Treatment of refractory polymyalgia rheumatica with etanercept: An open pilot study




To investigate whether etanercept has a steroid-sparing effect in the treatment of patients with relapsing polymyalgia rheumatica (PMR).


The study group comprised patients with relapsing PMR who were not able to reduce their prednisone dosage below 7.5–10 mg/day and who had experienced corticosteroid-related side effects. Patients received injections of etanercept 25 mg twice weekly for 24 weeks, and were followed up for 3 additional months after treatment withdrawal. Patients regularly underwent clinical assessment, measurement of erythrocyte sedimentation rate and C-reactive protein level, and ultrasound (US) examination of the shoulders during the 9 months of the followup period.


All 6 enrolled patients responded to etanercept with sustained remission (improvement of at least 70% according to European League Against Rheumatism response criteria for PMR in 4 patients and at least 50% in 2 patients) and were able to significantly reduce their median prednisone daily dosage without experiencing a disease relapse (8.75 mg versus 2.5 mg; P = 0.026) at the end of the 9-month study period. US shoulder examination performed at the end of followup demonstrated a parallel reduction of glenohumeral and periarticular inflammation. A significant reduction in the cumulative prednisone dose 9 months before versus the 9-month study period was observed (mean ± SD 1,767 ± 524 mg versus 730 ± 182 mg; P = 0.028). Three patients developed nonsevere side effects: bacterial cystitis in 2 and influenza in 1.


These results, which should be confirmed in a controlled study, suggest that etanercept may be a safe and useful corticosteroid-sparing agent in relapsing PMR.


Corticosteroids are the drug of choice in treating polymyalgia rheumatica (PMR). An initial dosage of 10–20 mg/day of prednisone or equivalent is adequate in most cases, in the absence of associated giant cell arteritis (GCA) (1, 2). A treatment course of 1–2 years is often required. However, a subset of patients tend to have a more prolonged and relapsing course, which may require corticosteroid treatment for several years (3–6). Corticosteroid-related adverse events are frequently observed during treatment course. Previous investigations have demonstrated that of patients with PMR who have been treated with prednisone, 65% developed at least 1 adverse event (7). Corticosteroid-related side effects are strictly related to the cumulative steroid dose (7). Cytotoxic drugs, in particular methotrexate, have been used as glucocorticoid-sparing agents with conflicting results (8–11).

Etanercept is a recombinant human tumor necrosis factor (TNF) dimeric receptor fusion protein, which consists of the extracellular portion of two p75 receptors fused to the Fc portion of IgG1. This agent has been demonstrated to be effective and safe in the treatment of rheumatoid arthritis (RA), ankylosing spondylitis, or psoriatic arthritis (12–18).

Two pilot studies conducted by our group and others have reported the efficacy of anti-TNF therapy in patients with corticosteroid-resistant PMR (19, 20). In a series of patients with corticosteroid-resistant GCA, a condition strictly related to PMR, infliximab was able to induce a sustained disease remission (21). However, to date no data have been published evaluating the efficacy of other anti-TNFα agents, in particular etanercept, in patients with corticosteroid-resistant PMR. In this pilot study, we used etanercept to treat 6 patients with longstanding PMR and corticosteroid-related side effects whose disease could be controlled only by using corticosteroid daily dosages ≥7.5 mg.



Six patients with longstanding PMR who were followed at the rheumatologic unit of Reggio Emilia Hospital, Italy, were included in the study group. At diagnosis, all 6 patients satisfied the criteria for PMR defined by Healey (22). Rheumatoid factor was negative in all patients. Anti–cyclic citrullinated peptide antibodies were tested in 3 patients and were in the normal range (<20 units/ml).

The patients received a starting prednisone dosage of 12.5 mg/day and were regularly followed for a median of 49 months (range 48–96 months). After symptoms had been in remission for 1 month, the dosage was reduced to 10 mg/day. Small decrements of 2.5 mg every 4 weeks were successively scheduled until the lowest effective maintenance dose was reached. All 6 patients had relapsed every time prednisone dosage had been reduced to 7.5–10 mg/day and all of them had also experienced different corticosteroid-related adverse events (Table 1).

