Etanercept added to background methotrexate therapy in patients with rheumatoid arthritis: Continued observations




To observe the long-term safety and efficacy of combination therapy with etanercept and methotrexate in patients with rheumatoid arthritis (RA), and to determine whether the addition of etanercept allowed reductions in methotrexate or corticosteroid dosages while maintaining a clinical response.


Patients with RA who received methotrexate plus etanercept in a previous randomized, placebo-controlled trial were offered the opportunity to enroll in an open-label extension study for further evaluation of treatment with etanercept and methotrexate.


Seventy-nine of the 89 patients in the original blinded study enrolled in the extension study, and 65 of these patients continue in the study. Patients have received etanercept therapy for up to 47 months (median 44 months). The types and rate of adverse events noted during the extension trial were similar to those observed in the controlled trial. At the 3-year assessment, 77% of the 57 patients available for evaluation met American College of Rheumatology 20% (ACR20) criteria for improvement in RA, 47% met the ACR50 criteria, and 23% met the ACR70 criteria. Of the 36 patients assessed at 3 years in the extension study, 30 (83%) were able to decrease their dosages of corticosteroids, and 20 (56%) were able to discontinue corticosteroid therapy. At 3 years, the dosage of methotrexate was decreased in 41 of 66 patients (62%), and methotrexate therapy was discontinued in 19 patients (29%).


In this observational continuation study, the addition of etanercept to background methotrexate provided sustained clinical benefit over a median period of 44 months. With etanercept therapy, there were trends toward dosage reduction or discontinuation of methotrexate and corticosteroids, without apparent worsening of ACR response rates. Compared with the controlled trial, no increases in the rate of adverse events were observed.

Over the past several decades, methotrexate has become increasingly recognized among rheumatologists as the gold standard in treating patients with rheumatoid arthritis (RA) (1, 2). Compared with other, more traditional disease-modifying antirheumatic drugs (DMARDs), methotrexate has a more rapid onset of action, superior long-term effectiveness, and fewer cumulative toxicities (3–6). Studies show that 80% of patients discontinue other DMARDs after 2 years, while at least 50% continue to take methotrexate for >5 years (7–9). However, despite its advantages over traditional agents, methotrexate is associated with toxicity, and many patients—even with aggressive treatment—do not achieve adequate therapeutic responses (10).

Recently, a body of evidence has emerged that demonstrates the importance of more aggressive treatment of patients with early disease and those with persistent symptoms (11, 12). Newer data suggest that radiographic evidence of joint damage is apparent even in patients with very early disease (11–15). Therefore, it is important to consider implementation of disease-modifying drugs at earlier stages of disease. In addition, it has been shown that despite treatment with traditional DMARDs, patients with persistent disease continue to deteriorate radiographically and lose functional capacity (13–15). This finding is especially important considering that studies have demonstrated a direct correlation between declining functional status and increased rates of morbidity and mortality (16–20). Previous studies have shown that both methotrexate and etanercept (Enbrel; Immunex, Seattle, WA) have the ability to slow radiographic progression in addition to providing positive effects on functional status (13, 21). Therefore, the combination of etanercept plus methotrexate may be an important and beneficial treatment modality for patients with RA.

In a previous 6-month, randomized, double-blind, placebo-controlled study in patients with persistently active RA despite ongoing methotrexate therapy, etanercept added to background methotrexate was a well-tolerated and effective treatment and was significantly better than methotrexate alone (22). However, questions remained regarding whether the success observed in this earlier study could be sustained over longer treatment periods, and whether toxicities would emerge. We undertook the present open-label extension study to determine the longer-term safety and clinical benefit of using etanercept plus methotrexate, and to determine whether this addition allowed for a dose reduction or discontinuation of methotrexate or corticosteroids while maintaining a clinical response.



Patients eligible for the study had active RA (functional class I, II, or III as defined by the American College of Rheumatology [ACR] revised criteria) (23), despite at least 6 months of methotrexate therapy at a dosage of 15–25 mg per week (with a stable dosage administered for a minimum of 4 weeks). Upon completion of the controlled trial, all patients were given the opportunity to receive etanercept and methotrexate in an open-label extension trial (without interruption in therapy). Both the original randomized, double-blind trial of etanercept versus placebo given in addition to methotrexate therapy and the long-term extension study were approved by the institutional review board at each of the centers, and all patients gave written informed consent before study entry.


