁷ Uterine artery embolization is a procedure used to address HMB and bulk symptoms, but it does not remove UFs, and it is associated with ovarian dysfunction and the need for reintervention.¹¹ Hysteroscopic resection or myomectomy of submucosal UFs are also options; however, these have been associated with high blood loss and UF recurrence.^6,7 Hysterectomy is the only definitive treatment for UFs, and it is associated with significant morbidity and mortality; therefore, it is a last-line option for patients who have persistent symptoms despite other treatments.⁷

The objective of this report is to review and critically appraise the evidence submitted by the sponsor on the beneficial and harmful effects of oral relugolix-estradiol–norethindrone acetate (RGX-E2-NETA) (40 mg relugolix [RGX], 1 mg estradiol [E2], and 0.5 mg norethindrone acetate [NETA]) for the treatment of HMB associated with UFs in patients in the premenopausal stage. RGX-E2-NETA has not been reviewed before by Canada’s Drug Agency (CDA-AMC).

Perspectives of Patients, Clinicians, and Drug Programs

The information in this section is a summary of the input provided by the patient groups that responded to our call for input and from the clinical expert consulted for the purpose of this review.

Patient Input

CDA-AMC received 1 patient group submission from Canadian Women with Fibroids (CANFib), which had conducted a survey of 315 respondents with UFs in 2023. The submission included comments from the incoming Society of Obstetricians and Gynecologists of Canada president and an endorsement from the Women’s Health Coalition. CANFib noted wait times to consult a health care provider, new symptoms, and the ineffectiveness of current medications as constituting important unmet needs for patients. The group reported that most people with UFs experienced heavy bleeding and pain, absences from work, inability to function within their family unit, deficits in their social or professional lives, and eventual depression. Considering the ineffectiveness of the medications that patients have already tried and the long wait times to consult specialists, the respondents felt that more options and earlier treatment would benefit patients and reduce the need for follow-up visits and surgeries. Regarding the experience with currently available treatments, CANFib noted that numerous treatments and medications are available to treat UFs, but there is no cure other than surgical hysterectomy. According to the group, desirable outcomes for patients include access to an effective treatment that does not require a 3-month dosage commitment, regaining confidence, and being able to have meaningful family, social, and professional lives.

CANFib also conducted group discussions with 32 females who had experience with RGX-E2-NETA. Of these participants, 31 reported being “very happy” and 1 reported feeling “indifferent” with respect to how the drug improved symptoms. CANFib added that RGX-E2-NETA enabled those in the discussion group to lead full lives, be with their families, care for their children and/or parents, show up at work, and function without debilitating pain, heavy blood loss, or ongoing depression.

Clinician Input

Input From the Clinical Expert Consulted for This Review

According to the clinical expert consulted for this review, patients want new therapies that adequately and rapidly address symptoms without adverse effects, improve HRQoL, and are simple to administer, with quick resolution upon discontinuation.

Based on input from the clinical expert, first-line treatment of AUB or HMB due to UFs can include nonsteroidal anti-inflammatory drugs, hormonal contraceptives, tranexamic acid, and iron repletion. Once a diagnosis of UF-associated HMB has been confirmed, RGX-E2-NETA could be a treatment option in the absence of endometrial pathology, depending on the patient’s symptom severity, the location of UFs, and the patient’s goals. The clinical expert expects that RGX-E2-NETA will be used in practice to complement and facilitate less invasive medical procedures and surgical approaches, such as hysteroscopic procedures and fertility-preserving surgical approaches. As per the expert, RGX-E2-NETA may be used as a bridge treatment in patients who plan to have gynecological surgery within 3 months (to elevate hemoglobin and/or enable less invasive surgical methods); want to delay surgery for fertility; or anticipate menopause within a couple of years.

The clinical expert indicated that the patients most likely to respond to treatment with RGX-E2-NETA include adults in the premenopausal stage with HMB associated with UFs and bulk symptomatology secondary to UFs. It was also noted that patients with anemia secondary to severe AUB, severe pain, and ureteric obstruction by UFs are most in need of treatment. The patients best suited for treatment with RGX-E2-NETA would be identified based on clinical history, physical examination, and investigations. As part of confirming a diagnosis, endometrial and cervical pathologies must be ruled out based on the patient’s age and risk factors.

As per the clinical expert, the main outcomes used to assess treatment response in practice include qualitative clinical assessments for a reduction in primary symptomatology (i.e., reduced HMB), reduction in pain and bulk symptoms, return to normal function (e.g., activities of daily living and work), and fertility success, if this is the patient’s goal. Clinicians also focus on the patient’s satisfaction with the treatment, rather than quantitative measures, such as MBL volume. However, physicians may measure hemoglobin levels with the goal of increasing them to normal values in patients experiencing anemia. The expert explained that although the outcomes used in clinical trials may be informative, these are generally not used in practice (aside from measuring and improving hemoglobin levels).

The clinical expert stated that RGX-E2-NETA should be discontinued upon successful surgery or procedure, plans for pregnancy, arrival of menopause, a meaningful decline in bone mineral density (BMD), or intolerable adverse effects. Although the Health Canada product monograph states that the use of RGX-E2-NETA should be limited to 24 months, due to the risk of continued and irreversible bone loss, the expert suggested that treatment beyond 2 years may be considered if the patient is maintaining their goals for symptom improvement and has a healthy BMD. However, if RGX-E2-NETA is used to treat HMB and anemia in anticipation of a planned surgery or medical procedure, treatment duration tends to be shorter than 2 years. As noted by the expert, the benefit in preparation for surgery with respect to the reduction of the size of the uterus and fibroids is achieved after 3 months. If wait times are longer than the 3 months, benefit can be maintained by longer use of the treatment.

RGX-E2-NETA would be used in a community setting, acute care hospital, or specialty clinic. While diagnosis and initial prescription of the drug would be performed by an obstetrician-gynecologist, the expert noted that follow-up management and monitoring may be done by a primary care physician.

Clinician Group Input

No input was received from clinician groups for this review.

Drug Program Input

The drug programs identified the following jurisdictional implementation issues: relevant comparators, initiation of therapy, continuation or renewal of therapy, and system and economic issues. The clinical expert consulted provided input on the potential implementation issues raised by the drug programs in Table 4.

Clinical Evidence

Systematic Review

Description of Studies

Two phase III, double-blind, randomized controlled trials (RCTs) — the LIBERTY 1 trial (N = 388) and the LIBERTY 2 trial (N = 382) — of female patients aged 18 to 50 years with a confirmed diagnosis of UFs and HMB (n = 128 and 126, respectively) assessed the efficacy and safety of RGX-E2-NETA (40 mg RGX, 1 mg E2, and 0.5 mg NETA) once daily compared to placebo (n = 128 and n = 129, respectively) at 24 weeks.^12,13 Efficacy was measured using the primary end point of MBL volume responder analyses, defined as patients who experienced an MBL volume of less than 80 mL and a 50% or greater reduction from baseline MBL volume. Other clinically relevant outcomes included improvement in anemia, pain, symptom severity, and HRQoL.

The LIBERTY 1 and LIBERTY 2 trials were mostly identical in design, with the main differences being the countries where the studies took place and the order of statistical testing of secondary outcomes. The mean age across the studies was approximately 42 years. In both studies, the proportion of patients who were Black was lower in the RGX-E2-NETA groups than in the placebo groups (46.1% versus 51.2% in the LIBERTY 1 trial; 49.6% versus 57.4% in the LIBERTY 2 trial). Conversely, the proportion of patients who were white was higher in the RGX-E2-NETA groups than in the placebo groups (50.0% versus 44.1% in the LIBERTY 1 trial; 46.4% versus 38.0% in the LIBERTY 2 trial). Proportions of other races and ethnicities were generally balanced between treatment groups and across the studies. Mean hemoglobin concentration was similar across the treatment groups in both studies; however, anemia (i.e., hemoglobin concentration of less than 12 g/dL) was slightly more prevalent in the placebo groups than in the RGX-E2-NETA groups.

Efficacy Results

Responder Rate for Change in MBL Volume

Response to treatment was defined as having an MBL volume of less than 80 mL and a 50% or greater reduction from baseline in MBL volume over the preceding 35 days of treatment as measured using the alkaline hematin method.

In the LIBERTY 1 trial, 73.4% (95% confidence interval [CI], 64.9% to 80.9%) of patients who received RGX-E2-NETA responded to treatment versus 18.9% (95% CI, 12.5% to 26.8%) of patients who received placebo. The treatment difference in the responder rate for change in MBL volume at week 24 was 54.5% (95% CI, 44.3% to 64.7%; P < 0.0001) in favour of RGX-E2-NETA.

In the LIBERTY 2 trial, 71.2% (95% CI, 62.4% to 79.0%) of patients who received RGX-E2-NETA responded to treatment compared to 14.7% (95% CI, 9.1% to 22.0%) of patients who received placebo. The treatment difference in the responder rate for change in MBL volume at week 24 was 56.5% (95% CI, 46.6% to 66.5%; P < 0.0001) in favour of RGX-E2-NETA.

Improvement of Anemia

For the subset of patients who had a hemoglobin level of 10.5 g/dL or lower at baseline, response was defined as an increase of 2 g/dL or more at week 24.

In the LIBERTY 1 trial, 15 patients (50.0%) and 5 patients (21.7%) in the RGX-E2-NETA and placebo groups, respectively, experienced an increase of 2 g/dL or more in their hemoglobin levels. The treatment difference in the proportion of patients who experienced improvement in anemia at week 24 was 28.3% (95% CI, 3.7% to 52.8%; P = 0.0377) in favour of RGX-E2-NETA.

In the LIBERTY 2 trial, 19 patients (61.3%) and 2 patients (5.4%) in the RGX-E2-NETA and placebo groups, respectively, experienced an increase of 2 g/dL or more in their hemoglobin levels. The treatment difference in the proportion of patients who experienced an improvement in anemia at week 24 was 55.9% (95% CI, 37.3% to 74.5%; P < 0.0001) in favour of RGX-E2-NETA.

Improvement of Pain

For the subset of patients who had a maximum numerical rating scale (NRS) score of 4 points or greater during the 35 days before randomization, response to treatment was defined as a maximum NRS score of 0 points or 1 point for UF-associated pain over the preceding 35 days of treatment.

In the LIBERTY 1 trial, 33 patients (39.3%) and 11 patients (11.6%) in the RGX-E2-NETA and placebo groups, respectively, experienced a maximum NRS score of 0 points or 1 point for UF-associated pain. The treatment difference in the proportion of patients who experienced an improvement in pain to a maximum NRS score of 0 point or 1 point was 27.7% (95% CI, 15.4% to 40.0%; P < 0.0001) in favour of RGX-E2-NETA.

In the LIBERTY 2 trial, 34 patients (36.6%) and 17 patients (17.9%) in the RGX-E2-NETA and placebo groups, respectively, experienced a maximum NRS score of 0 points or 1 point for UF-associated pain. The treatment difference in the proportion of patients who experienced improvement in pain to a maximum NRS score of 0 point or 1 point was 18.7% (95% CI, 6.2% to 31.1%; P = 0.0056) in favour of RGX-E2-NETA.

Change in Uterine Fibroid Symptom Quality of Life Questionnaire Symptom Severity Scale Score and HRQoL Total Score

The Uterine Fibroid Symptom Quality of Life questionnaire (UFS-QoL) is a disease-specific instrument used to assess symptom severity and HRQoL in patients with UFs over the preceding month.¹⁴ The UFS-QoL Symptom Severity Scale (SSS) consists of 8 items rated on a 5-point Likert scale ranging from “not at all” to “a very great deal.” The HRQoL component includes 29 items rated on a 5-point scale ranging from “none of the time” to “all of the time.” Scores for these SSS and HRQoL subscales are linearly transformed to scales ranging from 0 points to 100 points. Higher SSS scores indicate more severe UF symptoms, while higher HRQoL total scores indicate better quality of life.

In the LIBERTY 1 trial, the least squares mean (LSM) changes from baseline to week 24 in UFS-QoL SSS scores were –30.9 points (95% CI, –36.1 points to –25.8 points) and –10.5 points (95% CI, –15.5 points to –5.5 points) for patients in the RGX-E2-NETA and placebo groups, respectively. The treatment difference in LSM change from baseline to week 24 in the UFS-QoL SSS score was –20.4 points (95% CI, –27.1 to –13.7 points) in favour of RGX-E2-NETA.

In the LIBERTY 2 trial, the LSM changes from baseline to week 24 in UFS-QoL SSS scores were –36.1 points (95% CI, –41.2 points to –31.0 points) and –13.7 points (95% CI, –18.8 points to –8.5 points) for patients in the RGX-E2-NETA and placebo groups, respectively. The treatment difference in LSM change from baseline to week 24 in UFS-QoL SSS score was –22.4 points (95% CI, –29.4 points to –15.4 points) in favour of RGX-E2-NETA.

In the LIBERTY 1 trial, the LSM changes from baseline to week 24 in UFS-QoL HRQoL total score were 38.0 points (95% CI, 32.4 points to 43.5 points) and 12.8 points (95% CI, 7.5 points to 18.2 points) for patients in the RGX-E2-NETA and placebo groups, respectively. The treatment difference in the LSM change from baseline to week 24 in UFS-QoL HRQoL total score was 25.1 points (95% CI, 17.9 points to 32.3 points) in favour of RGX-E2-NETA.

In the LIBERTY 2 trial, the LSM changes from baseline to week 24 in UFS-QoL HRQoL total score were 37.8 points (95% CI, 32.9 points to 42.6 points) and 13.8 points (95% CI, 8.9 points to 18.8 points) for patients in the RGX-E2-NETA and placebo groups, respectively. The treatment difference in the LSM change from baseline to week 24 in UFS-QoL HRQoL total score was 23.9 points (95% CI, 17.2 points to 30.7 points) in favour of RGX-E2-NETA.

Harms Results

In the LIBERTY 1 trial, 61.7% of patients who received RGX-E2-NETA and 66.1% of patients who received placebo had a treatment-emergent adverse event (TEAE). Headache (10.9% of patients in the RGX-E2-NETA group versus 15.0% of patients in the placebo group) and hot flush (10.9% of patients in the RGX-E2-NETA group versus 7.9% of patients in the placebo group) were the most common TEAEs. In the LIBERTY 2 trial, 60.3% of patients who received RGX-E2-NETA and 58.9% of patients who received placebo had a TEAE. Headache (8.7% of patients in the RGX-E2-NETA group versus 11.6% of patients in the placebo group) was the most common TEAE. In general, the types and frequencies of TEAEs were similar between the treatment groups for both studies and across studies.

In the LIBERTY 1 trial, 7 patients (5.5%) who received RGX-E2-NETA reported 10 serious adverse events (SAEs), and 2 patients (1.6%) who received placebo reported 2 SAEs. In the LIBERTY 2 trial, 1 patient (0.8%) who received RGX-E2-NETA reported 1 SAE, and 4 patients (3.1%) who received placebo reported 5 SAEs.

In the LIBERTY 1 trial, 7 patients (5.5%) who received RGX-E2-NETA stopped treatment due to 12 TEAEs, and 5 patients (3.9%) who received placebo stopped treatment due to 13 TEAEs. In the LIBERTY 2 trial, 3 patients (2.4%) who received RGX-E2-NETA stopped treatment due to 3 TEAEs, and 6 patients (4.7%) who received placebo stopped treatment due to 7 TEAEs.

There were no deaths reported in either study.

The treatment differences in the percentage change from baseline to week 24 in BMD of the lumbar spine were –0.41% (95% CI, –1.16% to 0.34%) in the LIBERTY 1 trial and –0.44% (95% CI, –1.23% to 0.34%) in the LIBERTY 2 trial.

Critical Appraisal

In both trials, a lower proportion of patients were Black and a higher proportion were white in the RGX-E2-NETA groups versus the placebo groups (race and ethnicity are risk factors for developing UFs),¹⁵ and mean MBL volume was higher in the RGX-E2-NETA group versus the placebo group. While the baseline imbalances could have biased the study results (i.e., the direction and magnitude of bias are unknown), the clinical expert consulted on this review did not expect the imbalances to meaningfully affect the interpretation of the results. There was a potential risk of bias in the outcomes for improvement in anemia and pain, and the treatment effects were notably different between the studies for both. The reason for the inconsistency was unclear, but it may have had to do with the smaller number of patients contributing data (the outcomes used subpopulations) and/or a loss of prognostic balance. Patients who had anemia at baseline or new anemia during the study started or continued iron therapy. While the iron supplementation protocol was consistent across the treatment groups and studies, it is unknown if improvement in anemia was a result of iron therapy or reduced MBL; information on changes in the amount of iron throughout the studies was not available to clarify this. Bias may also have arisen from the missing data (5% to 30%) for the anemia, pain, UFS-QoL, and BMD outcomes (explanations for why these data were missing were not available); this reduces certainty in the results. Considering the large difference in efficacy between the RGX-E2-NETA and placebo groups for the primary outcome, patients may have been able to infer which treatment they received in the study, which would have affected the subjective outcomes (i.e., pain, UFS-QoL, and harms).

According to the clinical expert, 2 eligibility criteria that do not reflect clinical practice were exclusions based on an upper age limit (patients had to be adults in the premenopausal stage) and receiving gynecological surgery or ablation procedures for UFs within the 6 months before treatment. Moreover, the definition of HMB used in the clinical trials, while acceptable for the setting, is not used in practice. Instead, HMB is identified through clinical assessment and described as AUB that interferes with a patient’s HRQoL and activities of daily living. Furthermore, although the study end points inform on outcomes of importance to patients (i.e., pain, bulk symptoms, and HRQoL), these are not used in practice. Rather than quantifying treatment effect by measuring MBL, clinicians assess patient satisfaction with how well the treatment is addressing their disease symptoms and whether the patient can participate in daily activities without interference from HMB and UFs. The 24-week study duration was deemed acceptable in terms of meeting the clinical end points; however, the clinical expert does not expect most patients to use RGX-E2-NETA as a long-term treatment. Most would use it instead as a bridging therapy to surgery or menopause. The expert indicated that there are rare instances in which treatment beyond 24 months may be warranted, such as in patients whose disease responds well to treatment without intolerable TEAEs; however, it is noted in the Health Canada product monograph that due to the risk of irreversible bone loss, use of the drug should be limited to 24 months.¹⁶ Should the drug be used for the longer term, regular BMD scans would be important to ensure that the patient is not experiencing clinically important bone loss. If this develops, discontinuation of the drug should be considered.

GRADE Summary of Findings and Certainty of the Evidence

For the pivotal studies and RCTs identified in the sponsor’s systematic review, Grading of Recommendations Assessment, Development and Evaluation (GRADE) was used to assess the certainty of the evidence for the outcomes considered most relevant to inform expert committee deliberations, and a final certainty rating was determined, as outlined by the GRADE Working Group.^17,18

Following the GRADE approach, evidence from RCTs started as high-certainty evidence and could be rated down for concerns related to study limitations (which refer to internal validity or risk of bias), inconsistency across studies, indirectness, imprecision of effects, and publication bias.

When possible, certainty was rated in the context of the presence of an important (nontrivial) treatment effect; if this was not possible, certainty was rated in the context of the presence of any treatment effect (i.e., the clinical importance is unclear). In all cases, the target of the certainty of evidence assessment was based on the point estimate and its location relative to the threshold for a clinically important effect (when a threshold was available) or to the null. The target of the certainty of evidence assessment was the presence or absence of an important effect, informed by the clinical expert consulted for this review, for responder rate for change in MBL volume, improvement of anemia, improvement of pain, change in UFS-QoL SSS score and HRQoL total score, and the presence or absence of any (non-null) effect for change in BMD at the lumbar spine.

For the GRADE assessments, findings from the LIBERTY 1 trial and the LIBERTY 2 trial were assessed together per outcome because these studies were similar in population, intervention, design, and outcome measures.

The selection of outcomes for GRADE assessment was based on the sponsor’s Summary of Clinical Evidence, consultation with a clinical expert, and the input received from patient groups and public drug plans. The following list of outcomes was finalized in consultation with expert committee members:

• responder rate for change in MBL volume

• improvement in anemia

• improvement in pain

• change in UFS-QoL (SSS score and HRQoL total score)

• change in BMD.

Table 2: Summary of Findings for RGX-E2-NETA Versus Placebo for Patients With HMB Associated With UFs

BMD = bone mineral density; CI = confidence interval; HMB = heavy menstrual bleeding; HRQoL = health-related quality of life; LSM = least squares mean; MBL = menstrual blood loss; NRS = numerical rating scale; RGX-E2-NETA = relugolix-estradiol–norethindrone acetate; SSS = Symptom Severity Scale; UF = uterine fibroid; UFS-QoL = Uterine Fibroid Symptom Quality of Life questionnaire.

Note: Study limitations (which refer to internal validity or risk of bias), inconsistency across studies, indirectness, imprecision of effects, and publication bias were considered when assessing the certainty of the evidence. All serious concerns in these domains that led to the rating down of the level of certainty are documented in the table footnotes.

^aRated down 1 level for study limitations because this outcome uses a subpopulation, and prognostic balance may have been lost. Not rated down for inconsistency because, although the magnitude of the between-group difference differs across the 2 studies, the point estimate of the between-group difference was considered clinically meaningful in each, according to the clinical expert consulted for the review. Rated down 1 level for imprecision because there was no known threshold for a clinically important effect, and the presence of an important effect was informed by the clinical expert consulted for this review. The point estimates in both studies suggest the presence of a clinically important effect; however, the lower bound of the 95% CI for the LIBERTY 1 trial is consistent with an effect that would not be considered clinically important for patients.

^bRated down 1 level for study limitations because this outcome uses a subpopulation, and prognostic balance may have been lost. Not rated down for inconsistency because, although the magnitude of the between-group difference differs across the 2 studies, the point estimate of the between-group difference was considered clinically meaningful in each, according to the clinical expert consulted for the review. Rated down 1 level for imprecision because there was no known threshold for a clinically important effect, and the presence of an important effect was informed by the clinical expert consulted for this review. The point estimates in both studies suggest the presence of a clinically important effect; however, the lower bound of the 95% CIs for the LIBERTY 1 trial and the LIBERTY 2 trial are consistent with an effect that would not be considered clinically important for patients.

^cRated down 1 level for study limitations because data were missing for 24% and 19% of patients in the RGX-E2-NETA and placebo groups, respectively, in the LIBERTY 1 trial, and for 23% and 26% of patients, respectively, in the LIBERTY 2 trial. Not rated down for imprecision because there was no known threshold for a clinically important effect, and the presence of an important effect was informed by the clinical expert consulted for this review. The point estimates in both studies suggest the presence of a clinically important effect, and although the lower bound of the 95% CIs for the LIBERTY 1 trial and the LIBERTY 2 trial are consistent with an effect that would not be considered clinically important for patients, the review team did not rate down because the lower bound of the CIs likely do not considerably cross the minimum threshold for a clinically important effect. Analysis of this outcome was not adjusted for multiplicity. The results are considered as supportive evidence.

^dRated down 1 level for study limitations because data were missing for 24% and 19% of patients in the RGX-E2-NETA and placebo groups, respectively, in the LIBERTY 1 trial, and for 23% and 26% of patients, respectively, in the LIBERTY 2 trial. Not rated down for imprecision because there was no known threshold for a clinically important effect, and the presence of an important effect was informed by the clinical expert consulted for this review; the point estimates and lower bound of the 95% CIs in both studies suggest the presence of a clinically important effect. Analysis of this outcome was not adjusted for multiplicity. The results are considered as supportive evidence.

^eRated down 1 level for study limitations because data were missing for 22% and 20% of patients in the RGX-E2-NETA and placebo groups, respectively, in the LIBERTY 1 trial, and for 25% and 26% of patients, respectively, in the LIBERTY 2 trial. Not rated down for indirectness because the follow-up was limited to 24 weeks, which may be insufficient to detect clinically important BMD loss; however, the clinical expert expects that most patients would not use RGX-E2-NETA for long-term treatment. Rated down 1 level for imprecision because there was no known threshold for a clinically important effect; the point estimate suggested a decrease with RGX-E2-NETA, while the 95% CI included the potential for no difference and an increase. Analysis of this outcome was not adjusted for multiplicity. The results are considered as supportive evidence.

Source: LIBERTY 1 trial and LIBERTY 2 trial clinical study reports and sponsor’s Summary of Clinical Evidence.^12,13,19

Long-Term Extension Studies

Description of Studies

The LIBERTY extension study was a single-arm, phase III, 28-week, open-label extension (OLE) study of patients who had completed either pivotal trial. In the OLE study, the long-term efficacy and safety of RGX-E2-NETA were evaluated for an additional 28 weeks (i.e., after the 24 weeks of treatment during either parent study, the LIBERTY 1 trial or LIBERTY 2 trial).