Table 1. Clinical and demographic characteristics of the patients
PatientAge, yearsSexPeripheral involvementDisease duration, monthsNumber of relapsesCorticosteroid-related adverse events
178M954Osteoporotic fractures, cataract
281F403Diabetes, hypertension
372F573Diabetes, osteoporotic fractures
463F363Osteoporotic fractures
576F242Hypertension, diabetes
680FHand extremity swelling with pitting edema182Osteoporotic fractures, hypertension

Treatment protocol.

The study protocol was reviewed and approved by the local ethics committee. Inclusion criteria were the following: relapsing PMR, at least 12 months prednisone treatment, presence of corticosteroid adverse events, and inability to reduce prednisone dosage below 7.5 mg/day. Exclusion criteria were as follows: 1) clinical and/or histologic evidence of GCA; 2) fulfillment of the 1987 revised American College of Rheumatology (formerly the American Rheumatism Association) criteria for RA (23); 3) presence of uncontrolled diabetes and hypertension, infection, and neoplasm including multiple myeloma (24); and 4) presence of active or inactive (latent) tuberculosis (TB), evaluated by detailed medical history (including personal history of TB, possible previous contacts with TB, and family history of TB), chest radiographs (performed in the 2 months prior to study entry), and tuberculin purified protein derivative test. Before entering the trial each patient was informed of the nature, duration, and purpose of the study, as well as all the potential benefits and drawbacks that could be expected. All participants gave their written informed consent. The protocol duration was 9 months. The patients received 48 subcutaneous injections of etanercept 25 mg twice weekly for 24 weeks and were subsequently followed up for 3 additional months.

The prednisone dosage was reduced after the first 4 weeks to 5 mg/day if a clinical remission was attained. At week 12 prednisone was reduced to 2.5 mg/day if clinical remission persisted.

In case of incomplete response, the prednisone dosage was maintained at 5 mg/day. At the 24-week visit prednisone was withdrawn in case of complete remission, whereas in case of incomplete response the prednisone dosage was maintained at 2.5 mg/day.

Patient evaluation.

Patients were evaluated for disease activity variables at baseline (T0; at the time of first injection), 6 weeks after the first injection (T1; at the time of the thirteenth injection), 12 weeks after the first injection (T2; at the time of the twenty-fourth injection), 24 weeks after the first injection (T3; at the time of etanercept withdrawal), and 36 weeks after the first injection (T4; end of followup), and in case of disease relapse/recurrence. At every scheduled visit and in case of relapse/recurrence, the following clinical variables were evaluated: pain (0–10-cm visual analog scale [VAS]), duration of morning stiffness (minutes), patient global assessment (0–10-cm VAS), medical global assessment (0–10-cm VAS), validated Italian version of the Health Assessment Questionnaire disability index (25), and shoulder abduction (0–3 scale) (26). Laboratory investigations were as follows: erythrocyte sedimentation rate (ESR), C-reactive protein (CRP) level measured by nephelometry (NA latex CRP kit; Behringwerke, Marburg, Germany; upper limit of the normal reference range: 5 mg/dl), complete blood count, and liver and kidney function tests.

At each visit the percentage of each patient's response according to the European League Against Rheumatism (EULAR) response criteria (27) and a Disease Activity Score (DAS), according to the criteria suggested by Leeb and Bird (26), were calculated.

The disease was considered to be in remission when all of the following were fulfilled: absence of systemic symptoms (fever, malaise, anorexia, weight loss), morning stiffness, girdles and neck pain, and peripheral synovitis plus normal values of CRP (<0.5 mg/dl) and ESR (<40 mm/hour). The disease was considered to have relapsed when all of the following were present: typical proximal musculoskeletal symptoms, morning stiffness ≥1 hour, ESR ≥40 mm/hour, and/or CRP level 0.5 mg/dl.

Patients were evaluated for any adverse events occurring during the study period. In addition, at each visit the following variables were assessed and recorded: complete physical examination, vital signs, concomitant therapy, and premature discontinuation.

Ultrasound examination.

Ultrasound (US) examination was performed at baseline and at the end of followup or in the case of disease flare. Shoulders were examined with multiple high-frequency probes (5–13 MHz) for the detection of subacromial bursitis (semiquantitative scale 0–3), glenohumeral synovitis (semiquantitative scale 0–3), and long biceps tendon tenosynovitis (semiquantitative scale 0–3) according to a standardized protocol (28).

