Efficacy and safety of a selective estrogen receptor β agonist, ERB-041, in patients with rheumatoid arthritis: A 12-week, randomized, placebo-controlled, phase II study


  • ClinicalTrials.gov identifier: NCT00141830.

  • Presented in part at the 71st Annual Scientific Meeting of the American College of Rheumatology, Boston, MA, November 2007.



Selective estrogen receptor β (ERβ) agonists have demonstrated relevant antiinflammatory effects in different animal models. This study aimed to compare the efficacy and safety of one of these agonists, ERB-041, in subjects with rheumatoid arthritis (RA).


A total of 291 patients with active RA receiving stable doses of methotrexate were randomized to receive 5, 25, or 75 mg of ERB-041 or placebo for 12 weeks. The primary end point was the American College of Rheumatology 20% improvement criteria (ACR20) at 12 weeks. Secondary end points included the ACR 50% improvement criteria (ACR50) and the ACR 70% improvement criteria (ACR70) responses, health outcomes measures, C-reactive protein (CRP) levels, and potential exposure-response relationships. Medical history, physical examination, and laboratory values were obtained at screening, baseline, and weeks 2, 4, 8, and 12.


No statistically significant difference for the ACR20 was found between the ERB-041 treatment and placebo groups (P = 0.518). Nor was a significant difference observed for ACR50 and ACR70 responses, health outcomes measures, CRP levels, and overall incidence of adverse events among all groups. Forty-four subjects (15.1%) discontinued the study and the rate of discontinuation was similar among the treatment groups. The most commonly reported treatment-emergent adverse events were headache (7.6%), nausea (6.2%), infection (4.8%), and bronchitis (4.1%). None of the adverse events was considered treatment related.


Although well tolerated and safe, ERB-041 failed to demonstrate antiinflammatory efficacy in RA patients, despite evidence of strong activity in preclinical arthritis models. These results suggest that selective ERβ agonists would not have effects on regulating inflammatory response in RA. Nevertheless, further studies are warranted to establish their efficacy in inflammatory arthritis.


Sex hormones modulate the onset and progression of connective tissue diseases, including rheumatoid arthritis (RA), in both clinical and in vivo studies (1). The peak RA incidence in women occurs around menopause, when ovary dysfunction leads to a fall in estrogen and progesterone levels (2). RA activity often decreases during the postovulatory phase of the menstrual cycle and pregnancy, when sex hormones are in high levels, whereas the disease flares in the postpartum period, when sex hormone levels drop. Also, nulliparity has been associated with an increased risk of developing RA (3). Moreover, the underlying immune response in RA seems to be influenced by sex hormones (1). However, the effect of exogenous estrogens on RA remains controversial, and furthermore, their associated cardiovascular and cancer risks represent a big concern (4).

The discovery of estrogen receptor β (ERβ), a second ER, has renewed the interest in the estrogen role and regulation in inflammatory diseases, including RA. Both ERs have a highly homologous DNA-binding domain, bind identical estrogen response elements in the promoter of target genes, and antagonize NF-κB transcriptional activity (5). However, increasing evidence has emerged of different ERβ biologic functions than those exerted by the previously known ERα, and its potential as a drug target. From animal studies, ERβ appears to play a minor role in mediating classic estrogen effects in the uterus, skeleton, hypothalamus/pituitary, and other target tissues. In contrast, a leading role for ERβ has been proposed in the ovaries, cardiovascular system, brain, and inflammation. Indeed, ERβ showed potent antiinflammatory action in animal models of arthritis, endometriosis, inflammatory bowel disease (IBD), and sepsis (6).

Selective ERβ agonists have been developed for a variety of clinical applications such as RA in an attempt to avoid classic estrogenic side effects. ERB-041 (2-[3-fluoro-4-hydroxyphenyl]-7-vinyl-1,3-benzoxazol-5-ol) is a selective ERβ agonist that has shown strong antiinflammatory effects in several arthritis, IBD, and acute inflammatory pain models. Remarkably, it did not have proliferative effects on the mammary gland or uterus (7, 8). Furthermore, in 2 randomized, double-blind, placebo-controlled, phase I studies, ERB-041 was safe and well tolerated (Wyeth Research: unpublished data).

The present study evaluated the effects of an ERβ agonist in RA for the first time. Therefore, this study aims to evaluate the efficacy and safety of 3 dose levels of oral ERB-041 administered daily for 12 weeks in subjects with active RA who have had inadequate response despite stable methotrexate (MTX) therapy.


