Abstract
Objective. To evaluate the usefulness of the polymyalgia rheumatica (PMR) activity score (PMR-AS) in guiding adjustment of glucocorticoid (GC) dosage.
Methods. Rheumatologists prospectively included patients receiving GC therapy for PMR. At each visit, they assessed disease activity using a visual analog scale for physician’s global assessment (VASph) and recorded whether a flare was diagnosed and/or the GC dosage was changed. In each patient, the PMR-AS was calculated using the formula of Leeb and Bird: C-reactive protein (mg/dl) + VAS pain score (0 to 10) + VASph (0 to 10) + (morning stiffness in min × 0.1) + elevation of upper limbs (0–3). We evaluated the correlation between PMR-AS and GC dosage changes in the group already treated with GC.
Results. We included 89 patients (mean age 74.6 ± 6.2 yrs; disease duration 1.6 ± 2.2 yrs), who had a total of 149 visits. PMR-AS was available for 137 visits. Of those, 124 involved patients already treated with GC, and 13 patients who started GC treatment. The Spearman correlation coefficient between PMR-AS values and GC dosage change was 0.58 (p < 0.001). In the group already treated with GC, when the PMR-AS was higher than 20, GC dosages were never decreased. When the PMR-AS was between 10 and 20, GC dosages were decreased in 4 patients, unchanged in 4, and increased by < 5 mg in 4 patients. When PMR-AS was < 10, GC dosages were generally decreased.
Conclusion. The PMR-AS is helpful for diagnosing flares of PMR and may also assist in everyday practice to decide how to change the GC dosage.
Polymyalgia rheumatica (PMR) is a disease of older individuals characterized by pain and morning stiffness of the shoulder and pelvic girdles. The diagnosis is challenging because the clinical symptoms are often atypical and the range of differential diagnoses is broad1. The main symptoms are persistent bilateral inflammatory pain for > 1 month with morning stiffness in the neck, shoulders, and pelvic girdle, and laboratory evidence of systemic inflammation [erythrocyte sedimentation rate (ESR) > 40 mm/h]2,3. Active, and often passive, range of motion limitation is common at the shoulders and hips. Other manifestations include asymmetric arthritis (knee and wrist), edema of the dorsum of the hands, and carpal tunnel syndrome4,5.
The treatment of PMR rests on prednisone. Both the starting dosage (usually 10–20 mg/day) and the tapering schedule6,7,8 vary widely across physicians and patients. The regimen suggested by the British Society of Rheumatology (BSR)6 for gradual glucocorticoid (GC) tapering is daily prednisolone (or prednisone) 15 mg for 3 weeks, then 12.5 mg for 3 weeks, then 10 mg for 4–6 weeks, then reduction by 1 mg every 4–8 weeks. For the BSR, relapse is the recurrence of symptoms of PMR, justifying increasing GC to a previous higher dose.
Objective disease activity criteria would help to determine when remission has occurred and will probably be available in the near future9. However, there is also a need for objective criteria for tailoring the prednisone dosage to disease activity until remission is achieved.
In 2003, the European Collaborating Polymyalgia Group of the European League Against Rheumatism developed response criteria and an activity score (the PMR-AS) based on 5 items: morning stiffness (min), shoulder elevation (0, > 90°; 1, 90°; 2, < 90°; and 3, 0°), physician’s global assessment on a 10-point visual analog scale (VASph), pain severity on a 10-point VAS (VASp), and C-reactive protein (CRP) level (mg/dl)10,11. The score generally ranges from 0 to 50. Scores > 7 indicate somewhat active disease, scores between 7 and 17 moderately active disease, and scores > 17 highly active disease.
The PMR-AS is well suited to everyday practice, and its usefulness has been improved by the determination of cutoffs for disease remission and relapse. Thus, PMR-AS values between 0 and 1.5 suggest disease remission12 and values > 9.35 (in patients with previously controlled disease) or having increased by 6.6 or more between 2 visits suggest disease relapse13. Studies consistently found that the PMR-AS was easy to use and valid when determined by rheumatologists or general practitioners (GP)12,13,14,15, despite considerable interobserver variability in the VASph score.