Statistical analysis.

Wilcoxon's test for paired data was used to compare continuous variables.

Index patient history.

A 70-year-old man developed gradual onset of aching and morning stiffness in the shoulder and hip girdles. At the same time he became anorexic with a weight loss of 5 kg, and his ESR was 75 mm/hour. A diagnosis of PMR was made. Prednisone 12.5 mg/day was started with resolution of pain within 72 hours.

The prednisone dosage was gradually tapered, but the patient could not be weaned off the steroids because of disease relapses at a prednisone dosage lower than 7.5–10 mg/day. Three years after the onset of steroid treatment he developed bilateral cataracts.

Eight years later the patient developed multiple dorsal vertebral fractures despite risedronate and calcium. The cumulative prednisone dosage at that time was 25.5 gm. Clinical examination showed hypotrophic skin with easy bruising, moon face, hypotrophic appendicular muscles, and accentuated dorsal kyphosis.

We decided to start anti-TNF therapy in an attempt to control the disease and reduce the daily steroid dose. Prior to the first injection of etanercept, CRP level and ESR were normal, global pain (VAS) was 70 mm, Leeb's PMR DAS score was 28.1, and daily prednisone dosage was 10 mg.

US examination of the shoulders showed bilateral distension of the subacromial bursae and of the long biceps tendon sheath. Synovial fluid could also be detected in the right glenohumeral joint.

After 6 weeks of etanercept treatment we were able to reduce the patient's prednisone dosage to 5 mg/day and after 12 weeks to 2.5 mg/day without disease relapse and with persistently normal ESR and CRP level. At week 24 etanercept was withdrawn and at the last visit at week 38 the patient was still in clinical remission (Leeb's DAS score 5.55, CRP level 0.13 mg/dl, ESR 24 mm/hour) with a prednisone dosage of 2.5 mg/day. US shoulder examination showed absence of subacromial bursa and/or long biceps tendon sheath distensions. During the 9-month study period the cumulative prednisone dosage was 1,085 mg compared with 2,730 mg in the 9 months before etanercept administration. No adverse events were observed during the treatment period.


The clinical and demographic data of the 6 patients who were enrolled in the study are shown in Table 1. All 6 patients were able to reduce their daily prednisone doses without relapses (Tables 2 and 3). One patient stopped prednisone at the end of etanercept treatment. Five patients had a Leeb's DAS score <7 at the end of the treatment period, indicating low disease activity.

Table 2. Clinical and laboratory course during the 6-month etanercept treatment*
PatientTime 06 monthsAdverse events
ESR, mm/hourCRP, mg/dlLeeb's DAS (26)PDN daily doseESR, mm/hourCRP, mg/dlLeeb's DAS (26)PDN daily dose
  • *

    ESR = erythrocyte sedimentation rate; CRP = C-reactive protein; DAS = Disease Activity Score; PDN = prednisone.

5130.5424.610120.193.512.5Bacterial cystitis
6321.5619.17.5352.5610.312.5Bacterial cystitis
Table 3. Main outcome measures as percentage variation (Δ): end of the study versus baseline*
  • *

    Values are the percentage unless otherwise indicated. HAQ = Health Assessment Questionnaire; EULAR = European League Against Rheumatism; see Table 2 for additional definitions.

  • Disease activity score according to Leeb (26).

  • EULAR response criteria (27).


Four patients had a sustained improvement of at least 70% according to the EULAR response criteria for PMR (27) and 2 patients had a sustained improvement of at least 50% at the end of the 9-month study period. One patient continued not taking prednisone at the end of the study period, while the other 5 maintained a sustained improvement with very low daily doses of prednisone (median 2.5 mg, range 1.25–2.5 mg). The patient who suspended prednisone relapsed 4 months after the end of the study period but again attained remission with prednisone 5 mg/day. US shoulder examination at the end of followup showed an almost complete resolution of glenohumeral and periarticular inflammation (Table 4, Figure 1).

Table 4. Ultrasound shoulder examination during the 6-month etanercept treatment*
PatientTime 06 months
Subacromial bursa, right/leftLong biceps tendon sheath, right/leftGlenohumeral joint, right/leftSubacromial bursa, right/leftLong biceps tendon sheath, right/leftGlenohumeral joint, right/left
  • *

    − = no distension; + = minimal accumulation of fluid (>1 mm); ++ = moderate distension without stretching peripheral structure; +++ = significant accumulation of fluid to stretch the walls of the structure.