Study design.

This was a phase II, double-blind, placebo-controlled, comparative study conducted at 55 sites worldwide from July 2005 to October 2006. The study consisted of a screening period not exceeding 4 weeks, a 12-week treatment period, and a 2-week followup period. Scheduled visits were at screening, baseline, 2, 4, 8, 12, and 14 weeks, and early withdrawal. Outpatients with active RA were enrolled and randomly assigned to 1 of 4 groups: group 1 received placebo and groups 2, 3, and 4 received either 5, 25, or 75 mg of ERB-041, respectively, administered orally once per day. Patients continued to take their weekly dosage of MTX (7.5–20 mg) throughout the study. This study was designed and performed in compliance with the International Conference on Harmonisation Guidelines for Good Clinical Practice.


All of the patients had to meet the American College of Rheumatology (ACR; formerly the American Rheumatism Association) criteria for RA (9) with ACR functional class I–III (10) and have a disease duration of at least 6 months, disease onset at age >16 years, and an age range of 18–80 years. Patients should have had active RA consisting of ≥6 swollen and ≥6 painful joints (28-joint count) and at least 1 of the following: erythrocyte sedimentation rate ≥28 mm/hour, high-sensitivity C-reactive protein (CRP) level ≥15 mg/liter (at screening), or morning stiffness ≥45 minutes. Patients could have received less than 10 mg/day of prednisone or equivalent for 3 months or more. Finally, patients should have been treated with a stable, well-tolerated dose and route of administration of MTX (7.5–20 mg oral, intramuscular, or subcutaneous) weekly for at least 12 weeks before the baseline visit, and were willing to remain on this fixed dose and route of administration for the duration of the study. Institutional review boards/independent ethics committees reviewed and approved the study protocol in accordance with the Helsinki Declaration. All of the patients provided written informed consent before their inclusion in the study.

Efficacy assessment.

Efficacy evaluations were performed at every study visit. The primary end point, an ACR 20% improvement criteria (ACR20) response, was calculated at Wyeth Research facilities. The secondary end points, including the ACR 50% improvement criteria (ACR50) and the ACR 70% improvement criteria (ACR70) responses, the patient global assessment of disease activity (11-category scale), the physician global assessment of disease activity (11-category scale), the Health Assessment Questionnaire (HAQ; scale 0–3) scores, and the acute-phase reactant CRP levels, were also assessed. CRP values were used for the calculation of ACR responses and the Disease Activity Score based on the 28-joint count and CRP level (DAS28-CRP). The DAS28-CRP was calculated at Wyeth Research facilities using an algorithm that included 3 variables: the number of swollen joints, the number of tender joints, and the CRP concentration (mg/liter).

Safety assessment.

Safety evaluations included the adequate birth control/contraception and individual subject safety assessments at scheduled visits. Safety assessments were based on reports of adverse events (AEs) and results of routine physical examinations and laboratory determinations. Hormonal tests, including estradiol (postmenopausal women only) and total and free testosterone (men), were assessed at baseline and at weeks 8 and 12 or early withdrawal. Subjects with clinically relevant AEs and/or abnormal laboratory test results at the final study visit were followed up with by telephone calls, additional site visits, and/or additional laboratory evaluations until resolved or stabilized. All of the AEs were monitored throughout the study and were assessed for their relationship to test article and severity.

Statistical analyses.

For baseline data, a one-way analysis of variance model with treatment group as a factor was used to compare the groups for all variables except nominal attributes (e.g., sex, race), which were compared by the Fisher's exact test or chi-square test. For continuous efficacy end point comparison, the analysis of covariance model was employed with treatment group as a factor and baseline data as a covariate. The logistic regression model with treatment group as a factor was used for binary efficacy end point comparison. For the primary analysis, the last observation carried forward (LOCF) technique was applied for missing data imputation. The incidence of total AEs and treatment-emergent AEs, as well as clinically important laboratory measurements during the study, were compared between the treatment groups using the Fisher's exact test.


Patient disposition and demographic characteristics.