The LIBERTY randomized withdrawal study (RWS) was a double-blind, phase III, placebo-controlled, 52-week study enrolling those who had completed the LIBERTY extension study and whose disease had responded to treatment at the week 48 visit (i.e., the patient had experienced an MBL volume of less than 80 mL and a 50% or greater reduction from the pivotal study baseline MBL). Patients were rerandomized in a 1-to-1 ratio to receive either oral RGX-E2-NETA or placebo once daily for up to 52 weeks. For patients whose HMB recurred during the LIBERTY RWS (i.e., MBL volume ≥ 80 mL), blinded treatment was stopped, and re-treatment with open-label RGX-E2-NETA was offered.

Efficacy Results

LIBERTY Extension Trial

At week 52 of the LIBERTY extension trial, 87.7% of patients (95% CI, 81.7% to 92.3%) who had received RGX-E2-NETA after the parent study baseline (i.e., the RGX-E2-NETA group) were responders, while 75.6% of patients (95% CI, 68.3% to 82.0%) who had received placebo in the parent studies, but switched to RGX-E2-NETA (i.e., placebo-to-RGX-E2-NETA group) during the OLE study, were responders.

For the anemia outcome (i.e., patients whose hemoglobin level was ≤ 10.5 g/dL at baseline and who experienced an increase of ≥ 2 g/dL), 23 patients of 39 patients (59.0%) in the RGX-E2-NETA group and 16 patients of 38 patients (42.1%) in the placebo-to-RGX-E2-NETA group experienced an improvement in anemia at week 52.

At week 52, the LSMs for the UFS-QoL SSS score were –37.3 points (SE = 2.5 points) in the RGX-E2-NETA group and –35.0 points (SE = 2.5 points) in the placebo-to-RGX-E2-NETA group. The LSMs for the UFS-QoL HRQoL total score were 40.4 points (SE = 2.5 points) in the RGX-E2-NETA group and 39.0 points (SE = 2.6 points) in the placebo-to-RGX-E2-NETA group.

LIBERTY RWS

At week 76, 78.4% of patients (95% CI, 69.3% to 85.1%) in the rerandomized RGX-E2-NETA group and 15.1% of patients (95% CI, 8.9% to 22.8%) in the rerandomized placebo group responded to treatment. The difference in responder rates between treatment groups was 63.4% (95% CI, 52.9% to 73.9%) in favour of RGX-E2-NETA.

At week 104, 69.8% of patients (95% CI, 59.7% to 77.8%) in the rerandomized RGX-E2-NETA group and 11.8% of patients (95% CI, 6.3% to 19.0%) in the rerandomized placebo group responded to treatment. The difference in responder rates between treatment groups was 58.0% (95% CI, 47.0% to 69.1%) in favour of RGX-E2-NETA.

At week 76, for the UFS-QoL SSS score, patients had changes from rerandomization baseline (i.e., week 52 of total study treatment) of 2.2 points (standard deviation [SD] = 18.2 points) in the rerandomized RGX-E2-NETA group and 14.1 points (SD = 21.1 points) in the rerandomized placebo group. For the UFS-QoL HRQoL total score, patients had changes from rerandomization baseline of –0.8 points (SD = 16.8 points) in the rerandomized RGX-E2-NETA group and –9.1 points (SD = 22.4 points) in the rerandomized placebo group.

At week 104, for the UFS-QoL SSS score, patients had changes from rerandomization baseline (i.e., week 52 of total study treatment) of 0.1 points (SD = 15.9 points) in the rerandomized RGX-E2-NETA group and –4.3 points (SD = 42.2 points) in the rerandomized placebo group. For the UFS-QoL HRQoL total score, patients had changes from rerandomization baseline of 1.1 points (SD = 15.3 points) in the rerandomized RGX-E2-NETA group and 8.2 points (SD = 42.1 points) in the rerandomized placebo group.

Harms Results

The LIBERTY Extension Trial

During the OLE study, 54.6% of patients in the RGX-E2-NETA group and 62.8% of patients in the placebo-to-RGX-E2-NETA group experienced at least 1 TEAE. Nasopharyngitis (6.1%) was the most common TEAE in the RGX-E2-NETA group, while hot flush (7.9%), headache (6.7%), and hypertension (6.1%) were the most common TEAEs in the placebo-to-RGX-E2-NETA group.

Overall, 1 patient (0.6%) in the RGX-E2-NETA group and 11 patients (6.7%) in the placebo-to-RGX-E2-NETA group experienced at least 1 SAE. No SAE was reported by more than 2 patients.

Two patients (1.2%) in the RGX-E2-NETA group and 9 patients (5.5%) in the placebo-to-RGX-E2-NETA group stopped treatment due to a TEAE. No TEAE leading to treatment discontinuation was reported by more than 1 patient.

There were no deaths in the OLE study.

At the end of the LIBERTY extension trial (i.e., 52 weeks of study treatment), the LSM percentage changes in BMD at the lumbar spine were –0.80 g/cm² (95% CI, –1.36 g/cm² to –0.25 g/cm²) for the RGX-E2-NETA group and –0.78 g/cm² (95% CI, –1.32 g/cm² to –0.23 g/cm²) for the placebo-to-RGX-E2-NETA group.

LIBERTY RWS

In the LIBERTY RWS, 58.6% of patients in the rerandomized RGX-E2-NETA group and 64.3% of patients in the rerandomized placebo group experienced at least 1 TEAE. Nasopharyngitis (11.2%) was the most common TEAE in the rerandomized RGX-E2-NETA group, while nasopharyngitis (10.7%), hot flush (7.1%), dysmenorrhea (7.1%), hypertension (5.4%), and upper respiratory tract infection (5.4%) were the most common TEAEs in the rerandomized placebo group.

Overall, 2 patients (1.7%) in the rerandomized RGX-E2-NETA group and 2 patients (1.8%) in the rerandomized placebo group experienced at least 1 SAE. No SAE was reported by more than 1 patient.

Two patients (1.7%) in the rerandomized RGX-E2-NETA group and 3 patients (2.7%) in the rerandomized placebo group stopped treatment due to a TEAE. No TEAE leading to treatment discontinuation was reported by more than 1 patient.

There were no deaths in the LIBERTY RWS.

At the end of the LIBERTY RWS (i.e., 104 weeks of study treatment), the LSM percentage changes in BMD at the lumbar spine were 0.81 g/cm² (95% CI, 0.20 g/cm² to 1.41 g/cm²) in the rerandomized RGX-E2-NETA group and 0.10 g/cm² (95% CI, –0.52 g/cm² to 0.71 g/cm²) in the rerandomized placebo group. The LSM percentage between-group difference was 0.71 g/cm² (95% CI, –0.13 g/cm² to 1.55 g/cm²).

Critical Appraisal

There is the possibility of selection bias in the LIBERTY extension study because eligible patients had to have successfully completed either the LIBERTY 1 trial or the LIBERTY 2 trial; thus, patients were more likely to be those who had responded to treatment and did not experience intolerable TEAEs. The single-arm, OLE study was noncomparative, preventing conclusions from being made about the benefits or harms attributable to RGX-E2-NETA versus any comparator. In addition, the open-label design could bias the assessment of subjective, patient-reported outcomes (i.e., UFS-QoL and harms). The lack of data beyond week 52 in the OLE study makes it challenging to draw conclusions on the long-term sustainability of the treatment effect. Treatment in the LIBERTY RWS was double-blinded, which is important, given the subjective nature of the patient-reported outcomes; however, there is the risk of unblinding of the treatment received due to the large difference in efficacy results between the active and placebo groups. In both extension studies, the reasons for missing data for each outcome were not available; as such, the effect of these missing data remains uncertain. In the LIBERTY RWS, there is a risk of bias due to missing outcome data, particularly at longer follow-up times for the UFS-QoL SSS score, HRQoL total score, and BMD.

Because patients who were enrolled in the extension studies must have successfully completed the parent trials, the extension studies have the same generalizability issues. Of note, the clinical expert confirmed that patients receiving RGX-E2-NETA would not be restricted by an upper age limit; rather, they would be restricted by status — that is, being in the premenopausal stage or having recent or planned gynecological surgery or procedures for UFs would not preclude patients from using the drug. The efficacy outcomes, while relevant to patients and practice, are not used in clinical settings (aside from measuring hemoglobin), therefore, the results in real-world clinical settings may differ from those observed in the extension studies.

Indirect Comparisons

The sponsor submitted a systematic review with a network meta-analysis (NMA), a type of indirect treatment comparison (ITC), to estimate the efficacy and safety of RGX-E2-NETA compared with LA (with or without add-back) for the treatment of HMB associated with UFs in people in the premenopausal stage.²⁰

An ITC was required because of a lack of studies directly comparing RGX-E2-NETA with LA (with and without add-back) (the only authorized treatment for UFs in the Canadian setting).

To support the NMA, the sponsor:

• conducted a systematic literature review (SLR) to identify RCTs to estimate the relative efficacy and safety of RGX-E2-NETA versus LA (with or without add-back)

• performed an NMA to assess the relative efficacy and safety of RGX-E2-NETA versus LA (without add-back) for the treatment of HMB associated with UFs in people in the premenopausal stage (a comparison of RGX-E2-NETA to LA with add-back was considered infeasible by the sponsor).

Outcomes of interest for the NMA included effectiveness, TEAEs, and tolerability. The specific effectiveness outcomes were MBL resolution and UFS-QoL (HRQoL total score and SSS score). TEAEs included hot flushes and headaches. Tolerability outcomes were discontinuations due to TEAEs.

Of 228 reports identified in the literature, 31 trials were included in the SLR, and 12 of these were included in the NMA.

Treatments included in the NMA to establish a network connection between RGX-E2-NETA and LA were ulipristal acetate (UPA) 5 mg, UPA 10 mg, and mifepristone (MFP) 10 mg. Although UPA 5 mg, UPA 10 mg, and MFP 10 mg are not considered relevant comparators, these were maintained in the NMA to establish network connections between RGX-E2-NETA and LA (without add-back). These comparators were excluded from the main results despite being retained to establish a network connection between RGX-E2-NETA and LA.

Efficacy Results

MBL Resolution

For the comparison of RGX-E2-NETA versus LA (without add-back), the point estimate for the relative risk (RR) for MBL resolution favoured RGX-E2-NETA; however, the 95% credible interval (CrI) was wide and included the possibility that either RGX-E2-NETA or LA (without add-back) could be favoured ███ █ █████ ███ ████ ████ ██ █████. A sensitivity analysis included a study with a different outcome definition of MBL. Overall, it reached the same conclusion as the base case.

Uterine Fibroid Symptom Quality of Life Questionnaire

Point estimates of the mean differences (MDs) in UFS-QoL HRQoL total scores and UFS-QoL SSS scores favoured LA (without add-back), but were lower than the estimated clinically meaningful threshold of 15 points suggested by the clinical expert. The 95% CrIs crossed the line of equivalence, which suggested that either RGX-E2-NETA or LA (without add-back) could be favoured; however, not all values within the 95% CIs may be clinically important: UFS-QoL total score MD of █████ ██████ ████ ████ █████ ██ ████ ███████ and UFS-QoL symptom severity MD of ████ ██████ ████ ████ █████ ██ █████ ███████.

Two sensitivity analyses were performed for the UFS-QoL HRQoL total score and UFS-QoL SSS score by first excluding 2 studies (with < 10% of patients who were Black), then excluding a study with high heterogeneity. The conclusions drawn from the sensitivity analyses were similar to those drawn from the base-case analysis, favouring LA (without add-back).

Harms Results

Hot Flushes

RGX-E2-NETA was associated with a lower risk of hot flushes compared with LA (without add-back), with an RR of ████ ████ █ ██ █████. A sensitivity analysis was conducted that included studies beyond 1 month after the end of treatment. This sensitivity analysis was consistent with the base case.

Headaches

For the outcome of headaches, the RR estimate was imprecise, suggesting that the direction of the effects are inconclusive. Comparing RGX-E2-NETA and LA (without add-back), the point estimate for the RR for headaches favoured LA (without add-back); however, the 95% CrI was wide and included the possibility that either RGX-E2-NETA or LA (without add-back) could be favoured ███ █ ███████ ███ ████ ████ ██ █████████. A sensitivity analysis conducted for the time point estimate of headaches found the estimate consistent with the base case, with a similarly high level of imprecision.

Discontinuation Due to TEAEs

The point estimate for the RR of the tolerability outcome, discontinuation due to TEAEs, favoured RGX-E2-NETA with a high degree of imprecision, with an RR of ████ ████ ████ ████ ██ █████. The wide 95% CrI that crosses the null also suggests that either RGX-E2-NETA or LA (without add-back) could be favoured. No sensitivity analysis was reported for this outcome.

Critical Appraisal

A systematic review of published literature and grey literature was conducted to identify studies for the NMA. Even though a prespecified protocol for the SLR and NMA could not be obtained, the CDA-AMC review team appraised the submitted SLR using the criteria from A MeaSurement Tool to Assess systematic Reviews 2.²¹ Overall, the searches were comprehensive, and the methods for data extraction and risk of bias appraisal were adequate to reduce the risk of error and bias in these procedures. Although study selection was undertaken by 2 independent reviewers with consensus, the rationale for why certain studies were excluded from the networks was inconsistent. Of the 31 studies identified through the SLR, 12 were retained for the NMAs. Although certain studies of comparators not used in Canada were included in the networks to allow connections between RGX-E2-NETA and LA, other studies with the same comparators were excluded (e.g., Bagaria [2009], comparing placebo and MFP 10 mg, was included to connect the network, whereas other studies, including a study with an MFP 10 mg treatment group, were excluded because the treatments were not of interest). The selective inclusion of studies in the network may result in risk of bias due to missing results in the syntheses, the magnitude and direction of which cannot be predicted.

A Bayesian NMA included some outcomes (i.e., effectiveness, safety, and tolerability), but not all. Outcomes deemed relevant by patients and clinicians that were not included in the constructed networks were the efficacy outcomes of improvement in anemia and pain and the harms outcomes of BMD. BMD as a safety outcome was not included due to lack of comparative data. As such, this NMA cannot inform on the comparative effects of RGX-E2-NETA and LA (without add-back) for these outcomes.

There are limitations to both the fixed-effects and random-effects models used in this NMA. The models were selected based on better model fit (i.e., smaller deviance information criterion). Given that most outcomes were modelled with fixed effects, we expect that the true magnitude of uncertainty was underestimated. Because the effect estimates for these end points were already affected by important imprecision, reliance on the results of the unselected models would not have changed the review team’s conclusions about the estimated effects.

A key limitation of this NMA is the fact that the clinically relevant comparator, LA, did not contain add-back. While hormonal add-back is a part of RGX-E2-NETA, according to the sponsor, LA with add-back was excluded from the NMA due to feasibility reasons. This means that RGX-E2-NETA compared with LA (without add-back) does not reflect how the actual comparator is prescribed in Canada.

The sponsor found that RGX-E2-NETA had a reduced RR of hot flushes compared with LA. All other efficacy, safety, and tolerability outcomes crossed the null, with wide CrIs. These suggest that either RGX-E2-NETA or LA (without add-back) could be favoured.

There were other notable limitations of the overall NMA, given that the assumptions of exchangeability, similarity of outcomes, and consistency (coherence) could not be satisfied.

Overall, due to the absence of a comparison of RGX-E2-NETA versus LA (with add-back), the high degree of imprecision represented by RR or MD CrIs that can favour either RGX-E2-NETA or LA (without add-back), and the inability to satisfy the key assumptions required for an NMA (i.e., exchangeability, homogeneity, and consistency), definitive conclusions about the relative treatment effects of RGX-E2-NETA versus LA (with add-back) cannot be drawn based on this NMA.

Studies Addressing Gaps in the Evidence From the Systematic Review

There were no studies addressing gaps in the systematic review evidence submitted by the sponsor.

Conclusions

Evidence from 2 phase III, double-blind RCTs (the LIBERTY 1 trial and the LIBERTY 2 trial) suggests that patients aged 18 to 50 years with confirmed UFs and HMB (defined as an MBL volume of 160 mL or greater in 1 cycle or 80 mL or greater per cycle for 2 cycles) who received RGX-E2-NETA once a day for 24 weeks were more likely to experience treatment response for HMB compared to those who received placebo. Patients receiving RGX-E2-NETA may also have clinically meaningful improvements in anemia, UF-associated pain, symptom severity, and HRQoL (as measured using the UFS-QoL). However, the evidence for these outcomes was uncertain due to an increased risk of bias, lack of known thresholds for a clinically meaningful effect, and missing data. Longer-term evidence suggested that treatment effects were maintained for reduction in MBL and the patient-reported outcomes of symptom severity and HRQoL (with up to 104 weeks of study treatment), as well as for improvements in anemia (with up to 52 weeks of study treatment). Limitations with the study designs and numerous discontinuations increase the uncertainty of the longer-term results. It is also likely that patients who experienced benefits without intolerable harms remained in the studies longer. Results from the ITC of RGX-E2-NETA versus LA (without add-back) indicated that either treatment could be favoured for efficacy outcomes. Additionally, LA with add-back was not included in the ITC, and definitive conclusions about its treatment effect relative to that of RGX-E2-NETA could not be drawn. The incidence and types of harms were generally balanced and similar between the treatment groups in the pivotal studies and across the studies. Longer-term evidence did not indicate new safety signals; however, expert opinion noted that assessment of long-term bone health is important to inform decisions regarding prolonged treatment with RGX-E2-NETA. The Health Canada product monograph notes that use of RGX-E2-NETA should be limited to 24 months due to the risk of bone loss; expert opinion indicated that use in clinical practice would be short-term.

Introduction

The objective of this report is to review and critically appraise the evidence submitted by the sponsor on the beneficial and harmful effects of oral RGX-E2-NETA (40 mg RGX, 1 mg E2, and 0.5 mg NETA) for the treatment of HMB associated with UFs in patients in the premenopausal stage.

Disease Background

The contents in this section have been informed by materials submitted by the sponsor and by clinical expert input. The following information has been summarized and validated by the review team.

UFs, also known as myomas or leiomyomas, are benign neoplasms of uterine smooth muscle and are the most common tumours in females.¹ Patients with UFs can present with 1 predominant fibroid or with clusters of many fibroids.²² The exact etiology of UFs is unclear, but estrogen and progesterone play an important role in UF initiation and growth.¹ UFs are typically diagnosed in people between menarche and menopause. The main risk factors are younger age at menarche, late onset of menopause, African ancestry, and a child-bearing history of few or no children.¹

A large proportion of patients with UFs are asymptomatic.⁵ A common symptom is AUB, particularly HMB, which is reported by 50% to 80% of patients with symptomatic UFs in surveys and trials.^2,3 Infertility, abdominal bloating, constipation, increased urinary frequency, incontinence, dyspareunia, anemia, and fatigue are among other symptoms, all of which can affect patients’ HRQoL.¹ It has been reported that the risk of poor mental health outcomes is higher among patients with diagnosed UFs, particularly those who experience pain-related symptoms or undergo hysterectomy, compared to those without diagnosed UF.²³ For research purposes, HMB is defined as more than 80 mL of blood loss during each menstrual cycle and based on direct measurement of MBL using the alkaline hematin method.⁴ However, from a clinical practice point of view, it is defined as a volume that interferes with the patient's quality of life, including physical, social, and emotional aspects. Self-reported HMB can be inaccurate because it is based on the patient's perception of volume. It should be noted that some patients think their heavy volume is “normal.”²⁴

Prevalence estimates of UFs vary widely depending on the study methods applied and the characteristics of the populations assessed. A large, international, internet-based survey was conducted by Zimmermann et al. (2009) in which 21,746 females across 8 countries participated and provided patient-reported UF prevalence data. It included 2,500 females aged 15 to 49 years from Canada, among whom the self-reported prevalence of diagnosed UFs was 5.5%.⁵ An online survey of people in Canada published by Laberge et al. (2016) that included 9,413 females aged 20 to 49 years found a self-reported, physician-diagnosed UF prevalence of 4.1%. The authors noted that scores for all HRQoL domains were lower among females with moderate or severe UF symptoms than in those with mild UF symptoms. Additionally, the authors identified responses that were consistent with moderate or severe UFs in females who made no report of physician-diagnosed UF; some of these females may be experiencing substantial burden due to undiagnosed UFs.²⁵

Diagnosis of UFs is based on a pelvic examination and the finding of an enlarged, mobile uterus with an irregular contour. Diagnosis can be confirmed through pelvic ultrasonography. Ultrasonography is the most widely used diagnostic tool to confirm UFs due to its availability, ease of use, utility in visualizing genital tract structures, and cost-effectiveness.¹⁵

Standards of Therapy

The contents in this section have been informed by materials submitted by the sponsor and by clinical expert input. The following information has been summarized and validated by the review team.

According to the literature and the clinical expert consulted on this review, the overall goals of treatment are to reduce menstrual bleeding, allow for activities of daily living, increase hemoglobin levels, address pain and bulk symptoms, and improve HRQoL.^6,7 Other reasons noted by the clinical expert for using treatment to reduce the size of UFs include enabling less invasive surgery approaches, delaying surgery, reducing the need for surgery until menopause (when symptoms tend to subside), and reducing the impact of UFs on fertility. Treatment can include surgical, procedural, and pharmacological options. As per the expert, the choice of treatment depends on the clinical presentation (i.e., symptoms, severity, and size and location of UFs), the patient’s age, and the patient’s goals (e.g., desire to preserve fertility and/or the uterus).

For patients who do not desire fertility, hysteroscopic resection or myomectomy of submucosal UFs may be options. Each has benefits and drawbacks.⁶ Patients with UFs that are not submucosal are offered treatments to improve bleeding, but these do not reduce the size of the UFs. First-line pharmacological options include combined estrogen-progestin contraceptives, progestin-releasing intrauterine systems, and tranexamic acid (for patients who wish to avoid hormonal contraceptives or to treat only when symptoms occur). However, the use of hormonal contraceptives for treating UFs is off-label, and HMB is often inadequately controlled with hormonal contraceptives alone.⁸ For patients whose disease does not respond to first-line treatments, GոRΗ antagonists and agonists may be used with or without add-back therapies (i.e., E2 and NETA), which are used to mitigate hypoestrogenic side effects and protect the endometrium.⁹ Some medications, such as RGX-E2-NETA, contain add-back therapy in the combination. GnRH agonists (e.g., LA) are often used as short-term, preoperative therapies (i.e., for 3 months to 6 months) with concomitant iron therapy to reduce the size of UFs and improve anemia before surgery.¹⁰ Some caveats with the use of GnRH agonists are the initial flare effect (i.e., symptom exacerbation), which is not seen with GոRΗ antagonists, and the fact that once the drug is stopped, UF regrowth often occurs within 3 months.⁷ Uterine artery embolization can treat HMB and bulk symptoms, but does not remove UFs and is associated with ovarian dysfunction and the need for reintervention (rates between 14% and 38% within 5 years).¹¹ Progesterone receptor modulators are generally not available in most countries for treating fibroids, and use is restricted due to increased risk of hepatic toxicity. Hysterectomy is the only definitive treatment for UFs, but it is associated with significant morbidity and mortality; therefore, it is a last-line option for patients who have persistent symptoms despite other treatments.⁷

For patients who desire fertility with submucosal UFs, hysteroscopic myomectomy is minimally invasive and can relieve symptoms. Myomectomy allows for the uterus to be retained, but is associated with a higher risk of blood loss than hysterectomy.⁷ In addition, UFs have a 15% recurrence rate after myomectomy; risk varies with age, number of fibroids, uterine size, and childbirth after the procedure.⁷ Patients with UFs that are not submucosal may be offered laparoscopic or open abdominal myomectomy. Drug treatments are associated with adverse effects when used for long durations, and symptoms quickly return when treatment is discontinued.

Before September 2020, UPA was a hormonal drug indicated for the treatment of UFs. It was associated with morphological changes in the endometrium and an increased risk of rare but serious liver injury, and it was subsequently withdrawn from the Canadian market.^26-28 With UPA no longer available, physicians continue to treat UFs with tranexamic acid, GոRΗ agonists and antagonists (with add-back therapy), hormonal contraceptives, medical procedures, and surgery, as previously described.⁹

Drug Under Review

Key characteristics of RGX-E2-NETA are summarized in Table 3 along with those of other drug treatments available for patients with HMB associated with UFs who wish to maintain fertility, preserve the uterus, and reduce symptoms while avoiding surgery or other procedures, as suggested by the sponsor.