Patients were assessed monthly during the controlled trial and every 3–4 months during the extension trial. Evaluations included measurement of vital signs, physical examinations, hematology and chemistry profiles, urinalysis, determination of adverse events, and assessment of the dosages of methotrexate and corticosteroids. After 12 months in the extension trial, only clinically significant adverse events were required to be reported. Reports of injection site reactions were not routinely collected during the extension study, because their incidence, mild intensity, and association with etanercept have been well documented in this and other previous controlled trials. Measures of disease activity were recorded at each visit and included tender and swollen joint counts, physician's and patient's global assessments, visual analog scale for pain, disability as measured by the Health Assessment Questionnaire (HAQ) (24), and the C-reactive protein level.


Patients who received methotrexate plus etanercept in the controlled trial continued to receive the same treatment in the extension trial. Patients who received methotrexate plus placebo in the controlled trial received etanercept and methotrexate upon entering the extension trial. Patients received etanercept, 25 mg subcutaneously, twice weekly and methotrexate, orally or subcutaneously, weekly. After 12 weeks of etanercept therapy in the open-label study, dose reductions of methotrexate and corticosteroids were allowed at the discretion of each investigator. If an investigator noted a subjective increase in a patient's disease activity after dosage reduction, the dosages of methotrexate or corticosteroid were returned to previous levels.

Statistical analysis.

Adverse events per patient-year were calculated as the total number in each event category divided by total time in the extension study (summed over the 79 patients enrolled in the extension study). Total time in the study was computed in days and then divided by 365.25 to obtain patient-years. Efficacy data were collected over time in both the blinded and extension studies and include time receiving etanercept only. Improvements were calculated from baseline values at the start of etanercept therapy. Changes in the dosage of prednisone or methotrexate at 3 years (relative to the dosages at baseline of the extension study) were assessed statistically, using Wilcoxon's signed rank test.


During the controlled trial, 59 patients were originally randomized to receive methotrexate plus etanercept. Of these patients, 2 withdrew due to adverse events unrelated to etanercept, and 4 patients elected not to enter the extension trial. In addition, 26 of the 30 patients who were originally randomized to receive methotrexate plus placebo entered the extension trial and received etanercept plus methotrexate. Of the 4 patients who did not enter the extension phase, 2 discontinued treatment because of lack of efficacy, 1 discontinued due to an adverse event, and 1 patient was lost to followup. Thus, 85 of the original 89 patients were eligible, and 79 of them enrolled in the extension trial and received treatment. Patient distribution is shown graphically in Figure 1.

Figure 1.

Patient distribution from original blinded study into the extension trial.

Efficacy analyses at 1, 2, and 3 years relative to the start of etanercept therapy include those patients who were evaluable and whose visits fell within the 3-month window for each time point. Because analyses of methotrexate and prednisone tapering were performed using larger (1-year) windows relative to baseline of the extension study, the number of patients evaluated for efficacy and the number evaluated for tapering of other drugs differed. By design, patients who received placebo during the randomized study received etanercept for shorter periods, because their exposure to etanercept began during the extension trial.


Baseline demographics and clinical characteristics of the patients who entered the extension trial are presented in Table 1. At baseline in the original blinded trial, the median age of patients was 50 years. Most patients had long-standing RA, with a median duration of 12 years. The mean dosage of methotrexate was 17.5 mg/week, and 44 of 79 patients were receiving concomitant corticosteroid therapy (mean dosage of prednisone 6.4 mg/day).

Table 1. Baseline demographic and clinical characteristics of 79 patients who entered the extension trial*
  • *

    At baseline in the randomized, placebo-controlled trial. DMARD = disease-modifying antirheumatic drug; NSAID = nonsteroidal antiinflammatory drug.

Age, years 
 Mean (median)49 (50)
Sex, % women87
Race, % 
 African American5
Rheumatoid factor positive, %87
Duration of rheumatoid arthritis, years 
 Mean (median)14 (12)
No. of tender joints 
 Mean (median)28 (27)
No. of swollen joints 
 Mean (median)21 (18)
No. of DMARDs used 
 Mean (median)2.9 (3.0)
Concomitant therapy, % 
  Mean dosage (range 1.3–10 mg/day)6.4 mg/day
  Mean dosage (range 7.5–25 mg/wk)17.5 mg/week


In this ongoing extension trial, 79 patients with RA who began the study while receiving background methotrexate have now received treatment with etanercept for a median of 44 months (mean 38 months), representing a total exposure time of 250 patient-years. During the study, 14 patients withdrew for reasons that included subject refusal (n = 5), lack of efficacy (n = 3), adverse event (n = 2), physician decision (n = 2), death (n = 1), and other (n = 1). At the time of this analysis, 76 (96%) of the 79 patients were evaluable for 1-year efficacy, 63 (80%) were evaluable for 2-year efficacy, and 57 (72%) were evaluable for 3-year efficacy.


