REVIEW ARTICLE Treatment of Polymyalgia Rheumatica A Systematic Review Jose´ Herna´ndez-Rodrı´guez, MD, PhD; Maria C. Cid, MD, PhD; Alfons Lo´pez-Soto, MD, PhD; Georgina Espigol-Frigole´, MD; Xavier Bosch, MD, PhD Background: Polymyalgia rheumatica (PMR) treat- lapses and shorter therapy than were lower doses; start- ment is based on low-dose glucocorticoids. Glucocorti- ing prednisone doses of 15 mg/d or lower were associ- coid-sparing agents have also been tested. Our objective ated with lower cumulative glucocorticoid doses than were was to systematically examine the peer-reviewed litera- higher starting prednisone doses; and starting predni- ture on PMR therapy, particularly the optimal glucocor- sone doses higher than 15 mg/d were associated with more ticoid type, starting doses, and subsequent reduction regi- glucocorticoid-related adverse effects. Slow prednisone mens as well as glucocorticoid-sparing medications. dose tapering (Ͻ1 mg/mo) was associated with fewer re- lapses and more frequent glucocorticoid treatment ces- Methods: We searched Cochrane Databases and sation than faster tapering regimens. Initial addition of MEDLINE (1957 through December 2008) for English- oral or intramuscular methotrexate provided efficacy at language articles on PMR treatment (randomized trials, doses of 10 mg/wk or higher. Infliximab was ineffective prospective cohorts, case-control trials, and case series) as initial cotreatment. that included 20 or more patients. All data on study de- sign, PMR definition criteria, medical therapy, and dis- Conclusions: The scarcity of randomized trials and the ease outcomes were collected using a standardized high level of heterogeneity of studies on PMR therapy protocol. do not allow firm conclusions to be drawn. However, PMR remission seems to be achieved with prednisone treat- Results: Thirty studies (13 randomized trials and 17 ob- ment at a dose of 15 mg/d in most patients, and reduc- servational studies) were analyzed. No meta-analyses or tions below 10 mg/d should preferably follow a tapering systematic reviews were found. The PMR definition cri- rate of less than 1 mg/mo. Methotrexate seems to exert teria, treatment protocols, and outcome measures dif- glucocorticoid-sparing properties. fered widely among the trials. Starting prednisone doses higher than 10 mg/d were associated with fewer re- Arch Intern Med. 2009;169(20):1839-1850 OLYMYALGIA RHEUMATICA tabolite, prednisolone, considered to be (PMR) is a syndrome char- equipotent at equivalent doses, are uni- acterized by aching and versally used in PMR. Other currently used morning stiffness in the glucocorticoids include methylpredniso- shoulder and pelvic girdles lone and deflazacort (not available in the Pand neck in persons 50 years or older.1,2 United States). Systemic manifestations such as low- grade fever, fatigue, and weight loss are fre- CME available online at quently present, as are increased acute- www.jamaarchivescme.com phase reactants including high erythrocyte and questions on page 1827 Author Affiliations: sedimentation rate (ESR), C-reactive pro- Departments of Autoimmune tein (CRP) levels, and anemia of chronic An initial prednisone dosage of 10 to and Systemic Diseases disease.1,2 20 mg/d is deemed appropriate for most (Drs Herna´ndez-Rodrı´guez, Treatment with glucocorticoids is the patients who have PMR without associ- Cid, and Espigol-Frigole´) and preferred therapy for PMR.1,2 Before the ated giant cell arteritis (GCA).1,2,6 Symp- Internal Medicine glucocorticoid era, the occasional self- toms usually resolve completely after a few (Drs Lo´pez-Soto and Bosch), limiting nature of PMR was evidenced by days. Most patients require at least 2 years Clinical Institute of Medicine spontaneous improvements in some pa- of treatment, but others have a more and Dermatology, Hospital 3,4 Clı´nic, Institut d’Investigacions tients, and musculoskeletal symptoms chronic, relapsing, or refractory course re- Biomèdiques August Pi i Sunyer were treated with nonsteroidal anti- quiring steroid treatment for much (IDIBAPS), University of inflammatory drugs (NSAIDs).3,5 Today, longer.1,2 The adverse effects of long- Barcelona, Barcelona, Spain. prednisone and its principal active me- term glucocorticoid therapy are common (REPRINTED) ARCH INTERN MED/ VOL 169 (NO. 20), NOV 9, 2009 WWW.ARCHINTERNMED.COM 1839 ©2009 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/27/2021 and sometimes deleterious in pa- ing GCA or other inflammatory ized trials.29,30 Observational stud- tients with PMR.1,2,7 To reduce the conditions), patients lost to follow-up, ies included 8 case series (5 total cumulative dose of glucocor- follow-up duration, and treatment- retrospective15-19 and 3 prospec- ticoids and their adverse effects, related adverse effects. All data were tive20-22), 3 retrospective case- some researchers have investigated reviewed and confirmed by one of us control trials,23-25 and 3 prospective (J.H.