Two hundred ninety-one subjects were enrolled. The treatment groups were well balanced with respect to the number of subjects and demographic and baseline disease characteristics (Table 1). At baseline, the mean ± SD duration of RA was 9.57 ± 8.61 years and was similar between the groups. IgM rheumatoid factor was positive in 74.3% of subjects. Two hundred forty-seven subjects completed the study and a total of 44 subjects discontinued it. Thirty-four subjects (15.6%) who received ERB-041 and 10 subjects (13.7%) who received placebo were withdrawn from the study. There was no difference among the treatment groups in the rate of discontinuation (P = 0.165).

Table 1. Baseline demographic and disease characteristics of the 291 patients with rheumatoid arthritis*
 Placebo (n = 73)ERB-041
5 mg (n = 73)25 mg (n = 71)75 mg (n = 74)
  • *

    Values are the mean ± SD unless otherwise indicated. DMARD = disease-modifying antirheumatic drug; CRP = C-reactive protein; HAQ = Health Assessment Questionnaire; DAS28 = Disease Activity Score based on the 28-joint count.

  • Excluding methotrexate.

Age, years55 ± 12.055 ± 12.052 ± 10.254 ± 10.9
Women, %93.279.581.781.1
White, %
Weight, kg72.17 ± 18.075.30 ± 19.872.80 ± 16.774.30 ± 15.6
Disease duration, years9.8 ± 9.710.1 ± 8.88.7 ± 8.09.7 ± 7.8
Rheumatoid factor positive, %73.276.778.668.9
Glucocorticoid use at baseline, %
History of DMARD use, %19.217.821.114.9
CRP level, mg/liter17.8 ± 25.716.9 ± 18.017.0 ± 15.218.0 ± 17.8
Patient global assessment (range 0–10)6.2 ± 2.06.0 ± 2.16.4 ± 2.06.3 ± 2.0
Physician's global assessment (range 0–10)5.8 ± 1.56.0 ± 1.65.9 ± 1.76.0 ± 1.6
HAQ score (range 0–3)1.6 ± 0.51.4 ± 0.71.6 ± 0.71.5 ± 0.7
DAS28 CRP5.6 ± 0.85.5 ± 0.85.6 ± 0.85.6 ± 0.7

Efficacy results.

The results from the LOCF analysis of the primary efficacy end point, ACR20 at 12 weeks, for the modified intent-to-treat population showed at week 12 that 27.4% of subjects in the placebo group, 35.6% of subjects in the 5 mg of ERB-041 group, 35.2% of subjects in the 25 mg of ERB-041 group, and 27.0% of subjects in the 75 mg of ERB-041 group achieved an ACR20 response (P = 0.518). Likewise, the ACR50 and ACR70 response rates at 12 weeks were not different among the treatment groups (Figure 1).

Figure 1.

Clinical response to ERB-041 in patients with active rheumatoid arthritis during the 12-week clinical trial. Proportion of patients who achieved the American College of Rheumatology 20% (ACR20), 50% (ACR50), and 70% improvement criteria (ACR70) in 12 weeks. The patients were randomized into groups receiving placebo or 5, 25, or 75 mg of ERB-041 once daily.

Other secondary efficacy variables such as the DAS28 and CRP level had similar outcomes (Figure 2). Health outcomes measurements (HAQ and patient global assessment of disease activity, physician global assessment of disease activity, and pain and fatigue measured on a visual analog scale) also lacked a statistically significant difference between the treatment groups.

Figure 2.

Clinical response to ERB-041 in patients with active rheumatoid arthritis during the 12-week clinical trial. Change in mean Disease Activity Score in 28 joints (DAS28) and C-reactive protein (CRP; mg/liter) levels over the 12 weeks of the study in the groups receiving placebo or 5, 25, or 75 mg of ERB-041 once daily.

Safety results.

ERB-041 was generally well tolerated. There was no difference among the treatment groups in the overall incidence of treatment-emergent AEs (P = 0.763), as shown in Table 2. However, the incidences of increased serum glutamic pyruvic transaminase (SGPT) levels (P = 0.034) and RA flares (P = 0.021) were different. More subjects receiving 25 mg of ERB-041 experienced an increase of SGPT levels compared with those receiving 75 mg of ERB-041 (P = 0.012). The majority of laboratory test results were National Cancer Institute grade 0, 1, or 2. Also, more subjects experienced RA flares in the placebo and 25 mg ERB-041 groups compared with those in the 5 mg ERB-041 group (P = 0.028 and P = 0.013, respectively).