Previous observations have demonstrated that etanercept plus methotrexate was well tolerated (22). In this open-label extension trial, combination therapy continued to be well tolerated, with most patients experiencing only mild adverse events. Compared with the frequency of any type of adverse event during the controlled study, the frequency was not significantly increased when patients received long-term etanercept and methotrexate (Table 2).

Table 2. Adverse events observed in at least 5% of patients*
Adverse eventControlled trialOpen-label extension trial
Placebo plus MTX (n = 30)Etanercept plus MTX (n = 59)Etanercept plus MTX (n = 79)
  • *

    Values are the number of events per patient-year. MTX = methotrexate.

  • First year of the extension trial only; after 12 months, only significant adverse events were required to be reported.

Abdominal pain0.080.110.09
Increased cough0.410.080.03
Mouth ulcer0.330.080.04
Accidental injury0.250.080.04
Bone pain0.1600.01

Infectious events occurred at a rate similar to that in the previous controlled study. Four patients experienced infections that required hospitalization, a rate of 0.02 infections per patient-year that did not increase with more prolonged exposure. All patients recovered and continued to participate in the trial.

The rate of serious adverse events did not increase with extended exposure to etanercept and methotrexate therapy. One death was reported (a 68-year-old woman with a history of hypertension died of myocardial infarction). During the study, cancer was diagnosed in 3 patients. In 1 patient, symptoms of parotid gland non-Hodgkin's lymphoma developed after 140 days of combination therapy; this patient discontinued participation in the study, was treated with radiation therapy, and has had no further evidence of disease. In another patient (with a 10–20–pack-year smoking history), squamous cell carcinoma of the larynx developed after he received therapy for 491 days; after tumor resection and radiation therapy, the patient resumed therapy with etanercept and has had no further symptoms of disease. In the third patient, breast cancer was diagnosed after 787 days of therapy; this patient temporarily discontinued participation in the study in order to undergo a mastectomy and chemotherapy, but later resumed therapy with etanercept. No patient developed systemic lupus erythematosus or any other additional autoimmune rheumatic disease.

Clinical benefit.

The clinical benefit of etanercept was sustained for the duration of therapy in patients who added etanercept to background methotrexate therapy. ACR criteria for improvement in RA (25) were achieved throughout the study. Of the 57 patients assessed at 3 years, 44 (77%) had achieved ACR 20% improvement, 27 (47%) had achieved ACR 50% improvement, and 13 patients (23%) had achieved ACR 70% improvement (Figure 2). Patients had a median of 27 tender and 18 swollen joints at baseline and a median of 3 tender and 6 swollen joints at the 3-year assessment (Figure 3). Additionally, a substantial number of patients achieved complete resolution in individual disease activity measures during extended etanercept therapy. Among the 57 patients, 11 (19%) had no tender joints, 6 (11%) had no swollen joints, and 10 (18%) had HAQ disability scores of zero.

Figure 2.

American College of Rheumatology 20% improvement (ACR20), ACR50, and ACR70 rates over time. Patients were not always available for every evaluation. Decreasing numbers represent patients who discontinued the study or did not reach a particular time point.

Figure 3.

Median number of tender and swollen joints during the study. Tender and swollen joint counts decreased rapidly from baseline of the blinded study and remained stable over time. Patients were not always available for every evaluation. Decreasing numbers represent patients who discontinued the study or did not reach a particular time point.

Reduction of corticosteroids and methotrexate.