We recently reported that the PMR-AS may help both GP and rheumatologists to tailor the GC dosage to disease activity15. However, we did not evaluate whether changes in the GC dosage correlated with PMR-AS values.
If the range of prescription is variable, a guideline should help physicians to standardize the GC dosage. If the tapering or increase of corticosteroid dosage is proportional to PMR-AS level, rheumatologists and GP should use this tool for GC dosage. Here, our objectives were to investigate correlations between prednisone dosage changes made at the discretion of the clinicians without using the PMR-AS, and to find the mean change of prednisone dosage that should be recommended for various levels of PMR-AS values in everyday clinical practice.
MATERIALS AND METHODS
Study rheumatologists
The patients were recruited by 26 rheumatologists working with the Inflammatory Joint Disease Working Group of the French Society for Rheumatology (Club Rhumatismes et Inflammation). The rheumatologists recruited consecutive patients meeting the inclusion criteria.
Inclusion criteria
Patients were eligible if they were taking GC therapy for PMR. Because no specific diagnostic criteria are available, we defined PMR for our study as a diagnosis of PMR by a rheumatologist in the absence of other diseases that might mimic PMR.
Study questionnaire
Age and date of PMR diagnosis were recorded for each patient. At each visit, the rheumatologist recorded the PMR-AS items (VASp, VASph, morning stiffness, upper limb elevation, ESR, and CRP); whether the patient had synovitis, limb girdle pain, fever, nocturnal awakenings due to PMR, and/or weight loss; whether a flare was diagnosed; and whether the prednisone dosage was changed. By definition, a flare is a change from inactive to active disease. For our study, a flare was diagnosed if the patient experienced either the first bout of active disease (transition from health to symptomatic disease) or a transition from controlled to uncontrolled disease while on prednisone therapy, in the opinion of the rheumatologist.
PMR-AS
We calculated the PMR-AS as described by Leeb and Bird10:
The rheumatologist neither calculated nor used the PMR-AS during our study.
Statistical analysis
We computed the Spearman correlation coefficient to evaluate the correlation between PMR-AS and the prednisone dosage change between 2 visits. To illustrate the prednisone dosage changes compared to PMR-AS values, we categorized prednisone dosages in increments of 5 mg/day and PMR-AS changes in increments of 5 points.
RESULTS
The 26 participating rheumatologists recruited 89 patients (mean age 74.6 ± 6.2 yrs, disease duration 1.6 ± 2.2 yrs), who had a total of 149 visits. On those visits, 137 had PMR-AS available, including 13 first visits and 124 followup visits.
Rheumatologists generally started the treatment at 15 mg but on occasion at higher or lower dosages, according to the disease activity. We found a good correlation between changes of prednisone dosage and PMR-AS score values in the whole group (r = 0.58; p < 0.001; Figure 1).
Correlation between polymyalgia rheumatica activity score (PMR-AS) and change of prednisone dosage (mg/day) when the treatment was started and during the followup.
Table 1 shows changes of prednisone dosage according to PMR-AS values in the group of 124 patients. When the PMR-AS was not greater than 5 (71 visits), the prednisone dosage was usually decreased (59 visits) and was less often left unchanged (11 visits) or increased by < 5 mg/day (1 visit); it was never increased by > 5 mg/day. When the PMR-AS value was between 5 and 10 (30 visits), the prednisone dosage was either decreased (19 visits), unchanged (9 visits), or increased by < 5 mg/day (2 visits). When the PMR-AS value was between 10 and 20, the prednisone dosage was also decreased (4 visits), unchanged (4 visits), or increased (4 visits) by < 10 mg/day. When the PMR-AS was > 20 (11 visits), the prednisone dosage was usually increased by > 5 mg/day and less often left unchanged (3 visits including 2 with PMR-AS > 20). Both patients having PMR-AS higher than 20 for whom the prednisone dosage was unchanged had a discordance between VASp (> 5) and VASph (< 5) and a normal CRP. Another pathology explained the pain, and so the discordance.