Myfembree is a combination of RGX, E2, and NETA. RGX is a nonpeptide, GnRH receptor antagonist that binds competitively to GnRH receptors in the anterior pituitary gland, reducing the secretion of luteinizing hormone and follicle-stimulating hormone.¹⁶ This reduction of follicle-stimulating and luteinizing hormone concentrations limits the production of estrogen and progesterone, respectively, reducing the bleeding associated with UFs and the pain associated with endometriosis. E2 is an agonist of nuclear estrogen receptor subtypes with organ-specific effects. Exogenously administered E2 in RGX-E2-NETA may alleviate symptoms, such as BMD loss and vasomotor symptoms, that can occur due to a decrease in circulating estrogen concentrations from RGX alone. NETA is a synthetic progestin that acts as an agonist of progesterone receptors and reduces the estrogen-induced risk of endometrial hyperplasia.

The recommended dose is 1 tablet of RGX-E2-NETA to be taken orally once daily.¹⁶ It is recommended that the administration of RGX-E2-NETA be initiated as early as possible after the onset of menses, but no later than 5 days after menses have started. The Health Canada product monograph states that use of RGX-E2-NETA should be limited to 24 months due to the risk of continued, irreversible bone loss.

RGX-E2-NETA is indicated in women in the premenopausal stage for the management of HMB associated with UFs and for the management of moderate to severe pain associated with endometriosis.¹⁶ RGX and E2 have been reviewed by CDA-AMC as individual products and in combination, such as E2 with progesterone. However, the combination RGX-E2-NETA has not been reviewed before.

The sponsor’s reimbursement request is for the management of HMB associated with UFs in women in the premenopausal stage.

Table 3: Key Characteristics of RGX-E2-NETA and Leuprolide Acetate

BMD = bone mineral density; E2 = estradiol; FSH = follicle-stimulating hormone; GnRH = gonadotropin-releasing hormone; IM = intramuscular; LH = luteinizing hormone; NETA = norethindrone acetate; RGX = relugolix; RGX-E2-NETA = relugolix-estradiol–norethindrone acetate; UF = uterine fibroid.

^aHealth Canada–approved indication.

Source: Product monographs for Myfembree and Lupron Depot.^16,29

Perspectives of Patients, Clinicians, and Drug Programs

The full patient and clinician group submissions received are available in the consolidated patient and clinician group input document for this review on the project website.

Patient Group Input

This section was prepared by the review team based on the input provided by patient groups.

CDA-AMC received 1 patient group input submission from CANFib. The organization provides a forum for females in Canada with UFs to discuss their symptoms, talk to peers, find out more about UFs, and discuss navigation of the Canadian medical system while learning more about the treatments and medications approved by Health Canada for UFs. CANFib also works with Canadian pharmaceutical outreach, researchers, insurance companies, and specialists to push for new ideas that become future treatments and solutions.

CANFib conducted a survey in 2023. Respondents included 315 females with UFs. The patient group noted that 59% of the respondents had to wait from 7 months to more than 13 months to visit their health care provider. Based on the survey results, a new symptom or a current medication that was not working were the top reasons to return to the specialist. Additionally, pain was the top reason (n = 266) for seeking medical help, followed by excessive bleeding (n = 249); these symptoms led 73% of respondents to take at least 2 days off work per month. CANFib explained that 261 respondents reported consulting the internet for information on UF relief; that 84% of those who consulted medical professionals felt they left the consultation still “uninformed” about the condition; and that 86% felt they were not given the full list of Health Canada–approved treatments and medications. CANFib highlighted that because the top reason to return to a specialist was ineffectiveness of medications — and given the long wait times from diagnosis to treatment and for appointments — more options and earlier treatment will benefit patients and reduce the need for follow-up visits and surgery.

CANFib also conducted group discussions with 32 females with UFs who had experience with RGX-E2-NETA. Among these participants, 31 reported being “very happy” about the effects of RGX-E2-NETA on their symptoms, while 1 reported feeling “indifferent.” Two had been able to work again. Two out of the 32 respondents had access to RGX-E2-NETA through extended insurance coverage, while 30 respondents had to pay out of pocket for the medication; 3 respondents said they could or would continue to afford the medication, even for a lengthy period. CANFib added that RGX-E2-NETA enabled the females in the discussion group to lead full lives, spend time with their families, care for children and/or parents, show up at work, and function without debilitating pain, heavy blood loss, or ongoing depression.

CANFib explained that, based on member discussions, most females reported primary symptoms of heavy bleeding and pain; other common events are lost time at work, inability to function within a family unit, a diminished social or professional life, and eventual depression. The discussion group described the constant need for open access to a bathroom at work or at home. Members described how their relationships suffered from the immense pressure of feeling as though they must carry on as if they do not have symptoms.

Regarding currently available treatments, CANFib noted that although there are treatments and medications available to treat UFs, there is no cure other than surgical hysterectomy. According to the organization, desirable outcomes include access to an effective treatment that does not require a 3-month dosage commitment, regaining confidence, and being able to enjoy family, social, and professional lives.

This patient group submission also included comments from the incoming Society of Obstetricians and Gynecologists of Canada president and an endorsement from the Women’s Health Coalition.

Clinician Input

Input From the Clinical Expert Consulted for this Review

All CDA-AMC review teams include at least 1 clinical specialist with expertise regarding the diagnosis and management of the condition for which the drug is indicated. Clinical experts are a critical part of the review team and are involved in all phases of the review process (e.g., providing guidance on the development of the review protocol, assisting in the critical appraisal of clinical evidence, interpreting the clinical relevance of the results, and providing guidance on the potential place in therapy). The following input was provided by 1 clinical specialist with expertise in the diagnosis and management of HMB associated with UFs.

Unmet Needs

According to the clinical expert consulted on this review, patients seek new therapies that can adequately and rapidly address symptoms without adverse effects, improve HRQoL, offer simpler administration, with quick resolution upon discontinuation.

Place in Therapy

Based on input from the clinical expert, first-line treatments for AUB can include nonsteroidal anti-inflammatory drugs, hormonal contraceptives (including progestins and levonorgestrel intrauterine systems), tranexamic acid, and iron repletion as further investigations take place to identify the cause of the AUB. Once a diagnosis of UFs (with HMB) is confirmed, RGX-E2-NETA would be a treatment option in the absence of endometrial pathology, depending on the symptoms, severity, and location of the UFs and on the patient’s goals. The clinical expert expects that RGX-E2-NETA will be used in practice to complement and facilitate less invasive medical procedures and surgical approaches, such as hysteroscopic procedures and fertility-preserving surgical approaches. As per the expert, RGX-E2-NETA may be used as a bridge treatment in patients who plan to have gynecological surgery within 3 months (to elevate hemoglobin and enable less invasive surgical methods), want to delay surgery for fertility, or anticipate menopause within a couple of years.

Patient Population

The clinical expert indicated that patients receiving RGX-E2-NETA would be adults in the premenopausal stage, given that the symptoms of UFs tend to diminish after menopause. Additionally, the patients most likely to respond to treatment with RGX-E2-NETA include those with HMB associated with UFs and bulk symptomatology secondary to UFs. The expert also noted that patients with anemia secondary to severe AUB, severe pain, and ureteric obstruction by UFs are most in need of treatment.

The patients best suited for treatment with RGX-E2-NETA would be identified based on clinical history, physical examination, and investigations (e.g., Papanicolaou test, biopsy, complete blood count, pregnancy testing, and sonohysterography hysteroscopy). As part of confirming a diagnosis, endometrial and cervical pathology must be ruled out based on the patient’s age and risk factors.

Assessing the Response to Treatment

As per the clinical expert, the main outcomes used to assess treatment response in practice include qualitative clinical assessments for a reduction in primary symptomatology (i.e., reduced HMB), reduction in pain and bulk symptoms, return to normal function (e.g., activities of daily living, work), and fertility success, if this is the patient’s goal. Clinicians also focus on the patient’s satisfaction with the treatment rather than on quantitative measures, such as MBL volume. However, physicians may measure hemoglobin levels with the goal of increasing them to normal values in patients experiencing anemia. The expert explained that although the outcomes used in clinical trials may be informative, these are generally not used in practice (aside from measuring and improving hemoglobin levels).

Discontinuing Treatment

The clinical expert stated that RGX-E2-NETA should be discontinued upon successful surgery or procedure, plans for pregnancy, menopause, a meaningful decline in BMD, or intolerable adverse effects. Although the Health Canada product monograph states that use of RGX-E2-NETA should be limited to 24 months due to the risk of continued, irreversible bone loss, the expert suggested that treatment beyond 2 years may be considered if the patient is maintaining their goals for symptom improvement and has a healthy BMD. However, if RGX-E2-NETA is used to treat HMB and anemia in anticipation of a planned surgery or medical procedure, the treatment duration tends to be less than 2 years. As noted by the expert, the benefit in preparation for surgery with respect to the reduction of the size of the uterus and fibroids is achieved after 3 months. If wait times are longer than 3 months, benefit can be maintained by longer use of the treatment.

Prescribing Considerations

RGX-E2-NETA would be used in a community setting, acute care hospital, or specialty clinic. While diagnosis and initial prescription of the drug would be performed by an obstetrician-gynecologist, the expert noted that follow-up management and monitoring may be done by a primary care physician.

Clinician Group Input

No input was received from clinician groups for this review.

Drug Program Input

The drug programs provide input on each drug being reviewed through the reimbursement review processes by identifying issues that may affect the programs’ ability to implement a recommendation. The implementation questions and corresponding responses from the clinical expert consulted by for this review are summarized in Table 4.

Table 4: Summary of Drug Plan Input and Clinical Expert Responses

AUB = abnormal uterine bleeding; BIA = budget impact analysis; CDEC = Canadian Drug Expert Committee; E2 = estradiol; E2-NETA = estradiol and norethindrone acetate; GnRH = gonadotropin-releasing hormone; HMB = heavy menstrual bleeding; HRQoL = health-related quality of life; ITC = indirect treatment comparison; LHRH = luteinizing hormone-releasing hormone; MBL = menstrual blood loss; NETA = norethindrone acetate; RGX = relugolix; RGX-E2-NETA = relugolix-estradiol–norethindrone acetate; UF = uterine fibroid.

Clinical Evidence

The objective of this Clinical Review Report is to review and critically appraise the clinical evidence submitted by the sponsor on the beneficial and harmful effects of oral RGX-E2-NETA in the treatment of HMB associated with UFs in patients in the premenopausal stage. The focus will be placed on comparing RGX-E2-NETA to relevant comparators and identifying gaps in the current evidence.

A summary of the clinical evidence included by the sponsor in the review of RGX-E2-NETA is presented in 4 sections, with our critical appraisal of the evidence included at the end of each. The first section, the systematic review, includes pivotal studies and RCTs that were selected according to the sponsor’s systematic review protocol. Our assessment of the certainty of the evidence in this first section using the GRADE approach follows the critical appraisal of the evidence. The second section includes sponsor-submitted long-term extension studies. The third section includes indirect evidence from the sponsor. The fourth section is meant to include additional studies considered by the sponsor to address important gaps in the systematic review evidence; however, there were no such studies.

Included Studies

Clinical evidence from the following is included in the review and appraised in this document:

• 2 pivotal RCTs identified in the systematic review

• 2 extension studies

• 1 ITC.

Systematic Review

The contents in this section have been informed by materials submitted by the sponsor. The following information has been summarized and validated by the review team.

Description of Studies

Two phase III, double-blind, placebo-controlled RCTs were included in the systematic review: the LIBERTY 1 trial (N = 388) and the LIBERTY 2 trial (N = 382).^12,13 Characteristics of the included studies are summarized in Table 5. The studies were mostly identical in design and were conducted to evaluate the efficacy and safety of RGX-E2-NETA in patients in the premenopausal stage for the management of HMB associated with UFs. The main differences between the studies were the countries where the studies took place and the order of statistical testing of secondary outcomes. The studies started with a screening period of up to 13 weeks to confirm eligibility through diagnosis of UFs, validation of HMB, assessment of BMD by dual-energy X-ray absorptiometry (DXA), and endometrial biopsies. Patients were randomized to 24 weeks of treatment, after which they could enrol directly into the 28-week OLE study followed by the 52-week RWS; those who did not were followed for safety monitoring for 30 days.^31,32 Although each study had 3 treatment groups, only the RGX-E2-NETA groups (n = 128 in the LIBERTY 1 trial, n = 126 in the LIBERTY 2 trial) and placebo groups (n = 128 in the LIBERTY 1 trial, n = 129 in the LIBERTY 2 trial) are described in this report. The group receiving RGX plus delayed estradiol and norethindrone acetate (E2-NETA) was included in the LIBERTY trials to provide an assessment of the requirement for E2 and NETA to mitigate the adverse effects of RGX monotherapy on BMD loss and vasomotor symptoms. Results are not reported herein because RGX monotherapy is not available in Canada.

Table 5: Details of Studies Included in the Systematic Review

BMD = bone mineral density; BPD = bleeding and pelvic discomfort; DXA = dual-energy X-ray absorptiometry; E2 = estradiol; E2-NETA = estradiol and norethindrone acetate; HMB = heavy menstrual bleeding; HRQoL = health-related quality of life; MBL = menstrual blood loss; NETA = norethindrone acetate; NRS = numerical rating scale; OLE = open-label extension; RCT = randomized controlled trial; RGX = relugolix; RGX-E2-NETA = relugolix-estradiol–norethindrone acetate; SSS = Symptom Severity Scale; TEAE = treatment-emergent adverse event; UF = uterine fibroid; UFS-QoL = Uterine Fibroid Symptom Quality of Life questionnaire; vs. = versus.

^aMain secondary end points included key secondary end points and those deemed by the sponsor to be most relevant to the reimbursement review.

Source: LIBERTY 1 trial and LIBERTY 2 trial clinical study reports and sponsor’s Summary of Clinical Evidence.^12,13,19 Details included in the table are from the sponsor’s Summary of Clinical Evidence.

Populations

Inclusion and Exclusion Criteria

Eligibility criteria were consistent between the studies. Eligible patients included females in the premenopausal stage aged 18 to 50 years with regularly occurring menstrual periods of 14 or fewer days’ duration with a cycle of 21 days to 38 days. Patients had to have had a confirmed UF diagnosis by transvaginal ultrasound, HMB associated with UFs (defined as MBL of ≥ 160 mL during 1 menstrual cycle or ≥ 80 mL per cycle during 2 cycles), and no significant endometrial pathology on biopsy. Individuals were not eligible if they had ultrasound readings demonstrating other pathology that could be responsible for or contributing to HMB; if they had known, rapidly enlarging UFs; if they had undergone gynecological surgery or ablation procedures for UFs within the 6 months before screening; if they had a history of or current osteoporosis, other metabolic bone disease, hyperparathyroidism, hyperprolactinemia, hyperthyroidism, anorexia nervosa, or low traumatic or atraumatic fracture; or if they had a history of medication use to treat BMD loss (other than calcium and vitamin D preparations).

Interventions

Patients were randomized 1 to 1 to 1 to receive oral RGX-E2-NETA (40 mg RGX, 1 mg E2, and 0.5 mg NETA), RGX plus delayed E2 and NETA, or matching placebo once daily in a fasted state for 24 weeks. Randomization was stratified by geographic region (North America versus rest of the world) and by mean screening MBL using the alkaline hematin method (< 225 mL versus ≥ 225 mL).

In the LIBERTY studies, patients received RGX with commercially available E2-NETA in a copackaged configuration. The Health Canada–approved, commercialized formulation of RGX-E2-NETA is a fixed-dose combination tablet consisting of all 3 drugs. According to the sponsor, a bioequivalence study was conducted that demonstrated bioequivalence between the fixed-dose combination tablet and the coadministration of a RGX 40 mg tablet and an E2-NETA tablet containing 1 mg of E2 and 0.5 mg of NETA.³⁷

Supplementation with calcium and/or vitamin D was permitted in the studies, but was not required. Patients with a hemoglobin level of less than 8.0 g/dL were excluded from the studies. During screening, patients with microcytic iron deficiency anemia (defined as a hemoglobin level ≥ 8 g/dL and ≤ 10 g/dL), a mean corpuscular volume below the lower limit of normal, and low serum iron and ferritin were started on iron therapy, which was continued during the studies. Patients who entered the screening period on iron therapy could continue it during the studies. Patients who developed new microcytic iron deficiency anemia during the studies (defined as a hemoglobin level ≤ 10 g/dL), a mean corpuscular volume below the lower limit of normal, and low serum iron and ferritin were started on iron therapy. If the hemoglobin level was less than 10 g/dL and the mean corpuscular volume was below the lower limit of normal, a ferritin and iron level was reported through the central lab.

Prohibited medications included drugs affecting the hypothalamic-pituitary-gonadal axis (e.g., GnRH analogues, estrogens, and progestins), drugs to treat BMD (e.g., bisphosphonates, specified antiepileptic drugs, glucocorticoids), drugs to treat blood loss (e.g., anticoagulants), and drugs with the potential to affect RGX absorption (e.g., P-glycoprotein inducers and moderate and strong P-glycoprotein inhibitors).

Outcomes

A list of efficacy end points assessed in this Clinical Review Report is provided in Table 6, followed by descriptions of the outcome measures in Table 7. Summarized end points are based on outcomes included in the sponsor’s Summary of Clinical Evidence as well as any outcomes identified as important to this review, according to the clinical expert consulted for this review and input from patient groups and public drug plans. Using the same considerations, we selected end points that were considered to be most relevant to inform expert committee deliberations and finalized this list of end points in consultation with members of the expert committee. All summarized efficacy end points were assessed using GRADE. Select notable harms outcomes considered important for informing the expert committee deliberations were also assessed using GRADE.

The following considerations went into the selection of the outcomes that were summarized in the report and assessed using GRADE:

• improvement of symptoms (e.g., MBL, anemia, pain) and HRQoL, which were important to the patient group that provided input and to the clinical expert consulted on this review

• measurement of MBL using the alkaline hematin method, which has been used as a primary end point in clinical trials assessing HMB.³⁸

Table 6: Outcomes Summarized From the LIBERTY 1 and LIBERTY 2 Trials

BMD = bone mineral density; DXA = dual-energy X-ray absorptiometry; HRQoL = health-related quality of life; MBL = menstrual blood loss; NRS = numerical rating scale;RGX-E2-NETA = relugolix-estradiol–norethindrone acetate; SSS = Symptom Severity Scale; UF = uterine fibroid; UFS-QoL = Uterine Fibroid Symptom Quality of Life questionnaire; vs. = versus.

^aStatistical testing for these end points was adjusted for multiple comparisons (e.g., hierarchal testing).

Source: Sponsor’s Summary of Clinical Evidence.¹⁹ Details included in the table are from the sponsor’s Summary of Clinical Evidence.

Responder Rate and Change in MBL Volume

Responder rate was defined as the percentage of participants in the RGX-E2-NETA group versus the placebo group who experienced an MBL volume of less than 80 mL and a 50% or greater reduction from baseline in MBL volume over the last 35 days of treatment as measured using the alkaline hematin method. The effects of RGX-E2-NETA and placebo on reducing HMB associated with UFs were also measured by the percentage change from baseline to week 24 in MBL volume.

Baseline MBL was defined as the average MBL during the 2 screening menstrual cycles used to assess the inclusion criteria before the date of the first dose of the study drug. The MBL during the final study month was defined as the total MBL during the last 35 days on treatment. To ensure collection of all feminine products used during that menstrual cycle, an interval of up to 35 days for measurement of the primary end point was selected to accommodate patients who continued to have cyclic bleeding on study treatment and whose natural cycles were at the upper end of the normal cycle duration range.

MBL volume was analyzed using the alkaline hematin method, which is the gold standard for objective measurement of blood loss using feminine sanitary products. It has been used as a primary end point in clinical trials assessing HMB.^4,39 Used feminine products are pummelled in a 5% sodium hydroxide solution to convert hemoglobin to alkaline hematin, and the resulting hematin absorbance is compared against calibration curves coming from a diluted venous blood sample from the same patient.³⁹

The upper limit of normal MBL has been suggested to be 80 mL.⁴⁰ Based on the literature, a clinically meaningful change in MBL has been estimated to be a 22% reduction for patients with HMB, as measured using the alkaline hematin method.⁴¹ This estimate was calculated based on a receiver-operating characteristic curve using data from a multicentre, double-blind, placebo-controlled RCT of a medication for HMB that included some patients who had UFs.⁴¹

Improvement in Anemia

A baseline hemoglobin concentration of 10.5 g/dL or lower is below the lower limit of normal for females (i.e., 12 g/dL).⁴² An increase of 2 g/dL or greater is expected following the transfusion of 2 units of red blood cells; the transfusion is expected to result in a clinically meaningful improvement in hemoglobin concentration.^43,44 The proportion of patients who had an increase in hemoglobin level of at least 2 g/dL at week 24 was reported for the subset of patients who had a baseline hemoglobin value of 10.5 g/dL or lower. Clinical laboratory testing (including for ferritin and iron) was conducted at every visit (i.e., every 4 weeks).

NRS for UF-Associated Pain

Daily e-diary entries included assessments of UF-associated pain during the preceding 24 hours using an 11-point NRS (ranging from 0 to 10, where higher scores indicated worse pain). The maximum NRS score for pain at week 24 or end-of-treatment was calculated as the maximum NRS score during the last 35 days on study treatment. A minimal important difference (MID) for the NRS was not identified in the literature.

Uterine Fibroid Symptom Quality of Life Questionnaire

The UFS-QoL is a disease-specific instrument used to assess symptom severity and HRQoL in patients with UFs during the preceding month.¹⁴ The SSS consists of 8 items rated on a 5-point Likert scale ranging from “not at all” to “a very great deal.” The HRQoL component of the instrument includes 29 items rated on a 5-point scale ranging from “none of the time” to “all of the time.” Higher SSS scores indicate more severe UF symptoms, while higher HRQoL total scores indicate better quality of life. Scores for the SSS and HRQoL subscales are linearly transformed to a scale ranging from 0 points to 100 points:

• Symptom severity raw scores were transformed using the following formula: actual raw score minus lowest possible raw score, divided by total possible raw score range, multiplied by 100.

• Raw scores for scales other than symptom severity were transformed using the following formula: highest possible score minus actual raw score, divided by total possible raw score range, multiplied by 100.

The sponsor did not identify any literature-based MID estimate for this end point; however, a 20-point or greater decrease in the SSS has been associated with patients feeling “much better.”⁴⁵

Patients completed the UFS-QoL at their baseline, week 12, and week 24 visits.

Safety

Safety outcomes included the proportion of patients who reported TEAEs (graded according to the Common Terminology Criteria for TEAEs), SAEs, withdrawals from treatment due to TEAEs, and deaths. TEAEs were collected from when the first dose of the study drug was administered until the follow-up visit approximately 30 days after the last dose of study drug (or upon initiation of another investigational drug or hormonal therapy affecting the hypothalamic-pituitary-gonadal axis or surgical intervention for UFs, whichever occurred first). Based on clinical expert opinion, loss of BMD at 24 months was deemed relevant to the current review.

Bone Mineral Density

The percentage change from baseline to week 24 in BMD at the lumbar spine (i.e., L1 to L4) was assessed using DXA. RGX monotherapy suppresses sex hormones through GnRH receptor antagonism, resulting in low estrogen levels, leading to declines in BMD — an important determinant for fracture risk.⁴⁶ To assess the effect of adding E2-NETA to RGX, BMD was assessed throughout the LIBERTY clinical program.

Patients were assessed using the same DXA apparatus at each site with the same scan mode. The central radiology laboratory collected and evaluated the scans for acceptability and assessed BMD as per the imaging charter. DXA scans were completed at the screening, week 12, and week 24 visits.

Table 7: Summary of Outcome Measures and Their Measurement Properties

HMB = heavy menstrual bleeding; HRQoL = health-related quality of life; MBL = menstrual blood loss; MID = minimal important difference; SD = standard deviation; SF-36 = Short Form (36) Health Survey; SSS = Symptom Severity Scale; UF = uterine fibroid; UFS-QoL = Uterine Fibroid Symptom Quality of Life questionnaire; vs. = versus.

Source: Request to sponsor for additional information.⁴⁸

Statistical Analysis

Statistical analyses are summarized in Table 8.