During the controlled trial, patients were not allowed to alter the dosages of any background medications. During the extension trial, after patients had received combination therapy for at least 3 months, corticosteroid and methotrexate dosages could be reduced at the discretion of the investigator. At baseline in this extension study, 44 of the 79 patients were receiving corticosteroids (mean dosage 6.4 mg/day). Thirty of the 36 patients (83%) assessed after 3 years in the extension study were able to decrease or discontinue their dosage, 20 (56%) were able to discontinue corticosteroid therapy, and 1 patient (3%) increased the dosage of prednisone (Figure 4). All 79 patients were receiving methotrexate at baseline (mean dosage 17.5 mg/week). Among 66 patients in the extension study who were assessed at 3 years, 41 (62%) were able to decrease or discontinue the dosage of methotrexate, 19 (29%) discontinued methotrexate therapy, and 5 patients (8%) increased the dosage of methotrexate (Figure 5). Among patients who decreased the dosage or discontinued background corticosteroids or methotrexate, the response to etanercept was maintained.

Figure 4.

Mean dosages of prednisone over time. ∗ = Determined by Wilcoxon's signed rank test.

Figure 5.

Mean dosages of background methotrexate over time. ∗ = Determined by Wilcoxon's signed rank test.


The previous controlled trial and this extension trial were the first studies to evaluate the combination of a tumor necrosis factor (TNF) inhibitor and full-dose methotrexate. This observational study demonstrates that during longer-term use of etanercept in patients receiving background methotrexate, the clinical benefit is maintained. Furthermore, although patients received open-label etanercept during the extension trial, which may have influenced their perceptions, there was no evidence of an increase in the frequency of adverse events during the first year. Patients who were taking methotrexate or corticosteroids were able to significantly reduce the dosages of these therapies when they were given etanercept, further decreasing the potential for toxicities and enhancing the benefit-to-risk ratio. Although no data were collected regarding the effect of such medication tapering on the clinical response, overall ACR response rates appeared to be sustained (Figure 2).

Despite the advantages of methotrexate over other traditional agents, many patients receiving methotrexate experience at least some degree of toxicity, and its use requires monitoring (26). The most common adverse events in the pre–folate supplementation era included nausea, vomiting, and mucous membrane ulceration, which were usually dose-related (10). Rare but serious adverse effects such as hepatotoxicity, pneumonitis, opportunistic infections, and development of certain types of malignancies (e.g., non-Hodgkin's lymphoma) have been associated with methotrexate therapy but are not clearly dose-related (26–29). Therefore, the ability to discontinue methotrexate may decrease the potential for toxicities. Twenty-nine percent of the patients in this study discontinued methotrexate when etanercept was added to their therapy.

Corticosteroids are also commonly used by patients with RA because of their well-established effects on the signs and symptoms of disease. Corticosteroids, like methotrexate, are associated with toxicities that are most often dose-related. These toxicities include altered glucose metabolism, suppression of the hypothalamic–pituitary axis, immunosuppression, osteoporosis, skin fragility, atherosclerosis, and cataract development. Reducing the dosage of corticosteroids is important, because even small decreases can lead to significant safety benefits (30). Of the patients in this study who were receiving concomitant corticosteroids, 83% decreased the dosage or discontinued use of corticosteroids when etanercept was added to the background methotrexate.

Both methotrexate and etanercept have been demonstrated to reduce radiographic progression in RA (13, 21). Improvement in radiographic progression may lead to further long-term benefits such as a reduction in joint arthroplasty and improved physical function. The effect of the combination of methotrexate and etanercept on radiographic progression is being studied.

TNF is a proinflammatory cytokine with a pivotal role in the pathogenesis of RA (31, 32). Because evidence exists that TNF plays an important role in the immune response, there is some concern regarding the potential for etanercept to enhance the rate of occurrence of infection. Other functions of TNF, such as its role in tumor surveillance, are less well defined and raise hypothetical concerns about the development of malignancies (31, 32). We observed no increase in these toxicities throughout the substantial treatment intervals reported here. Similarly, clinical trials using etanercept as monotherapy have not demonstrated an increased risk of serious infection with up to 43 months of therapy (33). To date, no increase in malignancy has been observed in the patients receiving etanercept, with patients now in their fifth year of use (34). Further observation is necessary to fully evaluate this risk.

In conclusion, the addition of etanercept therapy in patients receiving background methotrexate provides sustained clinical benefit and continues to be well tolerated with extended use. Although data on clinical response were not assessed coincident with tapering of methotrexate and corticosteroids, patients showed a trend toward lower dose requirements, thereby decreasing exposure and the potential for toxicities. Although these patients continued to have some evidence of active disease, ACR response rates were sustained.


Drs. Kremer, Bankhurst, Bulpitt, Fleischmann, and Jackson have received support from Immunex Corporation. Dr. Weinblatt has received research support from and served as an ad hoc consultant to Immunex Corporation