Table 2. Incidence of treatment-emergent adverse effects occurring in ≥5% of the subjects in any treatment group*
 Placebo (n = 73)ERB-041
5 mg (n = 73)25 mg (n = 71)75 mg (n = 74)
  • *

    Values are the number (percentage). SGPT = serum glutamic pyruvic transaminase; UTI = urinary tract infection; LFTs = liver function tests.

  • P < 0.05 overall by Fisher's exact test.

  • P < 0.05 for 25 mg ERB-041 and placebo vs. 5 mg ERB-041.

  • §

    P < 0.05 for 25 mg ERB-041 vs. 75 mg ERB-041.

Any adverse event46 (63.0)43 (58.9)48 (67.6)47 (63.5)
Headache3 (4.1)5 (6.8)5 (7.0)9 (12.2)
Nausea1 (1.4)4 (5.5)4 (5.6)9 (12.2)
Rheumatoid arthritis flare6 (8.2)0 (0.0)6 (8.5)2 (2.7)
SGPT level increased2 (2.7)2 (2.7)6 (8.5)§0 (0.0)
Infection4 (5.5)3 (4.1)2 (2.8)5 (6.8)
Diarrhea5 (6.8)1 (1.4)2 (2.8)3 (4.1)
Bronchitis3 (4.1)1 (1.4)4 (5.6)4 (5.4)
Abdominal pain2 (2.7)3 (4.1)4 (5.6)3 (4.1)
UTI1 (1.4)2 (2.7)4 (5.6)2 (2.7)
Abnormal LFTs1 (1.4)0 (0.0)1 (1.4)4 (5.4)

Five subjects experienced serious AEs during the study. There was no significant difference among the treatment groups in the incidence of serious AEs. Arterial thrombosis, abnormal liver function test, and arthritis flare occurred separately in 3 patients receiving 75 mg of ERB-041. Increased serum glutamic oxaloacetic transaminase (SGOT) and SGPT levels occurred in a patient receiving 25 mg of ERB-041. Finally, a patient receiving placebo experienced a serious AE associated with miscellaneous factors: evaluation of serum β human chorionic gonadotropin with a negative result. In addition, the presence of an AE was the most common reason for discontinuation. Fifteen ERB-041–treated subjects and 4 placebo-treated subjects were withdrawn from the study because of AEs. There were no significant differences among the treatment groups in any AE that led to withdrawal from the study, and the most common were increased SGPT and SGOT levels. No deaths occurred during this study.


This study has assessed the efficacy and safety of the highly selective ERβ agonist ERB-041 in patients with active RA despite MTX therapy. ERB-041 has been shown to exert antiinflammatory actions by selectively and potently stimulating ERβ in rheumatoid tissues (7). ERβ predominates in the synovium, articular cartilage, and cancellous bone, whereas ERα is predominant in cortical bone (11). However, both ERs are present in T cells and chondrocytes, and synovial and bone cells present in the RA joint (12). ERB-041 has been similarly effective in both the HLA–B27 transgenic rat model of IBD and the Lewis rat model of antigen-induced arthritis (AIA), with maximum response at 1 mg/kg/day administered orally. In the AIA rat model, ERB-041 rapidly improved clinical and histologic joint scores after approximately 10 days of treatment. Because of its highly selective ERβ agonism, ERB-041 lacks proliferative effects in the uterus and mammary tissues and does not inhibit ovulation or protect against bone loss in these animal models (7). However, ERβ agonists have been ineffective in other inflammatory models, including ovariectomized female mice that experienced type II collagen–induced arthritis, whereby estradiol as well as a specific ERα agonist, but not ERβ or G protein–coupled receptor 30 agonists, ameliorated the disease and associated osteoporosis (13).

Despite the encouraging preclinical data, a lack of efficacy among the 3 oral doses of ERB-041 after a 12-week course was unexpectedly observed in the present study. Neither the primary end point, the ACR20 response, nor the secondary end points were reached. Therefore, these data demonstrate for the first time that, although ERβ receptors are expressed in the synovium, ERβ agonists may have no significant effect on an inflammatory arthritis such as RA. It is unlikely that the ERB-041 failure to show efficacy was due to inadequate dosing, drug biodisponibility, or trial duration, as discussed below. However, it should be noted that the absence of positive clinical results sometimes depends on the translation of results from animal models to human disease, which must be done with caution (14).