Sample Size and Power Calculation

Assuming a 20% dropout rate, approximately 130 patients in the RGX-E2-NETA group and 130 patients in the placebo group would provide at least 99% power at a 2-sided 0.05 significance level to detect a 30% difference in responder rates between the RGX-E2-NETA group and the placebo group for the primary end points for each of the LIBERTY 1 trial and the LIBERTY 2 trial.⁴⁹ The assumed responder rate of 25% for the placebo group was within the range of the responder rates observed from similar phase III trials involving UFs.⁵⁰ The sample size and power calculations were based on a chi-square test.

Statistical Testing

The primary efficacy end point (i.e., the proportion of patients in the RGX-E2-NETA group versus placebo group who experienced an MBL volume of less than 80 mL and a 50% or greater reduction from baseline MBL volume over the preceding 35 days of treatment, as measured using the alkaline hematin method) was analyzed using a Cochran-Mantel-Haenszel test statistic for proportions and was stratified by the baseline mean MBL volume using the alkaline hematin method and geographic region. The between-group difference in responder rate, its 2-sided 95% CI, and P value were estimated using stratum-adjusted Cochran-Mantel-Haenszel proportions. Details on the statistical testing of secondary outcomes are summarized in Table 8.

Multiple Testing Procedure

A gatekeeping, mixed-sequence testing procedure was applied to maintain the family-wise type I error rate for the primary end point and 7 key secondary end points in the LIBERTY 1 trial and LIBERTY 2 trial (Figure 1 and Figure 2, respectively). The studies were considered positive if the treatment effects for the primary end points were statistically significant with a 2-sided P value of less than 0.05. If the P value for the primary end point (i.e., family 1, MBL responders) was less than 0.05, the first 4 key secondary end points (i.e., family 2, amenorrhea, percentage change in MBL, UFS-QoL bleeding and pelvic discomfort, and anemia responders or NRS responders) were tested sequentially in a similar manner. If the 2-sided P value was less than 0.05 for all of the first 5 end points tested, then the remaining 3 key secondary end points (i.e., family 3, NRS responders or anemia responders, percentage change in UF volume, and percentage change in uterine volume) were tested using the Hochberg step-up procedure.

From the Hochberg procedure, P values were calculated for the 3 end points and ranked from smallest to largest. The end point corresponding to the largest P value was tested first. If it was less than 0.05, then no further testing occurred, and it was concluded that all 3 end points were significant. Otherwise, the end point corresponding to the second largest P value was tested, and if the P value was less than 0.025, then no further testing occurred, and it was concluded that the end points corresponding to the middle and smallest P values were significant. Otherwise, the end point with the smallest P value was tested, and if the P value was less than 0.0167, then no further testing occurred, and it was concluded that only the end point with the smallest P value was significant. Otherwise, all 3 end points did not pass the statistical significance criterion at the 0.05 level.

The difference between the 2 sequences was that in the LIBERTY 1 trial, the fourth key secondary end point tested was the proportion of patients with a hemoglobin level of 10.5 g/dL or lower at baseline who experienced an increase of greater than or equal to 2 g/dL (part of family 2), and the fifth key secondary end point tested was the proportion of patients who experienced a maximum NRS score of less than or equal to 1 for UF-associated pain in the subset of patients with a maximum pain score of greater than or equal to 4 who had completed the daily e-diary during the 35 days before randomization (part of family 3), whereas in the LIBERTY 2 trial, these end points were reversed in sequence. According to the sponsor, the change in the order of hierarchical testing was made on the basis of the results of the LIBERTY 1 trial, before unblinding and data analysis in the LIBERTY 2 trial.

All other end points were assessed at a nominal 2-sided alpha of 0.05 without adjusting for multiple comparisons. Comparative statistics (i.e., 95% CIs and P values) were provided for RGX-E2-NETA versus placebo for all other secondary efficacy end points.

Data Imputation Methods

Details for how missing data were handled are summarized in Table 8. For the primary end point, imputations for missing or incomplete measures of responder status were based on duration of treatment exposure, adherence with feminine product collection compared to what was reported in the e-diary, adherence with e-diary entry, and reasons for no feminine product collection. A mixed-effects model for repeated measures was used to predict the percentage change in MBL volume from baseline through the fixed effects associated with covariates (i.e., the stratification factors of baseline MBL volume and geographic region, visit, treatment, and visit by treatment interaction) and random effects (from the individual patient). An unstructured variance-covariance matrix was assumed for each patient. Data imputation methods for the other outcomes of interest are summarized in Table 8.

Subgroup Analyses

Prespecified subgroup analyses for the primary end point included geographic region, baseline MBL volume, age, race or ethnicity, and baseline body mass index. Based on clinical expert opinion, none of the prespecified subgroups were deemed relevant to the review because these were not likely to identify a subpopulation of patients who might benefit more from treatment or to inform reimbursement decisions.

Sensitivity Analyses

Several sensitivity analyses of the primary end point were conducted to test the impacts of the following scenarios:

• Unvalidated feminine product use: This scenario analysis was conducted using the week 24 or end-of-treatment, validated MBL volume (excluding unvalidated products).

• Missing data due to inadequate collection of feminine products: No imputation was implemented for patients with less than 100% feminine product adherence and both a reported MBL volume less than 80 mL and a 50% or greater reduction from baseline.

• Early discontinuation: Patients who discontinued the study drug during the first 4 weeks for any reason — or between week 4 and week 12 due to a TEAE, surgery or other intervention for HMB, reported lack of efficacy, or bleeding complaints — were considered as nonresponders, while all other patients had their responder status defined using data from the week 24 or end-of-treatment assessment period using the last observation carried forward method.

• Duration of treatment exposure: The analysis used the “completers population,” defined as patients in the modified intention-to-treat (mITT) population who completed 24 weeks of study treatment.

• Protocol deviations: A per-protocol population analysis was used.

• Different imputation method for missing data: Using the mixed-effects model for imputing missing or partially missing MBL volumes at week 24 or end-of-treatment, a multiple imputation approach was used. An arbitrary missing pattern was assumed using Markov chain Monte Carlo imputation to generate a monotone missing pattern for the observed longitudinal MBL volume values. Imputation was performed separately by randomized treatment group, given the distinct bleeding patterns among the treatment groups. Normalizing transformations were applied to the statistics estimated from each imputed dataset before Rubin’s combination rules were applied. This combined estimation and statistical test accounted for the additional variability due to imputation to provide a robust assessment of the treatment effect.

Figure 1: LIBERTY 1 Trial Mixed-Sequence Testing Procedure for Primary and Key Secondary End Points

BL = baseline; BPD = bleeding and pelvic discomfort; EP = end point; Hgb = hemoglobin; MBL = menstrual blood loss; M-vol = myoma volume; NRS = numerical rating scale; PE = primary end point; UFS-QoL = Uterine Fibroid Symptom Quality of Life questionnaire; U-vol = uterine volume.

Source: LIBERTY 1 trial and LIBERTY 2 trial Statistical Analysis Plan.⁴⁹

Figure 2: LIBERTY 2 Trial Mixed-Sequence Testing Procedure for Primary and Key Secondary End Points

BL = baseline; BPD = bleeding and pelvic discomfort; EP = end point; Hgb = hemoglobin; max = maximum; MBL = menstrual blood loss; M-vol = myoma volume; NRS = numerical rating scale; PE = primary end point; UFS-QoL = Uterine Fibroid Symptom Quality of Life questionnaire; U-vol = uterine volume.

Source: LIBERTY 1 trial and LIBERTY 2 trial Statistical Analysis Plan.⁴⁹

Table 8: Statistical Analysis of Efficacy End Points for the LIBERTY 1 and LIBERTY 2 Trials

CMH = Cochran-Mantel-Haenszel; FP = feminine product; HRQoL = health-related quality of life; MBL = menstrual blood loss; MMRM = mixed-effects model for repeated measures; NRS = numerical rating scale; RGX-E2-NETA = relugolix-estradiol–norethindrone acetate; SSS = Symptom Severity Scale; UF = uterine fibroid; UFS-QoL = Uterine Fibroid Symptom Quality of Life questionnaire; vs. = versus.

^aDefined as the proportion of patients who experienced an MBL volume of less than 80 mL and a 50% or greater reduction from baseline MBL volume over the preceding 35 days of treatment.

^bDefined as patients who had an e-diary entry rate of at least 70%, no more than 3 consecutive days of bleeding with product use, and no more than 5 days’ total of bleeding or spotting with product use reported in the e-diary over the week 24 or end-of-treatment visit window.

Source: LIBERTY 1 trial and LIBERTY 2 trial clinical study reports, LIBERTY 1 trial and LIBERTY 2 trial Statistical Analysis Plan, and sponsor’s Summary of Clinical Evidence.^12,13,19,49 Details included in the table are from the sponsor’s Summary of Clinical Evidence.

Analysis Populations

The analyzed populations are defined in Table 9.

Table 9: Analysis Populations of the LIBERTY 1 and LIBERTY 2 Trials

mITT = modified intention to treat; PP = per protocol.

Source: Sponsor’s Summary of Clinical Evidence.¹⁹ Details included in the table are from the sponsor’s Summary of Clinical Evidence.

Results

Patient Disposition

Patient disposition is summarized in Table 10. In the LIBERTY 1 trial, 1,891 of 2,279 individuals were screened out; in the LIBERTY 2 trial, 2,517 of 2,899 individuals were screened out. The most common reasons for being screened out included not meeting the eligibility criteria (80.1% of patients in the LIBERTY 1 trial and 83.2% of patients in the LIBERTY 2 trial) and withdrawn consent (12.0% of patients in the LIBERTY 1 trial and 10.1% of patients in the LIBERTY 2 trial). In the LIBERTY 1 trial, 128 patients were randomized to each of the RGX-E2-NETA and placebo groups. The proportions of patients who discontinued from the study were similar between the groups, with 21.9% leaving the RGX-E2-NETA group and 17.2% leaving the placebo group. The most common reason for discontinuation from either treatment group was withdrawal by the patient (7.8% and 5.5% in the RGX-E2-NETA and placebo groups, respectively), with all reasons being generally balanced. Fewer patients in the RGX-E2-NETA group (60.9%) than the placebo group (68.5%) enrolled in the LIBERTY OLE study. In the LIBERTY 2 trial, 126 patients were randomized to receive RGX-E2-NETA, and 129 patients were randomized to receive placebo. A similar proportion of patients discontinued from either group: 18.3% discontinued from RGX-E2-NETA and 20.9% discontinued from placebo. The most common reasons for discontinuation were withdrawal by patient (10.3% of patients and 4.7% of patients in the RGX-E2-NETA and placebo groups, respectively) and loss to follow-up (3.2% of patients and 5.4% of patients in the RGX-E2-NETA and placebo groups, respectively). A greater proportion of patients in the RGX-E2-NETA group (67.5%) than in the placebo group (59.7%) enrolled in the LIBERTY OLE study.

Across the studies, patient disposition was generally similar, except for the proportions of patients who enrolled in the LIBERTY OLE study (as described previously).

Table 10: Summary of Patient Disposition From the LIBERTY 1 and LIBERTY 2 Trials

E2-NETA = estradiol and norethindrone acetate; mITT = modified intention to treat; NETA = norethindrone acetate; OLE = open-label extension; PP = per protocol; RGX = relugolix; RGX-E2-NETA = relugolix-estradiol–norethindrone acetate.

^aOne patient who was randomized to the placebo group did not receive the study drug and was excluded from the mITT and safety populations.

^bOne patient was excluded from the RGX-E2-NETA group of the mITT population because the patient was randomized in error before eligibility was confirmed and did not receive treatment.

^cOne patient was randomized to RGX plus delayed E2-NETA (not reported here) but received RGX-E2-NETA. This patient was included in the RGX-E2-NETA group for safety analyses.

Source: LIBERTY 1 trial and LIBERTY 2 trial clinical study reports and sponsor’s Summary of Clinical Evidence.^12,13,19 Details included in the table are from the sponsor’s Summary of Clinical Evidence.

Baseline Characteristics

Baseline characteristics are summarized in Table 11 and are limited to those that are most relevant to this review or were felt to affect the outcomes or interpretation of the study results. All patients were female, and the mean age across the studies was approximately 42 years. For both studies, the proportion of patients who were Black was lower in the RGX-E2-NETA groups than in the placebo groups (46.1% versus 51.2% in the LIBERTY 1 trial; 49.6% versus 57.4% in the LIBERTY 2 trial). Conversely, the proportion of patients who were white was higher in the RGX-E2-NETA groups than in the placebo groups (50.0% versus 44.1% in the LIBERTY 1 trial; 46.4% versus 38.0% in the LIBERTY 2 trial). Proportions of other races and ethnicities were generally balanced between treatment groups and across the studies. Baseline mean MBL volume ranged from 218.8 mL to 239.4 mL in the LIBERTY 1 trial and from 211.8 mL to 246.7 mL in the LIBERTY 2 trial; volumes were higher in the RGX-E2-NETA groups than in the placebo groups. Mean hemoglobin concentration was similar across the treatment groups in both studies. However, more patients in the RGX-E2-NETA group (32.1%) of the LIBERTY 1 trial had a hemoglobin concentration of less than 10.5 g/dL compared to the placebo group (22.0%), and fewer patients in the RGX-E2-NETA group (30.5%) had a hemoglobin concentration from 10.5 g/dL to 12 g/dL compared to the placebo group (44.1%). The proportions of patients with a hemoglobin concentration of less than 10 g/dL were generally similar between treatment groups in the LIBERTY 2 trial; however, there were more patients in the RGX-E2-NETA group (36.8%) with a hemoglobin concentration of 12 g/dL or more compared to the placebo group (27.1%). The proportion of patients reporting a baseline maximum NRS score of 4 points or more was lower in the RGX-E2-NETA group (65.6%) than in the placebo group (74.8%) in the LIBERTY 1 trial.

Table 11: Summary of Baseline Characteristics From the LIBERTY 1 and LIBERTY 2 Trials — mITT Population

BMD = bone mineral density; BMI = body mass index; HRQoL = health-related quality of life; MBL = menstrual blood loss; mITT = modified intention to treat; NRS = numerical rating scale; RGX-E2-NETA = relugolix-estradiol–norethindrone acetate; SD = standard deviation; SSS = Symptom Severity Scale; UF = uterine fibroid; UFS-QoL = Uterine Fibroid Symptom Quality of Life questionnaire.

Source: LIBERTY 1 trial and LIBERTY 2 trial clinical study reports and sponsor’s Summary of Clinical Evidence.^12,13,19 Details included in the table are from the sponsor’s Summary of Clinical Evidence.

Exposure to Study Treatments

Patient exposure to study treatments is summarized in Table 12. Durations of treatment exposure were similar across the studies, ranging from 22.0 weeks to 23.3 weeks (SD = 6.2 weeks to 7.4 weeks). Because patients received RGX (or its placebo) and E2-NETA (or its placebo) as separate pills, adherence was calculated based on each component. Adherence to the study drug ranged from 97.8% to 100.9% (SD = 5.1% to 25.8%) across the studies.

Table 12: Summary of Patient Exposure in the LIBERTY 1 and LIBERTY 2 Trials — Safety Population

E2-NETA = estradiol and norethindrone acetate; RGX = relugolix; RGX-E2-NETA = relugolix-estradiol-norethindrone acetate; SD = standard deviation.

^aPatients received RGX with commercially available E2-NETA in a copackaged configuration, or placebo of RGX and placebo of E2-NETA. To determine adherence from pill counts, each component was counted separately (i.e., RGX or placebo of RGX, and E2-NETA or placebo of E2-NETA).

Source: LIBERTY 1 trial and LIBERTY 2 trial clinical study reports and sponsor’s Summary of Clinical Evidence.^12,13,19

Concomitant Medications

Concomitant medications reported in more than 5% of patients in any treatment group are summarized in Table 13. More than 94% of patients in any treatment group reported some use of concomitant medication during the LIBERTY 1 trial and the LIBERTY 2 trial. There was an imbalance in ibuprofen use across groups in the LIBERTY 1 trial. Fifty-nine percent of patients in the RGX-E2-NETA group and 72.4% in the placebo group used ibuprofen during the study.

Table 13: Summary of Concomitant Medications From the LIBERTY 1 and LIBERTY 2 Trials — Safety Population

NOS = nitric oxide supplement; RGX-E2-NETA = relugolix-estradiol–norethindrone acetate.

^aReported in more than 5% of patients.

Source: LIBERTY 1 trial and LIBERTY 2 trial clinical study reports and sponsor’s Summary of Clinical Evidence.^12,13,19 Details included in the table are from the sponsor’s Summary of Clinical Evidence.

Protocol Deviations

In the mITT population in the LIBERTY 1 trial, 31.3% of patients in the RGX-E2-NETA group and 28.3% of patients in the placebo group had an important protocol deviation. In both groups, the most common protocol deviation was for missing key study procedures (18.0% in the RGX-E2-NETA group and 18.9% in the placebo group), such as missed endometrial biopsy (9.4% and 5.5%, respectively) and DXA not performed (1.6% and 6.3%, respectively). Aside from those described, the frequencies of reported deviations were generally balanced between the treatment groups.

From the mITT population in the LIBERTY 2 trial, 30.4% of patients in the RGX-E2-NETA group and 31.0% of patients in the placebo group had an important protocol deviation. In both groups, the most common protocol deviation was for missing key study procedures (17.6% in the RGX-E2-NETA group and 22.5% in the placebo group). Missed endometrial biopsy (12.0% and 11.6%, respectively) was the most frequently reported procedure. The frequencies of reported deviations were generally balanced between the treatment groups.

Efficacy

Key efficacy results are summarized in Table 14.

Responder Rate for Change in MBL Volume

Response to treatment was defined as having an MBL volume of less than 80 mL and a 50% or greater reduction from baseline in MBL volume over the preceding 35 days of treatment as measured using the alkaline hematin method.

In the LIBERTY 1 trial, 73.4% of patients (95% CI, 64.9% to 80.9%) who received RGX-E2-NETA compared to 18.9% of patients (95% CI, 12.5% to 26.8%) who received placebo responded to treatment. The treatment difference in the responder rate for change in MBL volume at week 24 was 54.5% (95% CI, 44.3% to 64.7%; P < 0.0001) in favour of RGX-E2-NETA.

In the LIBERTY 2 trial, 71.2% of patients (95% CI, 62.4% to 79.0%) who received RGX-E2-NETA compared to 14.7% of patients (95% CI, 9.1% to 22.0%) who received placebo responded to treatment. The treatment difference in the responder rate for change in MBL volume at week 24 was 56.5% (95% CI, 46.6% to 66.5%; P < 0.0001) in favour of RGX-E2-NETA.

When looking at the individual components that made up the response definition, in both studies, similar proportions of patients met each of the criteria necessary for a response.

Results of the 6 sensitivity analyses were similar to and supported those of the primary analysis for both the LIBERTY 1 trial and the LIBERTY 2 trial.

Improvement of Anemia

For the subset of patients who had a hemoglobin level of 10.5 g/dL or lower at baseline, response was defined as an increase of 2 g/dL or more at week 24.

In the LIBERTY 1 trial, of the 30 patients and 23 patients who were in the subset from the RGX-E2-NETA and placebo groups, respectively, 15 patients (50.0%) and 5 patients (21.7%), respectively, experienced an increase of 2 g/dL or more in their hemoglobin levels. The treatment difference in the proportion of patients who experienced an improvement in anemia at week 24 was 28.3% (95% CI, 3.7% to 52.8%; P = 0.0377) in favour of RGX-E2-NETA.

In the LIBERTY 2 trial, of the 31 patients and 37 patients who were in the subset from the RGX-E2-NETA and placebo groups, respectively, 19 patients (61.3%) and 2 patients (5.4%), respectively, experienced an increase of 2 g/dL or more in their hemoglobin levels. The treatment difference in the proportion of patients who experienced an improvement in anemia at week 24 was 55.9% (95% CI, 37.3% to 74.5%; P < 0.0001) in favour of RGX-E2-NETA.

Improvement of Pain

For the subset of patients who had a maximum NRS score of 4 points or greater during the 35 days before randomization, response to treatment was defined as a maximum NRS score of 0 points or 1 point for UF-associated pain over the last 35 days of treatment in this subset.

In the LIBERTY 1 trial, of the 84 patients and 95 patients who were in the subset from the RGX-E2-NETA and placebo groups, respectively, 33 patients (39.3%) and 11 patients (11.6%) patients, respectively, experienced a maximum NRS score of 0 points or 1 point for UF-associated pain. The treatment difference in the proportion of patients who experienced an improvement in pain to a maximum NRS score of 0 points or 1 point was 27.7% (95% CI, 15.4% to 40.0%; P < 0.0001) in favour of RGX-E2-NETA.

In the LIBERTY 2 trial, of the 93 patients and 95 patients who were in the subset from the RGX-E2-NETA and placebo groups, respectively, 34 patients (36.6%) and 17 patients (17.9%), respectively, experienced a maximum NRS score of 0 points or 1 point for UF-associated pain. The treatment difference in the proportion of patients who experienced an improvement in pain to a maximum NRS score of 0 points or 1 point was 18.7% (95% CI, 6.2% to 31.1%; P = 0.0056) in favour of RGX-E2-NETA.

Change in UFS-QoL SSS Score

In the LIBERTY 1 trial, the LSM changes from baseline to week 24 in UFS-QoL SSS score were –30.9 points (95% CI, –36.1 to –25.8 points) and –10.5 points (95% CI, –15.5 points to –5.5 points) for patients in the RGX-E2-NETA and placebo groups, respectively. The treatment difference in the LSM change from baseline to week 24 in UFS-QoL SSS score was –20.4 points (95% CI, –27.1 points to –13.7 points) in favour of RGX-E2-NETA.

In the LIBERTY 2 trial, the LSM changes from baseline to week 24 in UFS-QoL SSS score were –36.1 points (95% CI, –41.2 points to –31.0 points) and –13.7 points (95% CI, –18.8 points to –8.5 points) for patients in the RGX-E2-NETA and placebo groups, respectively. The treatment difference in the LSM change from baseline to week 24 in UFS-QoL SSS score was –22.4 points (95% CI, –29.4 to –15.4 points) in favour of RGX-E2-NETA.

Change in the UFS-QoL HRQoL Total Score

In the LIBERTY 1 trial, the LSM changes from baseline to week 24 in UFS-QoL HRQoL total score were 38.0 points (95% CI, 32.4 points to 43.5 points) and 12.8 points (95% CI, 7.5 points to 18.2 points) for patients in the RGX-E2-NETA and placebo groups, respectively. The treatment difference in the LSM change from baseline to week 24 in UFS-QoL HRQoL total score was 25.1 points (95% CI, 17.9 points to 32.3 points) in favour of RGX-E2-NETA.

In the LIBERTY 2 trial, the LSM changes from baseline to week 24 in UFS-QoL HRQoL total score were 37.8 points (95% CI, 32.9 points to 42.6 points) and 13.8 points (95% CI, 8.9 points to 18.8 points) for patients in the RGX-E2-NETA and placebo groups, respectively. The treatment difference in the LSM change from baseline to week 24 in UFS-QoL HRQoL total score was 23.9 points (95% CI, 17.2 points to 30.7 points) in favour of RGX-E2-NETA.

Table 14: Summary of Key Efficacy Results From the LIBERTY 1 and LIBERTY 2 Trials — mITT Population

CI = confidence interval; HRQoL = health-related quality of life; LSM = least squares mean; MBL = menstrual blood loss; mITT = modified intention to treat; NRS = numerical rating scale; RGX-E2-NETA = relugolix-estradiol–norethindrone acetate; SSS = Symptom Severity Scale; UF = uterine fibroid; UFS-QoL = Uterine Fibroid Symptom Quality of Life questionnaire; vs. = versus.

^aBased on exact binomial 95% CI (Clopper-Pearson).

^bAdjusted 95% CI adjusted for baseline MBL volume (< 225 mL, ≥ 225 mL) and geographic region (North America, rest of world).

^cP value based on Cochran-Mantel-Haenszel test stratified by baseline MBL volume (< 225 mL or ≥ 225 mL) and geographic region (North America or rest of world).

^dP value has been adjusted for multiple testing.

^eThe 95% CI for difference is based on the normal approximation.

^fP value is based on Cochran-Mantel-Haenszel test stratified by baseline MBL volume (< 225 mL, ≥ 225 mL).

^gThe LSM and P value are based on a mixed-effect model, with treatment, visit, region, baseline MBL, and treatment-by-visit interaction included as fixed effects. The multiple visits for each patient were the repeated measures as random effect within each patient and an unstructured covariance. The transformed score ranges from 0 to 100 based on a Likert scale (none of the time, a little of the time, some of the time, most of the time, and all the time). Lower scores indicate lower symptom severity, and higher scores indicate greater symptom severity.