-R.). 26-28 the addition of cytotoxic drugs and, Various proposed definition criteria cohort studies (Table 2 and more recently, biologic agents with for PMR (Table 1)8-11 and the authors’ Table 3). Deflazacort was ana- 31 potential glucocorticoid-sparing ef- own criteria are noted when used. In lyzed in 1 prospective case series fects to the PMR regimen.1 studies including patients initially diag- and 3 randomized trials.32-34 Meth- To our knowledge, no reports nosed with GCA, only those patients ylprednisolone35 and 6-methylpred- have summarized the evidence for with isolated PMR were analyzed when nisolone36 were investigated in 1 ran- glucocorticoid treatment or gluco- possible. In studies with patients ini- domized trial each. Eight studies corticoid-sparing therapies in PMR. tially considered to have PMR alone who used glucocorticoid-sparing agents; The present review systematically later developed symptoms suggestive of 5 used methotrexate (3 random- analyzes the reported evidence on GCA, confirmed or not by temporal ar- ized trials,37-39 1 retrospective case- tery biopsy, the number of patients with PMR therapy, especially the prefer- control trial,40 and 1 prospective co- isolated PMR at the end of the study was 41 entially used glucocorticoid, its op- specified. hort study ); and 2 randomized timal initial and maintenance doses Methodologic quality was evalu- studies (1 each) tested azathio- and tapering regimens, and gluco- ated independently by 3 of us (J.H.-R., prine42 and infliximab43 (Table 4 corticoid-sparing agents used. X.B., and G.E.-F.). Observational stud- and Table 5). Three studies ana- ies were evaluated according to the lyzed NSAIDs.19,23,44 “Strengthening the Reporting of Obser- METHODS vational Studies in Epidemiology QUALITY AND (STROBE) statement.”12 Quality and sus- DATA SOURCES HETEROGENEITY ceptibility to bias in observational stud- OF THE STUDIES AND SEARCHES ies were appraised using the criteria rec- ommended by Sanderson et al.13 The We systematically searched the Coch- quality of randomized trials was as- All studies used different diagnos- rane Database of Systematic Reviews, sessed using the scale proposed by Ja- tic PMR criteria (Table 1), out- Cochrane Central Register of Con- dad et al.14 Disagreements on data and come definitions (eg, relapse, recur- trolled Trials, and MEDLINE/PubMed the quality of selected studies were re- rence, and disease remission) for English-language articles published solved by discussion among all authors. (Table 6), scoring systems, medi- between 1957 and December 2008, using cations and routes of administra- the MeSH term polymyalgia rheumatica tion, initial dosages, tapering sched- in combination with the terms treat- DATA SYNTHESIS ment, glucocorticoids, prednisone, pred- AND ANALYSIS ules, and length of follow-up. Most nisolone, methylprednisolone, deflaza- studies were observational, and only cort, methotrexate, azathioprine, NSAIDs, According to the type of medication used 2 randomized trials could be con- and biological therapy. References of rel- to treat PMR, we analyzed glucocorti- sidered confirmatory studies with an evant articles retrieved were searched coids, glucocorticoid-sparing agents, and appropriate sample size calcula- manually. Studies that included 20 pa- NSAIDs. Treatments for initial remis- tion.38,43 This heterogeneity did not tients or more were selected. sion induction and maintenance phases allow a pooled estimator to be cal- were examined separately. culated or statistical heterogeneity DATA EXTRACTION AND to be tested. Study designs were QUALITY ASSESSMENT RESULTS therefore considered in the follow- ing order (listed from lowest to high- Two of us (J.H.-R. and X.B.) indepen- SEARCH RESULTS est evidence quality): case series, dently read titles and abstracts search- case-control studies, cohort stud- ing for articles on medical interven- We identified 784 citations. After re- ies, and randomized trials. tions in PMR. Articles considered to meet inclusion criteria, and those with incon- trieving 163 articles, 133 were ex- clusive abstracts were fully reviewed to cluded. We finally analyzed 30 MEDICATIONS USED FOR decide on their final inclusion. Three of studies with Ն20 patients (13 ran- POLYMYALGIA RHEUMATICA us (J.H.-R., A.L.-S., and X.B.) recorded domized trials 17 and observa- the types and initial doses of gluco- tional studies) (Figure 1). No meta- Glucocorticoids corticoids and other therapies tested, analyses or systematic reviews were subsequent tapering schedules, propor- found. Of