No. patients experiencing change of prednisone dosage (in steps of 5 mg/day) by change of PMR-AS.
Table 2 shows the means of prednisone dosage changes in patients previously treated with GC, tapered below a score of 15 and increases of 2.6, 5, and 10 mg if PMR-AS was above 15, 20, and 25, respectively.
Means of prednisone dosage changes in patients previously treated with glucocorticoids.
DISCUSSION
PMR affects older patients, who are at high risk for side effects of GC therapy. Accurate tailoring of the GC dosage to the needs of each patient may decrease the occurrence of diabetes mellitus, vertebral fractures, femoral neck fractures, hip fractures, and cardiovascular disorders12,13.
In everyday practice, rheumatologists usually diagnose PMR and start prednisone therapy. Then the GP provides followup and tapers the prednisone dosage, in the absence of clear guidelines. A simple tool to help decide how and when to adjust the prednisone dosage to disease activity would be of considerable usefulness to both rheumatologists and GP. The effectiveness of the PMR-AS in diagnosing disease flares has been established in many studies12,13. We recently showed that the PMR-AS may help both rheumatologists and GP to tailor the prednisone dosage to disease activity15. However, we did not evaluate whether the prednisone dosage change correlated with the PMR-AS value. The correlation shown here between the prednisone dosage changes and the PMR values suggest that the PMR-AS may indeed help rheumatologists and GP determine exactly how much prednisone each patient needs at a given time.
According to the BSR, relapse is the recurrence of symptoms of PMR, justifying increasing GC to previous higher doses. In our study, the prednisone dosages used for each 5-point increase in PMR-AS values varied somewhat, albeit within a narrow range. Our results suggest that this BSR tapering schedule may be appropriate when the PMR-AS value is not > 10 but that a stable dosage may be in order for PMR-AS values of 10 to 15, and that increased dosage may be in order otherwise. The following increases may be appropriate: PMR-AS 15 to 20, 2.5 mg/day; PMR-AS 20 to 25, 5 mg/day; and PMR-AS > 25, 10 mg/day. However, because some patients having a PMR-AS > 20 had another pathology to explain why that rheumatologist did not increase the prednisone, GP should be encouraged to refer patients to rheumatologists in cases of doubt, peculiarly when CRP level is normal.
Our study has limitations. First, only patients in France were included, and they were evaluated by French rheumatologists. Second, the number of patients remains too low (particularly for the patients with PMR-AS > 10) to conclude that it is useful to tailor the GC dosage in everyday practice in this subgroup of patients. Third, the lack of longterm followup hinders evaluation of the effect of the variable dosages.
Our results suggest that the PMR-AS may help to tailor the GC dosage in patients with PMR who are seen in everyday practice. Using the PMR-AS during followup as described would improve practice uniformity. Dosage tapering as recommended by the BSR should be started again when the PMR-AS falls below 10. A cohort study would be welcome to investigate whether dosage adjustment based on the PMR-AS value at each visit significantly diminishes the cumulative prednisone dosage from diagnosis to full recovery.
Acknowledgment
We are indebted to the following physicians for their participation in the study: J.P. Benais, L. Beraneck, A. Besson, B. Boyer, D. Caplanne, J.P. Chabellard, P. Chicault, J.F. Debarge, A.E. Dumel, E. Gibert, P. Leca, V. Lemaire, P. Legros, B. Loze, F. Magnan, J. Marcou, J.C. Marquestaut, S. Melac-Ducamp, Q. Vuong Dang, E. Sornay-Rendu, J. Taillandier, and C. Visticot. We thank the following organizations for participating in our study: Association des Rhumatologues Parisiens, Association des Rhumatologues de l’Essonne, Association des Rhumatologues du Val de Marne, Association des Rhumatologues du Finistère, Association des Rhumatologues du Val d’Oise, Club Rhumatismes et Inflammation, Fédération Française de Rhumatologie, and Société de Rhumatologie de l’Ouest.
- Accepted for publication September 20, 2011.