^hP value has not been adjusted for multiple testing.

ⁱThe LSM and P value are based on a mixed-effect model, with treatment, visit, region, baseline MBL, and treatment-by-visit interaction included as fixed effects. The multiple visits for each patient were the repeated measures as random effect within each patient and an unstructured covariance. The transformed score ranges from 0 to 100 based on a Likert scale (none of the time, a little of the time, some of the time, most of the time, and all the time). Higher scores are indicative of better HRQoL (high score = good).

Source: LIBERTY 1 trial and LIBERTY 2 trial clinical study reports and sponsor’s Summary of Clinical Evidence.^12,13,19 Details included in the table are from the sponsor’s Summary of Clinical Evidence.

Harms

Harms are summarized in Table 15.

Adverse Events

In the LIBERTY 1 trial, 61.7% of patients who received RGX-E2-NETA and 66.1% of patients who received placebo had a TEAE. Headache (10.9% of patients in the RGX-E2-NETA group versus 15.0% of patients in the placebo group) and hot flush (10.9% of patients in the RGX-E2-NETA group versus 7.9% of patients in the placebo group) were the most common TEAEs.

In the LIBERTY 2 trial, 60.3% of patients who received RGX-E2-NETA and 58.9% of patients who received placebo had a TEAE. Headache (8.7% of patients in the RGX-E2-NETA group versus 11.6% of patients in the placebo group) was the most common TEAE.

In general, the types and frequencies of TEAEs were similar between the treatment groups for both studies and across studies.

Serious Adverse Events

In the LIBERTY 1 trial, 7 patients (5.5%) who received RGX-E2-NETA reported 10 SAEs (ankle fracture, avulsion fracture, hematemesis, hypothyroidism, menorrhagia, pelvic pain, rhabdomyolysis, uterine leiomyoma, uterine myoma expulsion, and vitreous detachment), and 2 patients (1.6%) who received placebo reported 2 SAEs (acute psychosis and pneumonia).

In the LIBERTY 2 trial, 1 patient (0.8%) who received RGX-E2-NETA reported 1 SAE (cholecystitis), and 4 patients (3.1%) who received placebo reported 5 SAEs (anemia, necrotising fasciitis, radius fracture, road traffic accident, and syncope).

Withdrawals Due to Adverse Events

In the LIBERTY 1 trial, 7 patients (5.5%) who received RGX-E2-NETA stopped treatment due to 12 TEAEs (asthenia, paresthesia, metrorrhagia, increased alanine aminotransferase, anxiety, fatigue, headache, hot flush, irritability, menorrhagia, muscle twitching, and vomiting), and 5 patients (3.9%) who received placebo stopped treatment due to 13 TEAEs (asthenia, paresthesia, alopecia, constipation, depressed mood, muscular weakness, pelvic pain, pollakiuria, pruritus [generalized], urticaria, vitreous floaters, vulvovaginal burning sensation, and vulvovaginal pruritus).

In the LIBERTY 2 trial, 3 patients (2.4%) who received RGX-E2-NETA stopped treatment due to 3 TEAEs (metrorrhagia, breast cyst, and mood swings), and 6 patients (4.7%) who received placebo stopped treatment due to 7 TEAEs (anemia, arthralgia, gastrointestinal pain, muscle spasms, necrotising fasciitis, pain in extremity, and vaginal discharge).

Mortality

There were no deaths reported in either study.

Notable Harms

According to the clinical expert consulted on this review, loss of BMD was noted as a TEAE of special interest for treatment with RGX-E2-NETA. In the LIBERTY 1 trial, the LSM changes from baseline to week 24 in BMD at the lumbar spine were –0.36% (95% CI, –0.93% to 0.22%) among patients who received RGX-E2-NETA and 0.05% (95% CI, –0.52% to 0.62%) among patients who received placebo. The treatment difference between RGX-E2-NETA and placebo in mean change from baseline to week 24 in BMD at the lumbar spine was –0.41% (95% CI, –1.16% to 0.34%).

In the LIBERTY 2 trial, the LSM changes from baseline to week 24 in BMD at the lumbar spine were –0.13% (95% CI, –0.71% to 0.46%) among patients who received RGX-E2-NETA and 0.32% (95% CI, –0.26% to 0.89%) among patients who received placebo. The treatment difference between RGX-E2-NETA and placebo in mean change from baseline to week 24 in BMD at the lumbar spine was –0.44% (95% CI, –1.23% to 0.34%).

Table 15: Summary of Harms Results From the LIBERTY 1 and LIBERTY 2 Trials — Safety Population

BMD = bone mineral density; CI = confidence interval; LSM = least squares mean; RGX-E2-NETA = relugolix-estradiol–norethindrone acetate; SAE = serious adverse event; TEAE = treatment-emergent adverse event; vs. = versus.

^aOccurring in at least 5% of patients in any treatment group.

Source: LIBERTY 1 trial and LIBERTY 2 trial clinical study reports and sponsor’s Summary of Clinical Evidence.^12,13,19 Details included in the table are from the sponsor’s Summary of Clinical Evidence.

Critical Appraisal

Internal Validity

In the LIBERTY 1 and LIBERTY 2 trials, randomization was performed using an interactive web response system and stratified by geographic region and mean screening MBL volume. Patients, investigators, and sponsor staff were blinded to treatment, and matching placebo was used. Based on baseline characteristics, there was a slight imbalance in the proportions of patients who were Black and white between the treatment groups for both studies (i.e., there was a lower proportion of patients who were Black and a higher proportion of patients who were white in the RGX-E2-NETA groups versus the placebo groups). Race and ethnicity are risk factors for developing UFs, and the incidence of UFs and occurrence of clinically relevant UFs are higher in patients who are Black.¹⁵ The mean MBL volume was higher in the RGX-E2-NETA group than in the placebo group for both studies. It is possible that the baseline imbalances could have biased the study results; however, the direction and magnitude of bias are unknown. The clinical expert consulted on this review did not expect the imbalances to meaningfully affect the interpretation of the results.

Of the patients who were randomized, 17% to 22% of patients discontinued from the studies; the most frequently reported reasons were patient withdrawal, TEAEs, and loss to follow-up. Sensitivity analyses tested the primary outcome using different data-handling methods, and the results were similar to and supportive of those for the primary analysis. The anemia and pain end points were based on subsets of patients who met the low hemoglobin and high pain criteria, respectively, at baseline, and whose disease improved (per the trial definitions). These criteria reduced the number of patients who contributed data to the anemia improvement outcome (21% and 27% of patients in the LIBERTY 1 trial and the LIBERTY 2 trial, respectively) and the pain improvement outcome (70% and 74% of patients, respectively). The magnitudes of the treatment effects were notably different between the treatment groups; however, the reasons for these differences are not clear. It may have been due to the small number of patients contributing data and/or because randomization was not stratified for these variables. As a result, there is an increased risk of bias due to a loss of prognostic balance, which increases the uncertainty in these results. For the UFS-QoL outcomes, data were missing for 24% and 19% of patients in the RGX-E2-NETA and placebo groups, respectively, in the LIBERTY 1 trial, and for 23% and 26% of patients, respectively, in the LIBERTY 2 trial. These end points were analyzed using a mixed-effects model for repeated measures, which assumes that data were missing at random and that patients for whom data were missing would continue to change similarly to those with ongoing data — an assumption that is not verifiable. Moreover, there were no sensitivity analyses to inform on how results may differ under different assumptions about the missing data. Data were also missing for the anemia, pain, and BMD outcomes (5% to 30%), leading to greater uncertainty in these results than in the MBL response. Considering the large difference in efficacy between the RGX-E2-NETA and placebo groups for the primary outcome, patients may have been able to infer which treatment they received in the study. If so, this would affect subjective outcomes like pain, UFS-QoL, and harms. The reasons for missing data for each outcome were not available for review. As a result, it is difficult to pinpoint the extent of bias and how it may have affected the interpretation of the results.

Both the diagnosis of UFs and the treatment response, defined by the primary end point, were centrally reviewed. According to the FDA, MBL volume (measured using the alkaline hematin method) is an established primary end point in clinical trials for assessing HMB.³⁸ Patients who had anemia during screening were identified, and they started (or continued their pre-existing) iron therapy. While the iron supplementation protocol was consistent across the treatment groups and studies, it is unknown if improvement in anemia was a result of iron therapy or reduced MBL; information on changes in the amount of iron throughout the studies was not available to clarify this. The incidences of different types of anemia as a TEAE were lower in the RGX-E2-NETA groups versus the placebo groups in both studies (4 patients versus 6 patients, respectively, experienced anemia in the LIBERTY 1 trial, and 4 patients versus 11 patients, respectively, experienced anemia in the LIBERTY 2 trial). Pain is subjective and variable among individuals, which can make it challenging to interpret changes in this outcome; however, the variability is expected to be similar across treatment groups. There is evidence of validity, reliability, and responsiveness for the UFS-QoL in patients with UFs. There is a low risk of bias in the measurement of study outcomes that are objective or centrally reviewed. However, for the pain, UFS-QoL, and harms outcomes, there is some risk of bias due to their subjective nature and because patients may have been able to infer what treatment they were receiving based on the large difference in the primary efficacy end point between the RGX-E2-NETA and placebo groups.

The statistical testing procedure used in the studies was acceptable. For the UFS-QoL symptom severity and HRQoL total scores, multiplicity was not controlled for in the studies; as a result, there is an increased risk of type I error. While these end points are considered supportive evidence, the results align with those for the other end points. Furthermore, the clinical expert indicated that the subgroups from the studies were not likely to identify subpopulations of patients who would benefit more from treatment; the review team agreed that results from the subgroups would not meaningfully inform reimbursement decisions.

External Validity

Based on study eligibility criteria, the clinical expert confirmed that the diagnostic methods used in the studies were consistent with those used in practice in Canada. However, the expert noted that patients would be excluded based not on an upper age limit, but on whether they were adults in the premenopausal stage. Additionally, receiving gynecological surgery or ablation procedures for UFs within the 6 months before or after treatment would not be reasons to exclude a patient from receiving RGX-E2-NETA in practice. Patients must have had HMB (i.e., MBL of at least 160 mL during 1 cycle or at least 80 mL per cycle for 2 menstrual cycles) to be included; the clinical expert stated that this was an acceptable definition for a clinical trial setting, but clarified that it is not used in clinical practice. HMB is identified through clinical assessment and described as AUB that interferes with a patient’s HRQoL and daily activities. Based on the baseline characteristics, the study populations adequately represent patients who could receive RGX-E2-NETA in Canada. Approximately three-quarters of patients across the studies were from the US, and none of the trial sites were in Canada. However, the expert felt that this did not meaningfully limit generalizability to practice in Canada.

The expert indicated no concerns with the types or usage of concomitant medications in the studies. It was noted that, in practice, clinicians would recommend that patients receiving GnRH agonists be on calcium and vitamin D supplements; however, the use of these supplements was not expected to influence the study results. For the purposes of a drug reimbursement review, placebo is not relevant to clinical practice; thus, the LIBERTY trials cannot inform on the efficacy and safety of RGX-E2-NETA relative to acceptable comparator treatments. Indirect comparison of relevant treatments for HMB associated with UFs is discussed further in the indirect evidence section of this report.

The study outcomes are not used in practice (aside from hemoglobin testing for anemia); however, the clinical expert confirmed that these were reasonable for a clinical trial setting and were the most important outcomes to address for the disease. Input from the patient group submission identified bleeding, anemia, pain, and bulk symptoms as important to treat, and noted that improved HRQoL is also important. Rather than quantifying treatment effect by measuring MBL, clinicians assess patient satisfaction with how well the treatment is addressing their symptoms and whether the patient is able to participate in daily activities without interference from HMB and UFs.

The 24-week study duration was deemed acceptable for meeting the clinical end points, based on expert opinion. Although the clinical expert does not expect most patients to use RGX-E2-NETA as a long-term treatment (rather as a short-term bridging therapy until surgery), they indicated that there are rare instances where treatment beyond 24 months may be warranted to maintain symptom improvement, such as in patients who are approaching menopause or whose disease responds well to treatment without intolerable TEAEs, including significant bone loss. It is noted in the Health Canada product monograph that the use of RGX-E2-NETA should be limited to 24 months due to the risk of continued and irreversible bone loss. There were no data available for this review to evaluate the use of the drug beyond 24 months.¹⁶ The expert noted that, should the drug be used long-term, regular BMD scans would be important to ensure the patient is not experiencing a clinically important decline in BMD; if so, discontinuation should be considered.

GRADE Summary of Findings and Certainty of the Evidence

Methods for Assessing the Certainty of the Evidence

For the pivotal studies and RCTs identified in the sponsor’s systematic review, GRADE was used to assess the certainty of the evidence for the outcomes considered most relevant to inform expert committee deliberations. A final certainty rating was determined as outlined by the GRADE Working Group:^17,18

• High certainty: We are very confident that the true effect lies close to that of the estimate of the effect.

• Moderate certainty: We are moderately confident in the effect estimate — The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. We use the word “likely” for evidence of moderate certainty (e.g., “X intervention likely results in Y outcome”).

• Low certainty: Our confidence in the effect estimate is limited — The true effect may be substantially different from the estimate of the effect. We use the word “may” for evidence of low certainty (e.g., “X intervention may result in Y outcome”).

• Very low certainty: We have very little confidence in the effect estimate — The true effect is likely to be substantially different from the estimate of effect. We describe evidence of very low certainty as “very uncertain.”

Following the GRADE approach, evidence from the RCTs started as high-certainty evidence and could be rated down for concerns related to study limitations (referring to internal validity or risk of bias), inconsistency across studies, indirectness, imprecision of effects, or publication bias.

When possible, certainty was rated in the context of the presence of an important (nontrivial) treatment effect; if this was not possible, certainty was rated in the context of the presence of any treatment effect (i.e., the clinical importance is unclear). In all cases, the target of the certainty of evidence assessment was based on the point estimate and its location relative to the threshold for a clinically important effect (when a threshold was available) or to the null. The target of the certainty of evidence assessment was the presence or absence of an important effect, informed by the clinical expert consulted for this review, for responder rate for change in MBL volume, improvement of anemia, improvement of pain, change in UFS-QoL SSS score and HRQoL total score, and the presence or absence of any (non-null) effect for change in BMD at the lumbar spine.

For the GRADE assessments, findings from the LIBERTY 1 trial and LIBERTY 2 trial were assessed together per outcome because these studies were similar in population, intervention, design, and outcome measures.

Results of GRADE Assessments

Table 2 presents the GRADE summary of findings for RGX-E2-NETA versus placebo.

Long-Term Extension Studies

The contents in this section have been informed by materials submitted by the sponsor. The following information has been summarized and validated by the review team.

Description of Studies

This section contains 2 studies: the LIBERTY extension trial (a 28-week OLE study involving patients who had completed either pivotal trial) and the LIBERTY RWS (a 52-week study involving those who had completed the LIBERTY extension trial).^31,32

The objectives of the LIBERTY extension trial were to evaluate the long-term efficacy and safety of RGX-E2-NETA for up to 52 weeks (including the 24 weeks of treatment during either of the parent studies — i.e., the LIBERTY 1 trial or LIBERTY 2 trial). It was a multinational, single-arm, phase III, OLE study that enrolled eligible patients who had completed participation in 1 of the parent studies. The core study was divided into 2 periods: an open-label treatment phase (28 weeks) and a safety follow-up period (30 days). The LIBERTY extension trial was conducted at 150 sites in 9 countries (Belgium, Brazil, Chile, Czechia, Hungary, Italy, Poland, South Africa, and the US).

Patients with UFs who completed the LIBERTY extension trial and whose disease responded to treatment at the week 48 visit (i.e., patients who demonstrated an MBL volume of less than 80 mL and a 50% or greater reduction from baseline MBL in the pivotal study) could enrol in the LIBERTY RWS, an international, phase III, double-blind, placebo-controlled RWS of RGX-E2-NETA. Patients were randomized 1 to 1 to receive either oral RGX-E2-NETA or placebo once daily for up to 52 weeks. Randomization was stratified by geographic region (Europe, Latin America, North America, and rest of world), duration of RGX-E2-NETA exposure before randomization (28 weeks or 52 weeks), and pivotal study baseline MBL volume (< 225 mL or ≥ 225 mL). The study was designed to establish the long-term efficacy of RGX-E2-NETA at week 76 (24 weeks after randomization in this study) as a primary end point and at week 104 as a secondary end point. Additionally, this study aimed to investigate the effect of re-treatment with RGX-E2-NETA. The study was conducted at 92 sites in 9 countries (Brazil, Chile, Czechia, Hungary, Italy, Poland, South Africa, and the US).

Populations

Participants were eligible for the LIBERTY extension trial if they completed 24 weeks of study drug treatment in either the LIBERTY 1 trial or the LIBERTY 2 trial. To be eligible for the LIBERTY RWS, patients had to complete the LIBERTY extension trial and have disease that responded to treatment. In both studies, patients were not expected to undergo gynecological surgery or ablation procedures for UFs within the respective study periods, including during the safety follow-up periods. Additionally, participants had to have a negative urine pregnancy test at the baseline visits and agree to continue to use acceptable nonhormonal contraceptive methods consistently during the studies and for at least 30 days after the last dose of the study drug.

Patients were excluded if they had undergone myomectomy, ultrasound-guided laparoscopic radiofrequency ablation, or any other surgical procedure for UFs; uterine artery embolization; MRI-guided, focused ultrasound; or endometrial ablation for AUB at any time during the parent study. Finally, any patient with any contraindication to treatment with E2-NETA, or who met a withdrawal criterion in the parent study, was excluded. For the LIBERTY extension trial, an additional exclusion criterion was a BMD z score of less than –2.0 or a 7% or greater decrease in BMD from the parent study baseline at the lumbar spine, total hip, or femoral neck, based on the parent study week 24 DXA assessment.

The baseline demographic and clinical characteristics of patients enrolled in the LIBERTY extension trial were consistent across pivotal trial treatment groups (Table 16). Approximately 70% of the population was enrolled in North America (i.e., US). Patients who were Black represented approximately 50% of those enrolled.

Table 16: Summary of Baseline Characteristics From the LIBERTY Extension Trial (mITT Population)

BMD = bone mineral density; BMI = body mass index; HRQoL = health-related quality of life; MBL = menstrual blood loss; mITT = modified intention to treat; PBO = placebo; RGX-E2-NETA = relugolix-estradiol–norethindrone acetate; SD = standard deviation; SSS = Symptom Severity Scale; UFS-QoL = Uterine Fibroid Symptom Quality of Life questionnaire.

Source: LIBERTY extension trial Clinical Study Report and sponsor’s Summary of Clinical Evidence.^19,31 Details included in the table are from the sponsor’s Summary of Clinical Evidence.

The baseline (week 52 of the clinical program) demographics and clinical characteristics of patients rerandomized in the LIBERTY RWS were similar across treatment groups (Table 17).

Table 17: Summary of Baseline Characteristics of Patients Enrolled in the LIBERTY RWS (mITT Population)

BMD = bone mineral density; BMI = body mass index; HRQoL = health-related quality of life; MBL = menstrual blood loss; mITT = modified intention to treat; PBO = placebo; RGX-E2-NETA = relugolix-estradiol–norethindrone acetate; RWS = randomized withdrawal study; SD = standard deviation; SSS = Symptom Severity Scale; UFS-QoL = Uterine Fibroid Symptom Quality of Life questionnaire.

Note: Baseline values were defined by the pivotal study baseline value unless stated otherwise, and percentages were based on the total number of patients in each parent study treatment group or total.

Source: LIBERTY RWS Clinical Study Report and sponsor’s Summary of Clinical Evidence.^19,32 Details included in the table are from the sponsor’s Summary of Clinical Evidence.

Interventions

LIBERTY Extension Trial

All patients received RGX 40 mg, E2 1 mg, and NETA 0.5 mg for 28 weeks.

LIBERTY RWS

Patients received 1 of the following blinded oral study treatments for 52 weeks:

• RGX 40 mg, E2 1.0 mg, and NETA 0.5 mg

• RGX placebo plus E2-NETA placebo.

The blinding procedure was the same as that used in the LIBERTY 1 trial and LIBERTY 2 trial. For patients whose HMB recurred during the LIBERTY RWS (i.e., MBL volume ≥ 80 mL), blinded treatment was stopped, and re-treatment with open-label RGX-E2-NETA was offered.

Outcomes

The outcomes in the LIBERTY extension trial and LIBERTY RWS are outlined in Table 18 and Table 19, respectively, and are measured as described in the main report.

Table 18: Outcomes Summarized From the LIBERTY Extension Trial

HRQoL = health-related quality of life; MBL = menstrual blood loss; SSS = Symptom Severity Scale; UFS-QoL = Uterine Fibroid Symptom Quality of Life questionnaire.

Source: LIBERTY extension trial Clinical Study Report and sponsor’s Summary of Clinical Evidence.^19,31 Details included in the table are from the sponsor’s Summary of Clinical Evidence.

Table 19: Outcomes Summarized From the LIBERTY RWS

BMD = bone mineral density; DXA = dual-energy X-ray absorptiometry; HRQoL = health-related quality of life; MBL = menstrual blood loss; RGX-E2-NETA = relugolix-estradiol–norethindrone acetate; RWS = randomized withdrawal study; SSS = Symptom Severity Scale; UFS-QoL = Uterine Fibroid Symptom Quality of Life questionnaire.

Source: LIBERTY RWS Clinical Study Report and sponsor’s Summary of Clinical Evidence.^19,32 Details included in the table are from the sponsor’s Summary of Clinical Evidence.

Statistical Analysis

LIBERTY Extension Trial

Sample Size and Power Calculation

Because this was an OLE study, the sample size was determined by the number of patients who had completed either parent study (i.e., the LIBERTY 1 trial or LIBERTY 2 trial) and were eligible and willing to participate in this study. It was estimated that approximately 600 patients (i.e., 75% of the 780 patients who were randomized into the parent studies) would participate in the OLE study. Of note, the number of patients included those who completed the parent studies in the delayed RGX-E2-NETA arm that is not presented in this summary.

Statistical Testing

Only descriptive statistics were reported for the LIBERTY extension trial. The handling of missing data was performed as described for the parent studies, except that the last observation carried forward approach was used for the latest nonmissing MBL volume to determine response status. No sensitivity analysis was conducted.

Subgroup Analyses

In the LIBERTY extension trial, subgroup analyses of the primary efficacy end point were performed to assess whether the response rate was consistent across clinically important subgroups, including geographic region, MBL volume, age, race or ethnicity, UF volume, uterine volume, BMI, maximum NRS score for UF-associated pain, history of prior pregnancy, alcohol use, and smoking history at baseline and pivotal study baseline. Based on clinical expert opinion, none of the prespecified subgroups were deemed relevant to the review because these subgroups were not likely to identify a subpopulation of patients who might benefit more from treatment or to inform reimbursement decisions.

Analysis Populations

The analysis sets used in the LIBERTY extension trial are described in Table 20.

Table 20: Analysis Populations of the LIBERTY Extension Trial

OLE = open-label extension; RGX = relugolix.

Source: LIBERTY extension study Clinical Study Report and sponsor’s Summary of Clinical Evidence.^19,31 Details included in the table are from the sponsor’s Summary of Clinical Evidence.

LIBERTY RWS

The statistical analyses used in LIBERTY RWS are outlined in Table 21.

Table 21: Statistical Analysis of Efficacy End Points of the LIBERTY RWS

eCRF = electronic case report form; FP = feminine product; FPRR = feminine product return rate; HRQoL = health-related quality of life; IWRS = interactive web response system; MBL = menstrual blood loss; mITT = modified intention to treat; PP = per protocol; RGX = relugolix; RGX-E2-NETA = relugolix-estradiol–norethindrone acetate; RWS = randomized withdrawal study; SSS = Symptom Severity Scale; UFS-QoL = Uterine Fibroid Symptom Quality of Life questionnaire; vs. = versus.

^aThe week 76 completers population is defined as patients in the mITT population who completed 24 weeks of treatment.

Source: LIBERTY RWS Clinical Study Report and sponsor’s Summary of Clinical Evidence.^19,32 Details included in the table are from the sponsor’s Summary of Clinical Evidence.

Sample Size and Power Calculation

Given that the LIBERTY RWS was an extension of the LIBERTY extension trial, the sample size of this study was based on the number of patients who had completed either pivotal study (N = 780 patients) and participated in the OLE study (N = 600 patients, assuming a 20% dropout rate at 6 months from the pivotal studies). Assuming that the proportion of responders at week 52 in the OLE study would be 60%, it was estimated that approximately 360 patients would be randomized to this study.