Figure 1.

Top, right subacromial bursa distension before etanercept treatment (arrow). Bottom, marked reduction of the subacromial bursal inflammation after etanercept treatment (arrow) in a patient with polymyalgia rheumatica (patient 6).

The median daily dosage of prednisone during the 9-month study period decreased from 8.75 mg to 2.5 mg, the difference being statistically significant (P = 0.026). The cumulative prednisone dose of the 9-month period before the onset of etanercept therapy was significantly higher than that of the 9-month study period (mean ± SD 1,767 ± 524 mg versus 730 ± 182 mg; P = 0.028). Etanercept was well tolerated: 3 patients developed nonsevere side effects. Two adverse events were bacterial cystitis and an episode of influenza.


The present study encompassed a series of 6 patients with isolated PMR refractory to steroid therapy. Relapses are not uncommon in patients with isolated PMR and are seen in approximately 30–50% of patients (29–31). Relapses are generally observed when prednisone dosage is <7.5 mg/day (30). Speed of tapering (30) and genetic susceptibility (31) have been proposed to influence the risk of relapses. In addition to frequent relapses, patients from our series required an inappropriately high prednisone dosage, greater than 7.5–10 mg/day, leading to steroid-related side effects. Due to this, etanercept therapy was considered.

The results of this open pilot study suggest that etanercept could be useful as a steroid-sparing agent in PMR. Even if only 1 of our 6 patients was able to stop corticosteroid treatment, all patients maintained good disease control with a significant reduction in daily prednisone dose. US examination showed that etanercept treatment also reduced the inflammatory lesions of periarticular and articular structures.

All patients who entered the study had experienced severe corticosteroid-related side effects, particularly multiple osteoporotic vertebral fractures. Morbidity in patients with PMR is mainly related to major steroid-related complications, the most common being fractures (7). Methotrexate has been proposed as a corticosteroid-sparing agent in PMR, but the data are conflicting (8, 9). Only 2 randomized placebo-controlled studies have evaluated the efficacy of methotrexate in patients with PMR when administered with oral prednisone (10, 11). In the largest of these 2 studies, a corticosteroid-sparing effect of methotrexate was observed after 1 year of treatment; however, the methotrexate group did not have less steroid-related morbidity (11).

TNFα, which is released by macrophages and activated T lymphocytes, plays a major role in inflammatory response. The use of TNFα blockade in patients with PMR was thought to be justified based on the demonstration of elevated circulating levels of this cytokine in patients with PMR (32) and the successful treatment of chronic inflammatory arthropathies (12–18). Furthermore, 2 recent pilot studies from our group demonstrated the efficacy of infliximab as a steroid-sparing agent in patients with steroid-resistant PMR and GCA (19, 21). By contrast, one 24-week, double-blind, placebo-controlled, multicenter study failed to demonstrate a steroid-sparing effect of infliximab in patients with newly diagnosed PMR (33).

Previous studies have demonstrated that interleukin-6 (IL-6) is persistently elevated in the subgroup of patients with PMR with relapsing disease (29). Anti-TNF agents are able to down-regulate the production of IL-6 in RA cell synovial culture systems in vitro (34) and to reduce IL-6 serum levels in treated patients with PMR (19). We speculate that anti-TNF therapy might be more active as a steroid-sparing agent in patients with relapsing disease because of its capacity to suppress the incomplete corticosteroid-induced IL-6 inhibition.

Even if we cannot arrive at definitive conclusions from our study because of the limited number of participating patients and because of the open-label design, our encouraging results suggest that a larger placebo-controlled study should be performed to establish the efficacy of etanercept as a corticosteroid-sparing drug in patients with PMR with relapsing disease.


Dr. Salvarani 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 design. Catanoso, Macchioni, Boiardi, Pipitone, Salvarani.

Acquisition of data. Catanoso, Macchioni.

Analysis and interpretation of data. Catanoso, Macchioni, Boiardi, Salvarani.

Manuscript preparation. Catanoso, Macchioni, Boiardi, Pipitone, Salvarani.

Statistical analysis. Macchioni.