Even though marked differences exist in their mechanisms of action, estrogen therapy provides the closest available data for ERβ agonists in RA treatment. In this regard, the Women's Health Initiative clinical trials showed no statistically significant differences in the risk of developing RA or the severity of RA between postmenopausal women that received hormone therapy replacement (HRT) and placebo (15). Nevertheless, in another study, 2 years of HRT ameliorated arthritis and bone mineral density and was associated with slower progression of radiologic joint destruction in women with active RA (16). Therefore, HRT does not appear to be associated with an increased risk of RA flares and may actually improve disease activity. Moreover, its additional risk of venous thrombosis and breast cancer is an important factor that should also be considered. In addition, a phase IIa trial has recently demonstrated that a selective ERα agonist, Org 37663, induces estrogenic effects but lacks antirheumatic activity in postmenopausal female RA patients receiving stable MTX or sulfasalazine (17). A better understanding of the connections between female hormones and RA development or progression is still critical.

The possibility that the failure of ERB-041 in showing efficacy in human subjects with active RA in this study might be due to a species-specific ligand-binding profile of ERs could not be ruled out. Therefore, ERB-041 would show a potent antiinflammatory effect in rats, but not in humans. In addition, a partial drug-specific gene regulation has been described among selective ERβ agonists. This specificity could be relevant to the distinct biologic and clinical responses that ERβ-selective compounds might originate (18).

Preliminary pharmacokinetic analysis indicated ERB-041 concentrations approximately dose proportional up to 75 mg reaching steady-state levels earlier than week 4. Furthermore, at week 12, the proportion of responder patients to ERB-041 assessed by the ACR20 criteria in the 5 and 25 mg ERB-041 groups were slightly higher than that seen in the 75 mg ERB-041 group. Also, the maximal effect of ERB-041 on DAS28 score and CRP levels was observed at an earlier time point, week 8, with no further increase by week 12 in RA patients treated with 5 mg of ERB-041 rather than the groups treated with 25 and 75 mg of ERB-041. These facts support the conclusion that the bioavailability, the dose of the drug used, and the duration of the trial were adequate to study the potential efficacy of ERB-041 in RA patients.

ERB-041 showed a good safety profile and there were no unexpected safety findings in the subject population. In fact, the most common AEs were treatment emergent and none of them were considered treatment related. Headache, nausea, infections, and bronchitis were the most frequent treatment-emergent AEs among all of the groups. Furthermore, no significant differences in the treatment-emergent AE overall incidence or in AEs that led to withdrawal from the study were found between the treatment groups. Only 5 subjects experienced serious AEs during the study, and no significant difference among the treatment groups in the incidence of serious AEs was observed. However, AEs were the most common reason for discontinuation.

Obviously, further research needs to be performed for a better understanding of the role of ERβ agonists in inflammatory diseases. Indeed, clinical trials with more potent ERβ agonists, including WAY-202196 or WAY-214156 in patients with active RA, may provide additional valuable evidence. In addition, these agonists are less selective than ERB-041, which could be beneficial because antiinflammatory properties also have been shown in in vivo models through ERα modulation (13).

In conclusion, ERB-041 was generally well tolerated, showing expected safety findings in the studied population. However, despite showing promising preclinical results, ERB-041 failed to demonstrate efficacy for the treatment of RA. These results suggest that ERβ-selective stimulation would not have positive effects on regulating inflammatory response in patients with chronic arthritis such as RA. Nevertheless, further studies are clearly warranted to establish the efficacy of selective ERβ agonists in inflammatory arthritis.


All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be submitted for publication. Dr. Herrero-Beaumont had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Cutolo, Herrero-Beaumont.

Acquisition of data. Cutolo, Herrero-Beaumont.

Analysis and interpretation of data. Roman-Blas, Castañeda, Herrero-Beaumont.


Wyeth Pharmaceuticals facilitated the study design and operation of the study, collected the data, and reviewed and approved the manuscript prior to submission. The authors independently interpreted the results, wrote the manuscript, and had the final decision to submit it for publication.


We gratefully acknowledge the contributions of the following individuals to this study: Andrew DerMovsesian, PhD, and James A. Gurr, PhD, from Wyeth Research, Collegeville, Pennsylvania; and Alan Brodsky, MD, María A. Lázaro, MD, Damon L. Bass, DO, and Laurence Paolozzi, MD. We would also like to thank María I. Nevado, MD, from Wyeth Clinical Research, Madrid, Spain, for her valuable help in the final preparation of the manuscript.