With 360 patients, the study would have at least 90% power to detect a difference of 20% or greater between RGX-E2-NETA and placebo for the primary end point. The following assumptions were used to determine the sample size for this study:

• 2-sided type I error rate: 0.05

• randomization: 1 to 1

• responder rate for the placebo group: 40%

• responder rate for the RGX-E2-NETA group: 60%

• dropout rate: 20%.

Multiple Testing Procedure

In the LIBERTY RWS, to test whether RGX-E2-NETA was statistically superior to placebo for the primary efficacy end point and 3 key secondary end points, a fixed-sequence testing procedure was applied to maintain the family-wise type I error rate. Under this testing procedure, the primary end point was first tested at a 2-sided 0.05 significance level. If the P value for the primary end point was less than 0.05, then the 3 key end points would be tested sequentially: time to relapse, sustained responder rate at week 104, and amenorrhea rate at week 76.

Subgroup Analyses

In the LIBERTY RWS, subgroup analyses of the primary efficacy end point were performed to assess whether the response rate was consistent across clinically important subgroups, including geographic region, MBL volume, duration of exposure to RGX-E2-NETA before randomization, age, race or ethnicity, UF volume, uterine volume, body mass index, and maximum NRS score for UF-associated pain. Based on clinical expert opinion, none of the prespecified subgroups were deemed relevant to the review because the subgroups were not likely to identify a subpopulation of patients who might benefit more from treatment or to inform reimbursement decisions.

Analysis Populations

The analysis sets used in the LIBERTY RWS trial are described in Table 22.

Table 22: Analysis Populations of the LIBERTY RWS

E2-NETA = estradiol and norethindrone acetate; mITT = modified intention to treat; PP = per protocol; RGX = relugolix; RWS = randomized withdrawal study.

Source: LIBERTY RWS Clinical Study Report and sponsor’s Summary of Clinical Evidence.^19,32 Details included in the table are from the sponsor’s Summary of Clinical Evidence.

Results

Patient Disposition

Summaries of patient disposition in the LIBERTY extension trial and LIBERTY RWS are provided in Table 23 and Table 24, respectively. Of note, each pivotal study contributed approximately 50% of the patients enrolled in the LIBERTY extension trial. Patients who were randomized in the LIBERTY RWS included those who had received RGX-E2-NETA for 52 weeks (i.e., 24 weeks in 1 of the pivotal studies and 28 weeks in the LIBERTY extension trial); had been randomized to placebo in the pivotal studies and then received 28 weeks of open-label RGX-E2-NETA in the LIBERTY extension trial; or had received RGX monotherapy for 12 weeks in the pivotal study in the RGX-E2-NETA group and open-label RGX-E2-NETA for 28 weeks in the LIBERTY extension trial. Thus, although patients were randomized to the placebo group, they had at least some exposure to RGX-E2-NETA in a previous study. Of note, this placebo group contained patients who received open-label RGX-E2-NETA as rescue therapy after experiencing relapsing HMB during the LIBERTY RWS.

Exposure to Study Treatments

Exposures to open-label and randomized treatments are summarized in Table 25 and Table 26, respectively. In the LIBERTY extension trial, most patients in the RGX-E2-NETA-to-RGX-E2-NETA group received treatments for up to 52 weeks from the start of the pivotal trials, and most patients in the placebo-to-RGX-E2-NETA group were exposed for more than 24 to 36 weeks, which is reflective of patients’ transition into open-label RGX-E2-NETA after the pivotal trials.

Table 23: Patient Disposition in the LIBERTY Extension Trial

E2-NETA = estradiol and norethindrone acetate; PBO = placebo; RGX-E2-NETA = relugolix-estradiol–norethindrone acetate.

^aNumbers included patients from the delayed E2-NETA groups, which are not summarized in this report.

Source: LIBERTY extension trial Clinical Study Report and sponsor’s Summary of Clinical Evidence.^19,31

Table 24: Patient Disposition in the LIBERTY RWS

mITT = modified intention to treat; PBO = placebo; PP = per protocol; RGX-E2-NETA = relugolix-estradiol–norethindrone acetate; RWS = randomized withdrawal study.

Note: In the LIBERTY RWS, 1 patient was randomized in error and did not receive any treatment. In addition, 1 patient was randomized to placebo but received a dose of RGX-E2-NETA. This patient was considered an RGX-E2-NETA patient in the safety population.

Source: LIBERTY RWS Clinical Study Report and sponsor’s Summary of Clinical Evidence.^19,32 Details included in the table are from the sponsor’s Summary of Clinical Evidence.

Table 25: Summary of Patient Exposure to Study Drug in the LIBERTY Extension Trial (Safety Population)

PBO = placebo; RGX-E2-NETA = relugolix-estradiol–norethindrone acetate; SD = standard deviation.

^aIn the LIBERTY extension trial, cumulative duration and adherence are defined as a cumulative summary of exposure throughout the 52-week treatment period (parent study and extension study).

Source: LIBERTY extension trial Clinical Study Report and sponsor’s Summary of Clinical Evidence.^19,31 Details included in the table are from the sponsor’s Summary of Clinical Evidence.

In the LIBERTY RWS, most patients in the RGX-E2-NETA group (59.5%) were exposed to the study treatment for more than 48 weeks during the randomized treatment period, with an average exposure of 40.1 weeks (SD = 18.0 weeks). In contrast, the majority of patients in the placebo group (83.0%) discontinued placebo and were exposed to placebo for 16.9 weeks (SD = 12.2 weeks) (Table 26). Patients in both arms who had a return of their MBL volume to 80 mL or greater during the randomized treatment period could receive open-label RGX-E2-NETA as rescue medication. The mean re-treatment period was generally shorter in the RGX-E2-NETA group than in the placebo group (i.e., 27.6 weeks [SD = 12.8 weeks] versus 35.4 weeks [SD = 10.3 weeks], respectively). Most of the patients (86.4%) in the placebo group were exposed to re-treatment for more than 24 weeks due to return of HMB (MBL ≥ 80 mL) early in the LIBERTY RWS, reflecting the initiation of re-treatment within 1 month to 2 months after stopping treatment with RGX-E2-NETA at week 52 or baseline. Adherence to the study drug was high (> 98%) throughout the clinical program.

Table 26: Summary of Patient Exposure to the Study Drug in the LIBERTY RWS (Safety Population)

PBO = placebo; RGX-E2-NETA = relugolix-estradiol–norethindrone acetate; RWS = randomized withdrawal study; SD = standard deviation.

Source: LIBERTY RWS Clinical Study Report and sponsor’s Summary of Clinical Evidence.^19,32 Details included in the table are from the sponsor’s Summary of Clinical Evidence.

Concomitant Medications and Cointerventions

In the LIBERTY extension trial, a total of 464 patients (97.5%) received concomitant medications during the study, including 160 patients (98.2%) in the RGX-E2-NETA-to-RGX-E2-NETA group and 160 patients (97.6%) in the placebo-to-RGX-E2-NETA group. In the former group, the most frequently reported concomitant medications were vitamin and iron supplements and analgesics, with 61.3% of patients reporting the use of ibuprofen, 30.1% reporting the use of ferrous sulphate, 19.0% reporting the use of acetaminophen, and 17.2% reporting the use of iron. Additionally, the proportions of patients who reported using vitamin D and cholecalciferol were 12.9% and 10.4%, respectively. In the placebo-to-RGX-E2-NETA group, the most frequently reported concomitant medications were vitamin and iron supplements and analgesics, with 70.7% of patients reporting the use of ibuprofen, 30.5% reporting the use of ferrous sulphate, 25.0% reporting the use of acetaminophen, 17.1% reporting the use of iron, and 17.1% reporting the use of vitamins. Additionally, the proportions of patients who reported the use of vitamin D and cholecalciferol were 14.6% and 12.2%, respectively.

In the LIBERTY RWS, a total of 220 patients (96.5%) received concomitant medications during the randomized treatment period, including 109 patients (94.0%) in the RGX-E2-NETA group and 111 patients (99.1%) in the placebo group. The reported concomitant medications reflect vitamin and iron supplements, analgesics, and antihypertensive medications. Iron and hydrochlorothiazide were reported more frequently in the RGX-E2-NETA group, while ibuprofen, vitamins (vitamin nitric oxide supplement, vitamin D nitric oxide supplement, and vitamin B12 nitric oxide supplement), ranitidine, and ketoprofen were reported more frequently in the placebo group.

Subsequent Treatment

In the LIBERTY RWS, patients who experienced an MBL volume of 80 mL or more during the randomized, double-blinded phase with either RGX-E2-NETA or placebo entered the re-treatment period with open-label RGX-E2-NETA. More specifically, 27 patients (23.3%) in the RGX-E2-NETA group and 88 patients (78.6%) in the placebo group received re-treatment with open-label RGX-E2-NETA.

Efficacy

LIBERTY Extension Trial

Responder Rate Over the Preceding 35 Days of Treatment

As presented in Table 27, and consistent with the pivotal trials (in which 73% and 71% of participants on RGX-E2-NETA from the LIBERTY 1 and LIBERTY 2 trials, respectively, responded to treatment at week 24), 87.7% of participants who continued to be treated with RGX-E2-NETA for more than 52 weeks responded to treatment. Moreover, a response rate (75.6%) similar to that reported in the pivotal studies was observed for patients who were originally assigned to the placebo group and received open-label RGX-E2-NETA for 28 weeks in the LIBERTY extension trial. The proportions of patients meeting the criteria for the individual components of the composite primary end point (MBL < 80 mL over the last 35 days of treatment and ≥ 50% reduction in MBL from pivotal baseline over the last 35 days of treatment) were similar (87.1% and 89.0%, respectively, in the RGX-E2-NETA-to-RGX-E2-NETA group, and 75.6% and 81.1%, respectively, in the placebo-to-RGX-E2-NETA group). Results from the prespecified subgroups were consistent with those of the primary analysis.

Table 27: Responder Rate at Week 52 in the LIBERTY Extension Trial (Study Population)

CI = confidence interval; MBL = menstrual blood loss; PBO = placebo; RGX-E2-NETA = relugolix-estradiol–norethindrone acetate.

^aResponders were patients who experienced an MBL volume of less than 80 mL and a 50% or greater reduction from pivotal trial baseline over the last 35 days of treatment.

^bBased on exact binomial 95% CI (Clopper-Pearson).

Source: LIBERTY extension trial Clinical Study Report and sponsor’s Summary of Clinical Evidence.^19,31 Details included in the table are from the sponsor’s Summary of Clinical Evidence.

Table 28: Summary of Secondary Outcomes in the LIBERTY Extension Trial (Study Population)

PBO = placebo; RGX-E2-NETA = relugolix-estradiol–norethindrone acetate.

Source: Al-Hendy (2022).⁵¹

Proportion of Patients With a Hemoglobin Level of 10.5 g/dL or Lower at Baseline Who Experienced an Increase of 2 g/dL or More at Week 52

There were 39 patients (23.9%) in the RGX-E2-NETA-to-RGX-E2-NETA group and 38 patients (23.2%) in the placebo-to-RGX-E2-NETA group who had a hemoglobin concentration of 10.5 g/dL or lower at pivotal study baseline, representing the cohort of patients who had symptomatic anemia (Table 28). Patients in both pivotal study groups — the RGX-E2-NETA-to-RGX-E2-NETA group (23 patients [59.0%]) and the placebo-to-RGX-E2-NETA group (16 patients [42.1%]) — experienced an increase of more than 2 g/dL at week 52.

Change From Baseline to Week 52 in UFS-QoL SSS Score and HRQoL Total Score

In the RGX-E2-NETA-to-RGX-E2-NETA group, the LSM SSS score changed from baseline by –36.9 points (standard error [SE] = 2.2 points) at week 24; this reduction was maintained through week 52 (–37.3 points; SE = 2.5 points) (Table 29). In the placebo-to-RGX-E2-NETA group, the LSM SSS scores changed from baseline by –10.8 points (SE = 2.1 points) at week 24. After the initiation of open-label RGX-E2-NETA, scores changed, with an LSM reduction in SSS score of –35.0 points (SE = 2.5 points) at week 52.

Over the 52-week treatment, the RGX-E2-NETA group experienced an LSM improvement of 40.4 points (SE = 2.5 points) in UFS-QoL HRQoL total score, which was consistent with the LSM change from baseline score observed at week 24 (40.8 points; SE = 2.2 points) (Table 29). The improvement in overall HRQoL was similar (LSM change = 39.0 points [SE = 2.6 points] at week 52) in the placebo group upon receiving open-label RGX-E2-NETA for 28 weeks.

Table 29: Change From Baseline to Week 52 in UFS-QoL Symptoms Severity Score and HRQoL Total Score (Extension Study Population)

CFB = change from baseline; HRQoL = health-related quality of life; LSM = least squares mean; PBO = placebo; RGX-E2-NETA = relugolix-estradiol–norethindrone acetate; SE = standard error; SSS = Symptom Severity Scale; UFS-QoL = Uterine Fibroid Symptom Quality of Life questionnaire.

^aThe transformed score ranges from 0 points to 100 points. Higher scores indicate greater symptom severity. For the change from baseline, a negative score indicates improvement.

^bThe transformed score ranges from 0 points to 100 points. Higher scores indicate better HRQoL. For the change from baseline, a positive score indicates improvement.

Source: LIBERTY extension trial Clinical Study Report and sponsor’s Summary of Clinical Evidence.^19,31 Details included in the table are from the sponsor’s Summary of Clinical Evidence.

LIBERTY RWS

Sustained Responder Rate Through Week 76 and Week 104

At week 76, more patients in the RGX-E2-NETA group than in the placebo group (78.4% versus 15.1%; P < 0.0001) experienced a sustained treatment response after maintaining an MBL volume of less than 80 mL, as demonstrated in Table 30. Similar results were observed at week 104. Results from the sensitivity analyses and prespecified subgroup analyses were consistent with the primary analysis.

Table 30: Sustained Responder Rate at Week 76 and Week 104 in the LIBERTY RWS (mITT Population)

CI = confidence interval; MBL = menstrual blood loss; mITT = modified intention to treat; PBO = placebo; RGX-E2-NETA = relugolix-estradiol–norethindrone acetate; RWS = randomized withdrawal study; vs. = versus.

^aThe sustained responder rate at week 76 was defined as the proportion of patients who maintained an MBL volume of less than 80 mL through week 76. It was calculated as the Kaplan-Meier estimate of the cumulative probability of an MBL volume of less than 80 mL while on randomized treatment through week 76.

^bBased on the stratified test statistics through log-log transformation of the difference in survival curve at a fixed time point stratified by pivotal study baseline MBL volume (< 225 mL, ≥ 225 mL) and duration of prior exposure to RGX (28 weeks, 52 weeks). Aside from the end points listed in the testing order, P values (if provided) were not adjusted for multiplicity; thus, the values are at a nominal level of 0.05.

Source: LIBERTY RWS Clinical Study Report and sponsor’s Summary of Clinical Evidence.^19,32 Details included in the table are from the sponsor’s Summary of Clinical Evidence.

Proportion of Patients Who Responded (MBL Volume < 80 mL) to Re-Treatment With RGX-E2-NETA Among Those Whose MBL Volume Had Returned to 80 mL or Greater

As presented earlier, most patients (81%) who were rerandomized to the placebo group had their MBL volume return to 80 mL or greater. Following re-treatment with open-label RGX-E2-NETA, nearly all patients (97.8%) experienced treatment response within 2 menstrual cycles (Table 31). A smaller proportion of patients (23%) experienced disease relapse in the RGX-E2-NETA group rather than in the placebo group; most of these occurrences were single episodes of bleeding that did not recur with re-treatment.

Change From LIBERTY RWS Baseline to Week 104 in UFS-QoL SSS Score and HRQoL Total Score

As presented in Table 32, UFS-QoL SSS scores remained relatively stable from week 52 to week 104 in the RGX-E2-NETA group. For patients who received placebo, cessation of treatment was accompanied by the return of symptoms, as indicated by the higher SSS scores at week 64 (i.e., 12 weeks following rerandomization); these improved at subsequent time points as patients received re-treatment with RGX-E2-NETA.

HRQoL total scores remained relatively stable in the RGX-E2-NETA group. However, in the placebo group, HRQoL scores decreased at week 64 compared to week 52 or baseline because patients transitioned from RGX-E2-NETA to placebo. Upon reinitiation of RGX-E2-NETA, HRQoL total scores improved to a level similar to that of the group that had continued RGX-E2-NETA.

Table 31: Proportion of Patients Who Responded to Re-Treatment in the LIBERTY RWS (mITT Population)

CI = confidence interval; MBL = menstrual blood loss; mITT = modified intention to treat; PBO = placebo; RGX-E2-NETA = relugolix-estradiol–norethindrone acetate; RWS = randomized withdrawal study.

^aResponders were patients with an MBL volume of less than 80 mL during the re-treatment period. MBL volume was calculated as the sum of MBL volume from all feminine products (validated, unauthorized, and unvalidated) assessed using the alkaline hematin method at each visit.

^bBased on exact binomial 95% CI (Clopper-Pearson).

Source: LIBERTY RWS Clinical Study Report and sponsor’s Summary of Clinical Evidence.^19,32 Details included in the table are from the sponsor’s Summary of Clinical Evidence.

Table 32: Change From Baseline to Week 104 in UFS-QoL SSS Score and HRQoL Total Score (mITT population)

CFB = change from baseline; HRQoL = health-related quality of life; mITT = modified intention to treat; PBO = placebo; RGX-E2-NETA = relugolix-estradiol–norethindrone acetate; RWS = randomized withdrawal study; SD = standard deviation; SSS = Symptom Severity Scale; UFS-QoL = Uterine Fibroid Symptom Quality of Life questionnaire.

^aThe transformed score ranges from 0 points to 100 points. Higher scores indicate greater symptom severity. For the change from baseline, a negative score indicates improvement.

^bThe transformed score ranges from 0 points to 100 points. Higher scores indicate better HRQoL. For the change from baseline, a positive score indicates improvement.

Source: LIBERTY RWS Clinical Study Report and sponsor’s Summary of Clinical Evidence.^19,32 Details included in the table are from the sponsor’s Summary of Clinical Evidence.

Harms

Key harms results are presented in Table 33. Headaches and hot flushes remained the most common TEAEs. In the LIBERTY extension trial, the proportion of patients with 1 or more TEAEs was numerically smaller among patients who remained on RGX-E2-NETA compared with those who transitioned from placebo. In the LIBERTY RWS, a numerically larger proportion of patients reported 1 or more TEAEs in the placebo group compared with the RGX-E2-NETA group.

SAEs were reported infrequently in the LIBERTY extension trial and the LIBERTY RWS. In the LIBERTY extension trial, SAEs were reported for 1 patient (0.6%) in the RGX-E2-NETA-to-RGX-E2-NETA group and 11 patients (6.7%) in the placebo-to-RGX-E2-NETA group during the 28-week open-label period. The SAE in the RGX-E2-NETA group was a uterine hemorrhage, and it led to the patient leaving the study. In the placebo-to-RGX-E2-NETA group, 2 patients with an SAE of menorrhagia, 1 patient with metrorrhagia, and 1 patient with atrial fibrillation withdrew from the study.

In the LIBERTY RWS, SAEs were reported for 2 patients (1.7%) in the RGX-E2-NETA group and 2 patients (1.8%) in the placebo group. The 2 SAEs in the RGX-E2-NETA group were related to breast cancer and anemia, and led to study discontinuation; the SAEs in the placebo group included 1 case of transient global amnesia, which led to study discontinuation, and 1 case of myxoid liposarcoma; the latter patient remained on treatment.

No deaths occurred during either study. No pregnancies were reported in the LIBERTY extension trial; 1 participant in the placebo group of the LIBERTY RWS became pregnant, while no pregnancies were reported in the RGX-E2-NETA group. A respective 1.2% and 5.5% of patients in the RGX-E2-NETA-to-RGX-E2-NETA and placebo-to-RGX-E2-NETA groups discontinued the LIBERTY extension trial due to TEAEs. One patient (1.7%) from the RGX-E2-NETA group and 3 patients (2.7%) from the placebo group in the LIBERTY RWS discontinued the treatment due to a TEAE.

The changes from baseline in lumbar spine BMD for the LIBERTY extension trial and LIBERTY RWS trial are presented in Table 34. In the LIBERTY extension trial, the mean percentage change in BMD at the lumbar spine was –0.8% in the RGX-E2-NETA group, which was similar to the placebo group that transitioned to RGX-E2-NETA (–0.8%). In both treatment groups of the LIBERTY RWS, the percentage change in BMD was generally stable, with minimal changes over 1 year. The difference in mean percentage change from placebo was 0.7% (95% CI, –0.1% to 1.6%).

Table 33: Summary of Harms in the LIBERTY Extension Trial (Safety Population) and the LIBERTY RWS (Safety Population)

LSM = least squares mean; NR = not reported; OLE = open-label extension; PBO = placebo; RGX-E2-NETA = relugolix-estradiol–norethindrone acetate; RWS = randomized withdrawal study; SAE = serious adverse event; TEAE = treatment-emergent adverse event.

^aIn the LIBERTY extension trial, the harms are presented for the OLE study period.

^bIn the LIBERTY RWS, the harms are presented for the 52-week cumulative treatment period, which included the randomized treatment and re-treatment periods.

^cThe LSM percentage change is calculated from baseline to week 52 in the LIBERTY extension trial, and from week 52 or baseline to week 104 in the LIBERTY RWS.

Source: LIBERTY extension trial and LIBERTY RWS clinical study reports and sponsor’s Summary of Clinical Evidence.^19,31,32 Details included in the table are from the sponsor’s Summary of Clinical Evidence.

Table 34: Changes From Baseline in Lumbar Spine BMD at Week 52 in the LIBERTY Extension Trial (Safety Population) and at Week 104 in the LIBERTY RWS (Safety Population)

BMD = bone mineral density; CI = confidence interval; LSM = least squares mean; NA = not applicable; PBO = placebo; RGX-E2-NETA = relugolix-estradiol–norethindrone acetate; RWS = randomized withdrawal study.

Source: LIBERTY extension trial and LIBERTY RWS clinical study reports and sponsor’s Summary of Clinical Evidence.^19,31,32 Details included in the table are from the sponsor’s Summary of Clinical Evidence.

Critical Appraisal

Internal Validity

LIBERTY Extension Trial

Patients eligible for the LIBERTY extension trial must have successfully completed either the LIBERTY 1 trial or the LIBERTY 2 trial. Therefore, patients who continued in the OLE study were more likely to be those who had responded to treatment and did not experience intolerable TEAEs. As a result, the results may not be generalizable to all patients who may be treated with RGX-E2-NETA in clinical practice. The single-arm OLE study was noncomparative; this prevents conclusions from being made about the benefits or harms attributable to RGX-E2-NETA versus any comparator. In addition, the open-label designs could bias the assessment of subjective, patient-reported outcomes (i.e., UFS-QoL and harms). The lack of data beyond week 52 in the OLE study makes it challenging to draw conclusions about the long-term sustainability of the treatment effect.

In the single-arm OLE trial, all patients in the RGX-E2-NETA-to-RGX-E2-NETA group received 28 weeks of RGX-E2-NETA (resulting in a total of 52 weeks of active treatment for those who had received RGX-E2-NETA in the parent studies). The Health Canada product monograph recommends limiting treatment to 2 years due to the risk of continued and irreversible bone loss.¹⁶ According to the clinical expert consulted for this review, when RGX-E2-NETA is used as a preoperative treatment, patients are not likely to receive the drug for more than 2 years. Instead, treatment would be for a few months in preparation for surgery. The expert noted that if a patient is meeting their treatment goals without intolerable TEAEs or clinically significant bone loss, then treatment beyond 2 years could be an option; however, these would be rare cases. With extended treatment, BMD measurements at regular intervals would be necessary to ensure that a patient is not experiencing bone loss.

Up to 22% of patients discontinued the study. The reasons for missing data for each outcome were not available. As such, the effect of the missing data and its magnitude remains uncertain. Although up to 98% of patients had at least 1 concomitant medication, the clinical expert consulted by CDA-AMC had no concern about the impact of these medications on the treatment results.

LIBERTY RWS

No concerns were identified with respect to the methods used for randomization or allocation concealment. There were some minor demographic differences between the groups at baseline; however, the disease-related characteristics appeared balanced. The clinical expert agreed that any baseline discrepancies were unlikely to bias the findings. A fixed-sequence testing procedure was applied to maintain the family-wise type I error rate for the primary efficacy end point as well as the 3 key secondary end points. Although similar proportions of patients in the placebo group (24%) and RGX-E2-NETA group (23%) discontinued, it is unclear if the treatment groups remained balanced throughout the study. The study was double blind, which is important given the subjective nature of the patient-reported outcomes; however, there is the risk of unblinding due to the large difference in efficacy results between the active and placebo groups. There was evidence of validity, reliability, and responsiveness for both the UFS-QoL SSS score and the HRQoL total score. There were imbalances in the use of concomitant medications. For example, there were higher levels of iron, hydrochlorothiazide, metformin, and amlodipine in the RGX-E2-NETA group compared to the placebo group. However, the clinical expert did not think these imbalances would meaningfully affect the study results. There is a risk of bias due to missing outcome data, particularly at longer follow-up times (i.e., 76 weeks and 104 weeks), for the UFS-QoL SSS score and the HRQoL total score. At 76 weeks, data were missing for 25% of patients in the RGX-E2-NETA group and for 81% of patients in the placebo group. At 104 weeks, data were missing for 43% of patients in the RGX-E2-NETA group and for 93% of patients in the placebo group. Similarly, there is a risk of bias due to missing data for BMD in the lumbar spine at 104 weeks, given that data were missing for 37% to 39% of patients across groups.

External Validity

LIBERTY Extension Trial

Patients who were enrolled in the LIBERTY extension trial had completed the parent trials; therefore, the trial has the same generalizability issues as the parent trials. Although no study sites were located in Canada, the clinical expert did not have concerns with generalizing based on geography. The clinical expert noted that older age or having undergone recent gynecological surgery or ablation procedures for UFs would not exclude a patient from being treated with RGX-E2-NETA. Because these groups were not represented in the study, the results may not be generalizable to these patients. The efficacy outcomes assessed in the study were of importance to patients, but the study outcomes (aside from measuring hemoglobin) are not used in the clinical setting (e.g., MBL is not quantified using the alkaline hematin method; pain and HRQoL are not measured using instruments). The patient population of the OLE study may not be reflective of the broader, more heterogeneous population in terms of demographic and clinical characteristics, particularly given that only a subset of patients in the parent trials continued to the OLE study. Therefore, the results in a real-world clinical setting may differ from those observed in the OLE study.

LIBERTY RWS

There were no study sites in Canada, but the clinical expert did not have concerns about generalizing based on geography. However, the results may not be generalizable to patients with conditions that excluded participants from the study, such as having undergone medical or surgical procedures for UFs, having a weight that exceeded the limit of the DXA scanner (or a condition that precluded an adequate DXA measurement), or having or developing a contraindication to treatment with E2 or NETA (e.g., breast cancer or estrogen-dependent neoplasia). Additionally, the race and ethnicity proportions in the study may not reflect those of the broader population of patients in Canada. In that case, the results might not be generalizable, particularly for the ethnic groups in which UFs tend to be more clinically significant.

Indirect Evidence

The contents in this section have been informed by materials submitted by the sponsor and by clinical expert input. The following information has been summarized and validated by the review team.

The sponsor submitted a systematic review with an NMA, a type of ITC, to estimate the efficacy and safety of RGX-E2-NETA compared with LA with or without add-back for the treatment of HMB associated with UFs in people in the premenopausal stage.²⁰

Objectives for the Summary of Indirect Evidence

An ITC was required because of a lack of studies directly comparing RGX-E2-NETA with LA with or without add-back (the only authorized treatment for UFs in the Canadian setting).

Description of Indirect Comparisons

Objectives

To support the NMA, the sponsor:

• conducted an SLR to identify RCTs to estimate the relative efficacy and safety of RGX-E2-NETA versus LA (with or without add-back)

• performed an NMA to assess the relative efficacy and safety of RGX-E2-NETA versus LA (without add-back) for the treatment of HMB associated with UFs in people in the premenopausal stage.

SLR Methods

Table 35 shows the study selection criteria and key aspects of the methods for the SLR. No predetermined protocol was identified for the SLR. The patient population of interest was people in the premenopausal stage with a confirmed diagnosis of symptomatic UFs.

An initial search was conducted from inception of the databases until February 7, 2022. An update was performed on December 10, 2024.

According to the sponsor, an SLR was conducted following Cochrane guidance.⁵² The search strategy was peer-reviewed using the Peer Review of Electronic Search Strategies checklist.⁵³ Searches were performed for clinical trials of RGX-E2-NETA, LA, and several other historical comparators in MEDLINE and Embase through the Ovid platform. Historical comparators were treatments no longer in use in Canada.

A search of grey literature was conducted on Clinicaltrials.gov, the WHO International Clinical Trials Registry Platform, the European Union Clinical Trial Registry, Health Canada’s clinical trial databases, Google, and Google Scholar. Hand searches of reference lists of included trials and key SLRs were conducted during screening. Unpublished data from conference abstracts were included from the American College of Obstetricians and Gynecologists and the European Society of Human Reproduction and Embryology.

Two independent reviewers performed the study selection process, which comprised title and abstract screening and full-text article review. Discrepancies were resolved by consensus or a third reviewer. For inclusion in the NMAs, studies in the SLR had to meet the study selection criteria outlined in Table 35.

Data were extracted by a single reviewer using a predefined data grid. This information was validated by a second reviewer. A third reviewer resolved discrepancies to ensure the accuracy and integrity of the data.

Study-level risk of bias was assessed based on the Cochrane risk of bias tool for randomized trials⁵⁴ by 2 reviewers, and discrepancies were resolved through discussion. This tool addresses random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting, and other potential sources of bias. Each domain was rated as being at “low,” “high,” or “unclear” risk of bias.

Outcomes of interest for the NMA included effectiveness, TEAEs, and tolerability. The specific effectiveness outcomes were MBL resolution and UFS-QoL symptom severity and HRQoL total score. TEAEs included hot flushes and headaches. Tolerability outcomes were discontinuations due to TEAEs.

Of 228 reports identified in the literature, 31 trials were retained from the SLR, and 12 of these were included in the NMA.

Table 35: Study Selection Criteria and Methods for ITCs Submitted by the Sponsor

AE = adverse event; BMD = bone mineral density; GA = goserelin acetate; GnRH = gonadotropin hormone-releasing hormone; HRQoL = health-related quality of life; ITC = indirect treatment comparison; LA = leuprolide acetate; MBL = menstrual blood loss; MFP = mifepristone; RCT = randomized controlled trial; RGX-E2-NETA = relugolix-estradiol–norethindrone acetate; SLR = systematic literature review; SPRM = selective progesterone receptor modulator; SSS = Symptom Severity Scale; TEAE = treatment-emergent adverse event; UF = uterine fibroid; UFS-QoL = Uterine Fibroid Symptom Quality of Life questionnaire.

Source: Details included in the table are from the sponsor’s ITC Technical Report.²⁰

NMA Feasibility Appraisal

The sponsor conducted a feasibility assessment by evaluating the studies’ comparators, network connectivity, similarity of outcome definitions, time points assessed, baseline characteristics, NMA assumptions, and risk of bias.

Of the 31 trials included in the SLR, 12 were retained because these allowed a connection between RGX-E2-NETA and LA (without add-back). The NMA included UPA 5 mg, UPA 10 mg, and MFP 10 mg as treatments to establish a network connection between RGX-E2-NETA and LA (without add-back). MFP was excluded from the NMA because it is not indicated for the treatment of UFs in Canada; nor is it used for UF-related symptoms. UPA was withdrawn from the Canadian market in 2020 because it was associated with morphological changes in the endometrium and associated with increased risk of a rare but serious liver injury.^27,28,55 These comparators were excluded from the main results despite being retained to establish a network connection between RGX-E2-NETA and LA (without add-back).

The efficacy outcomes included were MBL resolution, amenorrhea, and change from baseline in UFS-QoL SSS scores and UFS-QoL HRQoL total scores.

Safety outcomes included were hot flushes and headaches. According to the sponsor’s NMA technical report, networks could not be constructed for the safety outcomes of BMD loss, hypertension, fatigue, or nausea involving RGX-E2-NETA and LA.

ITC Analysis Methods

Bayesian NMAs were conducted for each outcome, as outlined in Table 36.

No prespecified statistical analysis plan was identified for the NMA.

Table 36: Analysis Methods in ITC 1

DIC = deviance information criterion; EOT = end of treatment; HRQoL = health-related quality of life; LA = leuprolide acetate; MBL = menstrual blood loss; MFP = mifepristone; NMA = network meta-analysis; PBO = placebo; PSRF = potential scale reduction factor; RGX-E2-NETA = relugolix-estradiol–norethindrone acetate; SLR = systematic literature review; SSS = Symptom Severity Scale; TEAE = treatment-emergent adverse event; UFS-QoL = Uterine Fibroid Symptoms Quality of Life questionnaire; UPA = ulipristal acetate.

^aThe cumulative dose of LA administered after 3 months is equivalent for 3.75 mg per month and 11.75 mg every 3 months; both are reported as LA.

Source: Details included in the table are from the sponsor’s SLR and NMA report.²⁰

Results of the ITC

Summary of Included Studies

A summary of the included studies is presented in Table 37. Networks for all outcomes included nodes for RGX-E2-NETA, LA 3.75 mg, UPA 5 mg, UPA 10 mg, and placebo:

• The NMA for MBL included data from 4 studies (the LIBERTY 1, LIBERTY 2, PEARL 1, and PEARL 2 trials).

• The NMA for UFS-QoL (HRQoL total score and symptom severity) included data from 7 studies (the LIBERTY 1, LIBERTY 2, MYOMEX, VENUS 1, VENUS 2, PEARL 2, and PEARL 4 trials).

• The NMA for hot flushes included data from 5 studies (the LIBERTY 1, LIBERTY 2, PEARL 2, PEARL 4, and VENUS 1 trials).

• The NMA for headaches included 5 studies (the LIBERTY 1, LIBERTY 2, PEARL 1, PEARL 2, and PEARL 4 trials).

• The NMA for discontinuation due to adverse events included 8 studies (the LIBERTY 1, LIBERTY 2, ASTEROID 2, PEARL 1, PEARL 2, PEARL 4, VENUS 1, and VENUS 2 trials).

CDA-AMC selected outcomes that were considered most relevant to inform the expert committee deliberations and finalized this list of end points in consultation with members of the expert committee.

A sample network for MBL resolution is provided in Figure 3. The networks for UFS-QoL HRQoL total score, UFS-QoL SSS score, headaches, hot flushes, and discontinuations due to adverse events contained the same connections; however, the number of patients contributing to each node and the number of studies contributing to each comparison sometimes differed.

Figure 3: Evidence Network Diagram for MBL Resolution

LA = leuprolide acetate; PBO = placebo; REL-CT = relugolix combination therapy; UPA 5 = ulipristal acetate 5 mg; UPA 10 = ulipristal acetate 10 mg.

Source: Myfembree A Systematic Review and Network Meta-Analysis Report.²⁰

Table 37: Summary of Studies Included in the NMA — Reported Definitions for Outcomes

AE = adverse event; EOT = end of treatment; LA = leuprolide acetate; MBL = menstrual blood loss; NMA = network meta-analysis; PBAC = pictorial blood loss assessment chart; PBO = placebo; RGX-E2-NETA = relugolix-estradiol–norethindrone acetate; SSS = Symptom Severity Scale; UFS-QoL = Uterine Fibroid Symptoms Quality of Life questionnaire; UPA = ulipristal acetate.

Source: Myfembree A Systematic Review and Network Meta-Analysis Report.²⁰

Assessment of Similarity

Table 38 outlines an assessment of similarity across studies.

Table 38: Assessment of Similarity for Studies in the NMAs

AE = adverse event; CDA-AMC = Canada’s Drug Agency; EOT = end of treatment; HRQoL = health-related quality of life; IM = intramuscular; LA = leuprolide acetate; MBL = menstrual blood loss; NMA = network meta-analysis; PBAC = pictorial blood loss assessment chart; SC = subcutaneous; SD = standard deviation; SSS = Symptom Severity Scale; UFS-QoL = Uterine Fibroid Symptom Quality of Life questionnaire.

^aLiu (2017) was excluded from the CDA-AMC evaluation of the NMA because the study did not provide clinically meaningful outcomes.

Source: Myfembree A Systematic Review and Network Meta-Analysis Report.²⁰

Risk of Bias of the Included Studies

A Cochrane risk of bias assessment was provided by the sponsor. In 17% of the studies (2 out of 12 studies), the risk of bias was high due to allocation concealment, and in 1 of these 2 studies, the risk of bias was also high for performance (i.e., blinding of participants). The only 2 outcomes involving these studies were amenorrhea (Irahara [2019]) and all-cause discontinuation (Liu [2017]). The remaining 10 out of 12 included studies had either some concerns or low risk of concerns.

Efficacy Results

The estimated treatment effects of each outcome were quantified as RR with a 95% CrI (Table 39) or as MD with 95% CrI (Table 40).

Table 39: [Redacted]

CrI = credible interval; LA = leuprolide acetate; MBL = menstrual blood loss; RGX-E2-NETA = relugolix-estradiol–norethindrone acetate; RR = relative risk; TEAE = treatment-emergent adverse event.

^aSample sizes in each study are not reported.

^bFor MBL, RR of less than 1 favours LA. For hot flushes, headaches, and discontinuation due to TEAEs, RR of less than 1 favours RGX-E2-NETA.

Source: Myfembree A Systematic Review and Network Meta-Analysis Report.²⁰

Table 40: [Redacted]

CrI = credible interval; HRQoL = health-related quality of life; LA = leuprolide acetate; MD = mean difference; UFS-QoL = Uterine Fibroid Symptoms Quality of Life questionnaire.

^aSample sizes in each study are not reported.

^bFor UFS-QoL HRQoL total score, a negative MD favours LA. For UFS-QoL symptom severity, a positive MD favours LA.

Source: Myfembree A Systematic Review and Network Meta-Analysis Report.²⁰

MBL Resolution

For the comparison of RGX-E2-NETA versus LA (without add-back), the point estimate for the RR for MBL favoured RGX-E2-NETA; however, the 95% CrI was wide and included the possibility that either RGX-E2-NETA or LA (without add-back) could be favoured ███ █ █████ ███ ████ ████ ██ █████.

A sensitivity analysis was conducted for the outcome definition of MBL resolution by including the PEARL 4 trial, which had a different outcome definition than the LIBERTY 1 and LIBERTY 2 trials. The results of this sensitivity analysis were consistent with those of the base-case analysis, with a wide 95% CrI, and included the possibility that either RGX-E2-NETA or LA (without add-back) could be favoured ███ █ █████ ███ ████ ████ ██ █████.

The sponsor reported I² values of less than 50% for all direct and indirect comparisons of MBL resolution. For the comparison of UPA 5 mg versus UPA 10 mg for MBL resolution, I² was 33.5%.

Consistency of MBL resolution could not be assessed because this outcome did not have direct and indirect comparisons.

Uterine Fibroid Symptoms Quality of Life

In the absence of a published MID, expert opinion was considered to estimate a clinically meaningful threshold of 15 points. Based on this value, the MDs between RGX-E2-NETA and LA for the UFS-QoL HRQoL total and SSS scores likely do not meet the threshold; therefore, these are likely not different than one another.

Both point estimate MDs of the UFS-QoL HRQoL total and SSS scores favoured LA (without add-back); however, the magnitude of the MDs did not reach a clinically important threshold of 15, as suggested by the clinical expert. The 95% CrIs crossed the line of equivalence, suggesting that either RGX-E2-NETA or LA (without add-back) could be favoured:

███████ █████ ██████ ██ ██ █████ ██████ ████ ████ ██████ ██ ████ ████████████████ ███████ █████████ ██ ██ ████ ██████ ████ ████ █████ ██ █████ █████████

For UFS-QoL HRQoL total score, 2 sensitivity analyses were performed. The first excluded 2 studies in which less than 10% of patients were Black (the PEARL 2 and PEARL 4 studies), and the second excluded the PEARL 4 study, which showed a treatment effect in the opposite direction from that of the other studies (i.e., the PEARL 1, VENUS 1, and VENUS 2 studies).

The conclusions drawn from the sensitivity analyses were similar to those drawn from the base-case analysis, favouring LA (without add-back). The 95% CrIs crossed the line of equivalence, suggesting that either RGX-E2-NETA or LA (without add-back) could be favoured:

███████ █████ █████ ████████ ██ █ ███████ ████ █ ███ ████████ ███ ████ ███████ █████ ██████ ████ ████ ██████ ██ ████ ████████████████ █████ █████ ████████ ██ █████ ████ ████ ███████████████ █████ ██████ ████ ████ ██████ ██ ████ █████████

For the UFS-QoL SSS score, 2 sensitivity analyses were conducted using the same approach as that used for the UFS-QoL HRQoL total score. First, the authors excluded 2 studies in which less than 10% of patients were Black (i.e., the PEARL 2 and PEARL 4 studies); the second excluded the PEARL 4 study, which showed a treatment effect in the opposite direction of the other studies (i.e., the PEARL 1, VENUS 1, and VENUS 2 studies).

The conclusions drawn from the sensitivity analyses were similar to those drawn from the base-case analysis, favouring LA (without add-back); however, the magnitude of the MDs did not reach the clinically important threshold of 15. The 95% CrIs crossed the line of equivalence, suggesting that either RGX-E2-NETA or LA (without add-back) could be favoured:

███████ ███████ ████████ █████ ████████ ██ █ ███████ ████ █ ███ ████████ ███ ████ ███████ ████ ██████ ████ █████ ██████ ██ █████ ████████████████ ███████ ████████ █████ ████████ ██ █████ ████ ████ ███████████████ ████ ██████ ████ █████ █████ ██ █████ █████████

Key differences across studies were reported in UFS-QoL scores with an I² greater than or equal to 50%:

• UFS-QoL (HRQoL total score): UPA 5 mg to UPA 10 mg, with an I² of 72.7%

• UFS-QoL score (symptom severity): UPA 5 mg to UPA 10 mg, with an I² of 59.1%.

Consistency was assessed for 1 node only for the following outcomes using the node-splitting analysis:

• For UFS-QoL HRQoL total score, consistency was assessed between comparators UPA 10 mg and LA, with a P value of 0.61.

• For UFS-QoL SSS score, consistency was assessed between comparators UPA 10 mg and LA, with a P value of 0.55.

Harms Results

Hot Flushes

RGX-E2-NETA was associated with a lower risk of hot flushes than LA (without add-back), with an RR of | ████ ████ █ ██ ██████ A sensitivity analysis was conducted that included studies beyond 1 month after the end of treatment. The results of this sensitivity analysis were consistent with those of the base-case analysis, with an RR of ████ ████ ████ ████ ██ ██████

Headaches

For the outcome of headaches, the RR estimate was imprecise, suggesting that the direction of the effects are inconclusive. The point estimate for the risk ratio for headaches favoured LA without add-back; however, the 95% CrI was wide, including the potential that either treatment could be favoured ███ █ ███████ ███ ███ ████ ██ ██████████

A sensitivity analysis was conducted by including a trial (the VENUS 2 study) with a different time point estimate (i.e., reporting headaches for up to 2 menstrual cycles after the end of treatment). The results of this analysis were consistent with those of the base-case analysis, with a similarly high level of imprecision. This sensitivity analysis suggests that either RGX-E2-NETA or LA (without add-back) could be favoured ███ █ ███████ ███ ████ ████ ██ █████████

For the safety outcomes, in the comparison of UPA 5 mg versus UPA 10 mg, heterogeneity was low (I² = 0% for hot flushes and I² = 6.9% for headaches).

Discontinuation Due to TEAEs

The RR estimate of the tolerability outcome discontinuation due to TEAEs favoured RGX-E2-NETA with a high degree of imprecision, with an RR of ████ ████ ████ ████ ██ █████. The 95% CrI also includes values that suggest that either RGX-E2-NETA or LA (without add-back) could be favoured. No sensitivity analysis was reported for this outcome.

For discontinuations due to TEAEs, heterogeneity between RGX-E2-NETA and placebo was low (I² = 18.6%). Consistency between UPA 10 mg and placebo was supported (P = 0.35).

Critical Appraisal of ITC

A systematic review of published and grey literature was conducted to identify studies for the NMA. A prespecified protocol for the SLR and NMA could not be obtained. As such, it is unclear whether the analyses were preplanned and whether the reported analyses were selected based on their results (i.e., direction and/or magnitude of effect). The CDA-AMC review team appraised the submitted SLR using the criteria from A MeaSurement Tool to Assess systematic Reviews 2.²¹

Overall, the searches were comprehensive, and the methods for data extraction and risk of bias appraisal were adequate to reduce the risk of error and bias in these procedures. Although study selection was undertaken by 2 independent reviewers with consensus, there was inconsistent rationale for why certain studies were excluded from the networks. Of the 31 studies identified through the SLR, 12 studies were retained for the NMAs. Although certain studies of comparators not used in Canada were included in the networks to allow for connections between RGX-E2-NETA and LA, other studies with the same comparators were excluded. For example, Bagaria (2009) — comparing placebo and MFP 10 mg — was included to connect the network, whereas other studies that included a treatment group receiving MFP 10 mg were excluded because the treatments were not of interest. The selective inclusion of studies in the network may result in risk of bias due to missing results in the syntheses; the magnitude and direction of this bias cannot be predicted.

A Bayesian NMA included some outcomes, but not all, including effectiveness, safety, and tolerability. Outcomes deemed relevant by patients and clinicians that were not included in the constructed networks were efficacy outcomes of improvements in anemia and reduction in pain, and the harms outcomes of BMD. BMD as a safety outcome was not included due to lack of comparative data. As such, this NMA cannot inform on the comparative effects of RGX-E2-NETA versus LA for these outcomes.

There are limitations to both fixed-effects and random-effects models, both of which were used in this NMA. Models were selected based on better fit (i.e., a smaller deviance information criterion). Fixed-effects models were used for outcomes such as MBL resolution, hot flushes, headaches, and discontinuation due to TEAEs. Random-effects models with informative priors⁵⁶ were used for the UFS-QoL outcomes (i.e., HRQoL total and SSS scores). The selection of informative priors in this context was appropriate. Using fixed-effects models, the width of the 95% CrI may underestimate the uncertainty arising from the between-study heterogeneity because treatment effects between studies are assumed to be identical. Using random-effects models, accounting for between-study heterogeneity can lead to wider CrIs and greater uncertainty in estimating the effect sizes. Given that most outcomes were modelled with fixed effects, we expect that the true magnitude of the uncertainty was underestimated. Because the effect estimates for these end points were already affected by important imprecision, reliance on the results of the unselected models would not have changed the review team’s conclusions about the estimated effects.

A key limitation of this NMA is that the clinically relevant comparator LA did not contain add-back. Long-term use of LA is typically given with add-back to mitigate the hypoestrogenic effects of BMD loss and vasomotor symptoms. While hormonal add-back is a part of RGX-E2-NETA, according to the sponsor, LA with add-back was excluded from the NMA due to feasibility reasons. This means that RGX-E2-NETA compared with LA (without add-back) does not reflect how the actual comparator is prescribed in Canada.

The efficacy outcome of MBL, the safety outcome of headaches, and the tolerability outcome of discontinuation due to TEAEs crossed the null, with very wide CrIs. These results suggest that either RGX-E2-NETA or LA (without add-back) could be favoured, rendering the NMAs inconclusive for the outcomes.

For the HRQoL outcome measured using the UFS-QoL, the point estimates of the UFS-QoL HRQoL total score MD and UFS-QoL SSS score MD did not exceed an estimated 15-point threshold of clinical meaningfulness; both favoured LA (without add-back). The 95% CrI for both total score MD and SSS score MD also crossed the line of equivalence, suggesting that either RGX-E2-NETA or LA (without add-back) could be favoured. Another limitation of relying on the UFS-QoL was that the time point of measurement was unclear and could have differed across studies. Although the sponsor has stated that the effect of RGX-E2-NETA is relatively stable after 4 weeks of treatment, no supportive evidence was provided for the intervention or for the stability of the treatment effect for the comparator. In addition, there was insufficient detail on how stable treatment effect is defined across studies. This makes comparisons of HRQoL across studies more difficult. There were other notable limitations of the overall NMA because the assumptions of exchangeability, similarity of outcomes, and consistency (coherence) could not be satisfied.

The exchangeability (transitivity) assumption was likely not satisfied. First, the sponsor did not report a methodology on how effect modifiers were identified. The reported effect modifiers, such as ethnicity, fibroid volume, and MBL at baseline, were not reported across all included trials. Because the number of studies was small (i.e., < 10 studies per comparison), meta-regression methods were not feasible. As such, consistency in these effect modifiers across the included trials could not be comprehensively assessed. If reported, sensitivity analyses could not remove the 4 studies in which less than 10% of patients were Black because the studies were essential for connecting the nodes in the network. Second, the location of UFs, which directly affects MBL resolution, is an important effect modifier that was not assessed by the sponsor, according to the clinical expert consulted for this review. Location of UFs is graded using the International Federation of Gynecology and Obstetrics (FIGO) classification system. In this system, submucosal myomas are FIGO type 0, 1, or 2, and intramural myomas are FIGO type 3, 4, or 5.¹⁵

Potential imbalances in the distribution of effect modifiers among the included trials — such as location of UFs across studies (not reported), mean age of trial participants, and the percentage of patients who are Black — can cause bias and reduce the validity of the NMA.⁵⁷

Definitions of the relevant outcomes were variable, as reported for MBL resolution (Table 37). MBL used both objective alkaline hematin methods and a partially subjective pictorial blood loss assessment chart. The time points of outcome evaluation were heterogeneous (Table 37). The longer follow-up times in the LIBERTY 1 and LIBERTY 2 trials, in which assessment occurred in the last 35 days of treatment at week 24, could have facilitated a different MBL resolution than all the other studies, in which MBL resolution was evaluated at 12 weeks or 13 weeks.

The sponsor reported statistical heterogeneity using I², a quantification of the proportion of variation in posterior estimates due to among-study differences,⁵⁸ which was low overall across studies (< 50%). The largest magnitude of heterogeneity was reported for the outcomes of UFS-QoL (HRQoL total score and SSS score) for UPA 5 mg versus UPA 10 mg, which are no longer available in Canada. In the absence of additional information (e.g., estimated effects from the individual contributing studies), more nuanced appraisals of potential heterogeneity could not be undertaken. Therefore, the importance of such heterogeneity is uncertain.⁵⁹

The consistency (also known as coherence) assumption could not be assessed for all outcomes. The networks were sparse, consisting of many comparisons from a total of 12 studies, with a small number of studies contributing to each node. Consistency was assessed for outcomes using the node-splitting approach for the outcomes of UFS-QoL (HRQoL total score and SSS score) and discontinuation due to TEAEs. Consistency could not be assessed for MBL resolution in the absence of direct and indirect evidence. The sponsor’s reliance on statistical significance is insufficient to demonstrate the consistency assumption, given that statistical tests for incoherence have low power and may fail to detect incoherence when it is present.⁶⁰

Risk of bias assessment at the study level was evaluated using the Cochrane risk of bias tool for randomized trials.⁵⁴ In 17% of the studies (2 out of 12 studies), the risk of bias was high due to selection bias (allocation concealment); in 1 of these 2 studies, the risk of bias was high in performance (blinding of participants). The sponsor conducted sensitivity analyses removing studies with high risk of bias; therefore, the networks for the outcomes included in the current review did not include any studies with high risk of bias.

Notably, risk of bias was assessed at the study level rather than at the level of the reported effects (i.e., for each outcome of interest). This methodology ignores the fact that risk of bias can vary depending on the effect estimate being evaluated, particularly for such domains as performance, detection, attrition, and reporting bias. As such, the risk of bias reported by the sponsor for each study may not apply universally to MBL resolution, UFS-QoL HRQoL total score, UFS-QoL SSS score, or harms and tolerability outcomes.

Overall, because of the absence of an RGX-E2-NETA versus LA (with add-back) comparison, the high degree of imprecision represented by RR or MD CrIs that can favour either RGX-E2-NETA or LA (without add-back), and uncertainty related to the key assumptions required for an NMA (i.e., exchangeability, homogeneity, and consistency), definitive conclusions on the relative treatment effects of RGX-E2-NETA versus LA (without add-back) cannot be drawn based on this NMA.

Studies Addressing Gaps in the Systematic Review Evidence

There were no studies addressing gaps in the systematic review evidence submitted by the sponsor.

Discussion

Summary of Available Evidence

The evidence included in this review consisted of 2 pivotal studies identified through the sponsor’s systematic review (the LIBERTY 1 trial and the LIBERTY 2 trial), 2 extension studies (the LIBERTY extension trial and the LIBERTY RWS), and 1 sponsor-submitted ITC. There were no additional studies addressing gaps in the evidence submitted by the sponsor.

The LIBERTY 1 trial (N = 388) and LIBERTY 2 trial (N = 382) were phase III, double-blind, placebo-controlled RCTs evaluating the efficacy and safety of RGX-E2-NETA (40 mg RGX, 1 mg E2, and 0.5 mg NETA) (n = 128 and n = 126 in the LIBERTY 1 and LIBERTY 2 trials, respectively) once daily compared to placebo (n = 128 and n = 129, respectively) at 24 weeks. Eligible patients were females aged 18 to 50 years with a confirmed diagnosis of UFs and HMB (defined as an MBL volume of 160 mL or greater during 1 menstrual cycle or 80 mL or greater per cycle for 2 cycles), no gynecological surgery or procedure within 6 months before treatment or following enrolment, and no history or current metabolic bone disease or treatment for BMD loss. The primary outcome was MBL volume response, defined as patients who experienced an MBL volume of less than 80 mL and a 50% or greater reduction from baseline MBL volume. Other clinically relevant outcomes included improvements in anemia, pain, symptom severity, and HRQoL. The mean age across the studies was approximately 42 years. Baseline clinical characteristics were generally balanced between the groups, apart from slight imbalances in race and ethnicity (with a higher proportion of patients who were Black and a lower proportion of patients who were white in the placebo groups for both studies), MBL (between treatment groups for both studies), and pain scores (between treatment groups in the LIBERTY 1 trial).

The LIBERTY extension trial (N = 327) was a single-arm, phase III, 28-week OLE study involving patients who had completed either pivotal trial. In the OLE study, the long-term efficacy and safety of RGX-E2-NETA were evaluated for an additional 28 weeks (after the 24 weeks of treatment during either parent study). The LIBERTY RWS was a double-blind, phase III, placebo-controlled study enrolling patients who had completed the LIBERTY extension trial and responded to treatment. The LIBERTY RWS rerandomized patients 1 to 1 to receive either RGX-E2-NETA or placebo once daily for up to 52 weeks. For patients whose HMB recurred during the LIBERTY RWS (i.e., MBL volume ≥ 80 mL), blinded treatment was stopped, and re-treatment with open-label RGX-E2-NETA was offered. MBL response, improvement in anemia (in the LIBERTY extension trial only), and UFS-QoL scores were the main efficacy outcomes of interest. Baseline characteristics and imbalances were generally similar to those reported for the pivotal studies.

Due to the lack of evidence directly comparing RGX-E2-NETA with relevant therapies for the treatment of HMB associated with UFs, an ITC was summarized and appraised. An NMA was used to assess the relative efficacy and safety of RGX-E2-NETA versus LA (without add-back) for MBL response, UFS-QoL HRQoL total and SSS scores, safety, and tolerability. Other treatments (i.e., UPA and MFP) were used to establish the network connection between RGX-E2-NETA and LA; however, UPA and MFP were excluded in the comparison (because UPA has been withdrawn from the market, and MFP is not indicated for the treatment of UFs in Canada). Comparison to LA with add-back, which is used in clinical practice when LA is given for more than 3 months, was not possible because relevant studies that could connect to the networks were not available.

Interpretation of Results

Efficacy

UFs are a common type of uterine tumour and are associated with HMB, anemia, pain, and bulk symptomatology that interfere with HRQoL and daily living.¹ The overall goals of treatment are to reduce menstrual bleeding and other symptoms and increase hemoglobin levels to improve HRQoL and allow for uninterrupted activities of daily living.^6,7 Input from patients emphasized the need for safe and effective options that address the goals of treatment and do not require 3-month dosage commitments.

In clinical practice, HMB is defined as MBL that interferes with a person’s HRQoL. It is not quantified as it is in clinical trials (i.e., as more than 80 mL of blood loss during each menstrual cycle). After 24 weeks of treatment in the pivotal studies, high-certainty evidence demonstrated that more patients experienced disease response (i.e., an MBL volume of less than 80 mL and a 50% or greater reduction from baseline MBL volume) in the RGX-E2-NETA group compared to the placebo group. This treatment effect (between-group difference in the LIBERTY 1 trial = 54.5%; 95% CI, 44.3% to 64.7%; between-group difference in the LIBERTY 2 trial = 56.5%; 95% CI, 46.6% to 66.5%) was considered clinically meaningful, according to expert opinion, and statistically significant.

Because treatment continued for an additional 28 weeks in the OLE study, responder rates increased in the group of patients who continued to receive RGX-E2-NETA (87.7%; 95% CI, 81.7% to 92.3%); patients who switched from placebo to active treatment experienced disease response, but to a lesser degree (75.6%; 95% CI, 68.3% to 82.0%). However, the study lacked a randomized comparator, which precludes the ability to draw causal conclusions; and patients in the OLE study represented those who stayed on treatment and did not discontinue or have intolerable TEAEs. Responder rates in the LIBERTY RWS (in which patients were rerandomized) were similar to those observed in the pivotal studies at week 76 (63.4%; 95% CI, 52.9% to 73.9%) and week 104 (58.0%; 95% CI, 47.0% to 69.1%); however, only patients who had demonstrated a treatment response in the LIBERTY extension trial were eligible to enrol in the RWS. When RGX-E2-NETA was compared to LA (without add-back) in the NMA, the RR favoured the former; however, the results were inconclusive due to imprecision (i.e., wide 95% CrIs indicated that either treatment could be favoured). In addition, important methodological limitations prevent firm conclusions from being drawn regarding the comparative efficacy of the 2 treatments.

Improvement in anemia was assessed in a subset of patients whose baseline hemoglobin level was 10.5 g/dL or lower and who experienced an increase of more than 2 g/dL from baseline. In the pivotal studies, a greater proportion of patients who received RGX-E2-NETA compared to placebo experienced an improvement in anemia. The responder rate was notably lower in the LIBERTY 1 trial (between-group difference of 28.3%; 95% CI, 3.7% to 52.8%) compared to the LIBERTY 2 trial (between-group difference of 55.9%; 95% CI, 37.3% to 74.5%); and although it was unclear why, the difference could have been caused by the small number of patients contributing to the analysis and the loss of prognostic balance across treatment groups stemming from the use of a subpopulation in both studies. While the treatment effect of RGX-E2-NETA versus placebo in both studies was considered clinically meaningful, based on expert opinion, in the LIBERTY 1 trial, the lower bound of the 95% CI suggests that the treatment effect may not be clinically meaningful to patients. The evidence supporting this outcome was of low certainty due to the study limitations and imprecision previously described. The magnitudes of the responder rates in the OLE study were similar to those observed in the LIBERTY 2 trial. Improvement in anemia was not assessed in the LIBERTY RWS or the ITC.

Improvement in pain was evaluated in a subset of patients whose maximum pain NRS score was 4 points or greater before treatment and who experienced a maximum pain NRS score of 0 points or 1 point for UF-associated pain at the end of the study. Low-certainty evidence from the pivotal studies indicated that a greater proportion of patients who received RGX-E2-NETA compared to placebo experienced an improvement in UF-associated pain. The responder rate was higher in the LIBERTY 1 trial (between-group difference of 27.7%; 95% CI, 15.4% to 40.0%) compared to the LIBERTY 2 trial (between-group difference of 18.7%; 95% CI, 6.2% to 31.1%). As with the anemia improvement outcome, the results for pain improvement were based on a subpopulation in which prognostic balance may have been lost, thereby increasing the risk of bias. Although the treatment effect was considered clinically meaningful, the lower bound of the 95% CI in the LIBERTY 2 trial may include an effect that is not considered meaningful to patients; thus, the GRADE was rated down for imprecision. Improvements in pain were not assessed in the extension study or the ITC.

UF-associated HMB, whose symptoms contribute to the decreased HRQoL that patients experience, was highlighted as an important unmet need in both the patient group input and the clinical expert input. In the pivotal studies, patients who received RGX-E2-NETA compared to placebo showed a greater decline in symptom severity and increase in HRQoL, as measured using the UFS-QoL. The certainty of evidence for both outcomes was moderate due to the risk of bias arising from missing data (19% to 26% across the studies). Although there is no known threshold for a clinically important effect, the clinical expert indicated that the treatment effect was likely to be clinically meaningful to patients. The magnitude of effect in the OLE study was similar to that seen in the RGX-E2-NETA group from the pivotal studies for both change in symptom severity and HRQoL. In the LIBERTY RWS, patients randomized to RGX-E2-NETA continued to maintain the improvements they had gained from active treatment in the OLE study, while patients randomized to placebo showed an increase in symptom severity and a decline in HRQoL. Over time in the LIBERTY RWS, symptom severity and HRQoL tended to improve for the placebo group; however, there were increasing amounts of missing data with longer follow-up times, increasing the risk of bias for these outcomes. When RGX-E2-NETA was compared to LA (without add-back) in the NMA, the point estimate for the RR favoured LA for both the UFS-QoL SSS and HRQoL total scores; however, the 95% CrIs crossed the line of equivalence, suggesting that either treatment could be favoured. Significant methodological limitations prevent firm conclusions from being drawn regarding the comparative efficacy of the 2 treatments.

The ITC results indicated that either treatment could be favoured; imprecision and methodological limitations with the ITC preclude the ability to draw conclusions about the direction of treatment effect. The clinical expert noted that there are practical benefits associated with the use of RGX-E2-NETA, such as ease of administration (i.e., combined oral pills versus intramuscular injections that require separate oral add-back therapy), timing (LA must be timed carefully to prevent symptom flare), and quicker loss of effect once the drug is stopped (compared to a 3-month commitment with the LA injection). The expert was of the opinion that both treatments are expected to be used for similar purposes in practice (that is, for short-term use to address UF-associated HMB). As per the expert, RGX-E2-NETA would be used as a preoperative treatment to reduce the size of UFs, reduce HMB, and improve low hemoglobin in anticipation of surgery. The expert explained that doing so can facilitate less invasive surgical or medical procedures and reduce the risks of surgery (e.g., significant blood loss). RGX-E2-NETA could also be used to avoid surgery or as a bridging therapy before menopause, when symptoms tend to subside. The Health Canada product monograph states that the use of RGX-E2-NETA should be limited to 24 months due to the risk of irreversible bone loss. While treatment beyond 24 months may be an option for patients whose disease is responding well to treatment and who are not experiencing clinically significant bone loss, these are expected to be rare cases. In such instances, regular BMD assessments would be necessary to monitor bone health and inform decisions about continuing or stopping treatment.

The sponsor indicated that RGX-E2-NETA would be used in patients who wish to maintain fertility, preserve their uterus, and reduce their symptoms while avoiding surgery or other procedures. Patients in the studies were not expected to undergo gynecological surgeries or procedures within 6 months of enrolment. Considering the expected and current use of RGX-E2-NETA in practice (as per the clinical expert) and the fact that outcomes for fertility, uterus preservation, and avoidance of surgical or medical procedures were not investigated, evidence from the pivotal studies may not fully inform on the sponsor’s suggested place in therapy.

Harms

In the LIBERTY 1 and LIBERTY 2 trials, up to two-thirds of patients in each treatment group experienced at least 1 TEAE; in general, the types and frequences of TEAEs were similar between the treatment groups in either study and across the studies. The clinical expert consulted for this review indicated no concerns with the safety profile of RGX-E2-NETA and stated that TEAEs would generally be manageable for patients. Reports of SAEs and patients withdrawing from treatment due to TEAEs were low across the studies. There were no deaths.

TEAEs in the LIBERTY extension trial and the LIBERTY RWS were similar to those of the pivotal studies. There did not appear to be any new safety signals with continued use of RGX-E2-NETA. SAEs and withdrawals due to TEAEs were generally low across the studies. There were no deaths.

Based on the ITC results, RGX-E2-NETA was associated with a lower risk of hot flushes, but not headaches, compared to LA (without add-back). LA without add-back is known to exacerbate vasomotor symptoms (including hot flushes) and is a reason for having add-back included in RGX-E2-NETA. The observed difference may reflect the effect of add-back therapy in RGX-E2-NETA; thus, the difference in effect between the drugs in the ITC may not be relevant. Results comparing discontinuations due to TEAEs were uncertain due to imprecision, indicating that either treatment could be favoured for these outcomes. Because RGX-E2-NETA and LA (with or without add-back) have not been directly compared in a robust RCT — and the sponsor-submitted ITC did not include LA with add-back — it is not possible to say with uncertainty that the drugs have the same safety or efficacy. The indirect evidence available for review also had methodological issues, and the key assumptions of the NMAs could not be satisfied, preventing definitive conclusions from being made.

According to the clinical expert consulted on this review, loss of BMD was noted as a TEAE of special interest for treatment with RGX-E2-NETA. In both pivotal studies, there was a numerically larger decrease in BMD among patients who received RGX-E2-NETA compared to placebo after 24 weeks of treatment. With no known clinically meaningful threshold, it is unclear how significant the change in BMD is. Decreases in BMD were numerically larger at the end of the LIBERTY extension trial, suggesting that bone loss may continue with longer treatment times; conversely, BMD measures in the rerandomized RGX-E2-NETA group were numerically greater compared to the rerandomized placebo group in the LIBERTY RWS. The reason for the rebound is unclear, and the wide 95% CI for the between-group difference that includes 0 suggests the possibility that there is no difference between the treatment groups. Limitations in the study designs and missing data amplify the uncertainty in these estimates. The clinical expert stated that to mitigate BMD decline, preventive strategies for bone health are implemented with these types of drugs. Furthermore, it would be important to monitor BMD with prolonged use of RGX-E2-NETA; however, most patients are expected to use the treatment for a short duration (months rather than years). Bone loss may be less of a concern if treatment remains short.

Conclusion

Evidence from 2 phase III, double-blind RCTs (the LIBERTY 1 trial and the LIBERTY 2 trial) suggests that patients aged 18 to 50 years with confirmed UFs and HMB (defined as an MBL volume of 160 mL or greater in 1 cycle or 80 mL or greater per cycle for 2 cycles) who received RGX-E2-NETA once a day for 24 weeks were more likely to experience treatment response for HMB compared to those who received placebo. Patients receiving RGX-E2-NETA may also have clinically meaningful improvements in anemia, UF-associated pain, symptom severity, and HRQoL (as measured using the UFS-QoL). However, the evidence for these outcomes was uncertain due to an increased risk of bias, lack of known thresholds for a clinically meaningful effect, and missing data. Longer-term evidence suggested that treatment effects were maintained for reduction in MBL and the patient-reported outcomes of symptom severity and HRQoL (with up to 104 weeks of study treatment), as well as for improvements in anemia (with up to 52 weeks of study treatment). Limitations with the study designs and numerous discontinuations increase the uncertainty of the longer-term results. It is also likely that patients who experienced benefits without intolerable harms remained in the studies longer. Results from the ITC of RGX-E2-NETA versus LA (without add-back) indicated that either treatment could be favoured for efficacy outcomes. Additionally, LA with add-back was not included in the ITC, and definitive conclusions on its treatment effect relative to that of RGX-E2-NETA could not be drawn. The incidence and types of harms were generally balanced and similar between the treatment groups in the pivotal studies and across the studies. Longer-term evidence did not indicate new safety signals; however, expert opinion noted that assessment of long-term bone health is important to inform decisions regarding prolonged treatment with RGX-E2-NETA. The Health Canada product monograph notes that use of RGX-E2-NETA should be limited to 24 months due to the risk of bone loss; expert opinion indicated that use in clinical practice would be short-term.

References

1.Wise LA, Laughlin-Tommaso SK. Epidemiology of Uterine Fibroids: From Menarche to Menopause. Clin Obstet Gynecol. 2016;59(1):2-24. doi: 10.1097/GRF.0000000000000164 PubMed

2.Pron G, Cohen M, Soucie J, Garvin G, Vanderburgh L, Bell S. The Ontario Uterine Fibroid Embolization Trial. Part 1. Baseline patient characteristics, fibroid burden, and impact on life. Fertil Steril. 2003;79(1):112-119. doi: 10.1016/S0015-0282(02)04539-9 PubMed

3.Soliman AM, Margolis MK, Castelli-Haley J, Fuldeore MJ, Owens CD, Coyne KS. Impact of uterine fibroid symptoms on health-related quality of life of US women: evidence from a cross-sectional survey. Curr Med Res Opin. 2017;33(11):1971-1978. doi: 10.1080/03007995.2017.1372107 PubMed

4.Magnay JL, Nevatte TM, Dhingra V, O'Brien S. Menstrual blood loss measurement: validation of the alkaline hematin technique for feminine hygiene products containing superabsorbent polymers. Fertil Steril. 2010;94(7):2742-6. doi: 10.1016/j.fertnstert.2010.03.061 PubMed

5.Zimmermann A, Bernuit D, Gerlinger C, Schaefers M, Geppert K. Prevalence, symptoms and management of uterine fibroids: an international internet-based survey of 21,746 women. BMC Womens Health. 2012;12:6. doi: 10.1186/1472-6874-12-6 PubMed

6.Stewart EA. Uterine fibroids (leiomyomas): Treatment overview. UpToDate; 2024. Accessed February 6, 2025. https://www.uptodate.com

7.Vilos GA, Allaire C, Laberge P-Y, et al. The Management of Uterine Leiomyomas. J Obstet Gynaecol Can. 2015;37(2):157-178. doi: 10.1016/s1701-2163(15)30338-8 PubMed

8.Sayed GH, Zakherah MS, El-Nashar SA, Shaaban MM. A randomized clinical trial of a levonorgestrel-releasing intrauterine system and a low-dose combined oral contraceptive for fibroid-related menorrhagia. Int J Gynaecol Obstet. 2011;112(2):126-30. doi: 10.1016/j.ijgo.2010.08.009 PubMed

9.Tulandi T, Casper R. Management of Uterine Fibroids After the Withdrawal of Fibristal. J Obstet Gynaecol Can. 2021;43(7):797-799. doi: 10.1016/j.jogc.2020.12.002 PubMed

10.Laberge PY, Murji A, Vilos GA, Allaire C, Leyland N, Singh SS. Guideline No. 389-Medical Management of Symptomatic Uterine Leiomyomas - An Addendum. J Obstet Gynaecol Can. 2019;41(10):1521-1524. doi: 10.1016/j.jogc.2019.01.010 PubMed

11.Management of Symptomatic Uterine Leiomyomas: ACOG Practice Bulletin, Number 228. Obstet Gynecol. 2021;137(6):e100-e115. doi: 10.1097/aog.0000000000004401 PubMed

12.Myovant Sciences GmbH. Clinical Study Report: MVT-601-3001. LIBERTY 1: An International Phase 3 Randomized, Double-Blind, Placebo-Controlled Efficacy and Safety Study to Evaluate Relugolix Co-Administered with and without Low-Dose Estradiol and Norethindrone Acetate in Women with Heavy Menstrual Bleeding Associated with Uterine Fibroids [internal sponsor's report]. January 31, 2020.

13.Myovant Sciences GmbH. Clinical Study Report: LIBERTY 2: An International Phase 3 Randomized, Double-Blind, Placebo-Controlled Efficacy and Safety Study to Evaluate Relugolix Co-Administered with and without Low-Dose Estradiol and Norethindrone Acetate in Women with Heavy Menstrual Bleeding Associated with Uterine Fibroids [internal sponsor's report]. January 31, 2020.

14.Spies JB, Coyne K, Guaou Guaou N, Boyle D, Skyrnarz-Murphy K, Gonzalves SM. The UFS-QOL, a new disease-specific symptom and health-related quality of life questionnaire for leiomyomata. Obstet Gynecol. 2002;99(2):290-300. doi: 10.1016/s0029-7844(01)01702-1 PubMed

15.Stewart EA, Laughlin-Tommaso SK. Uterine fibroids (leiomyomas): Epidemiology, clinical features, diagnosis, and natural history. UpToDate; 2024. Accessed February 6, 2025. https://www.uptodate.com

16.Pfizer Canada ULC. Myfembree (relugolix, estradiol, and norethindrone acetate): 40 mg/1 mg/0.5 mg oral tablets [product monograph]. October 13, 2023.

17.Balshem H, Helfand M, Schünemann HJ, et al. GRADE guidelines: 3. Rating the quality of evidence. J Clin Epidemiol. 2011;64(4):401-6. doi: 10.1016/j.jclinepi.2010.07.015 PubMed