Is a high erythrocyte sedimentation rate necessary for diagnosing polymyalgia
rheumatica?
Miguel A. Gonzales-Gay1, Vicente Rodríguez-Valverde2, Ricardo Blanco2, Jose L.
Fernandez-Sueiro2, Jose Armona2, Manuel Figueroa3 and Victor M. Martinez-Taboada2,
Division of Rheumatology, Hospital Universitario "M. Valdecilla", Facultad de
Medicina, Universidad de Cantabria (Santander) 2, Hospital Xeral-Calde (Lugo)1, and H.
Nuestra Seńora de Aránzazu (San Sebastián)3, Spain
During many years, we have observed in our clinical practice patients with typical
symptoms of polymyalgia rheumatica (PMR) and erythrocyte sedimentation rate (ESR) values
lower than 40 mm/h which is classically considered by most authors as necessary for a
diagnosis of PMR. However this observation is not new as Ellis and Ralston had reported
more than 20 years ago , that a significant proportion of their patients with PMR had a
normal ESR. Our experience along with these observations prompted us to analyze the actual
frequency and features of patients who have been diagnosed as having PMR, with a ESR lower
than 40 mm/h. Our preliminary results were later on supported by Helfgott and Kieval and
the results of our multicenter study of a large cohort of patients with PMR have been
recently published . In our study the diagnosis of PMR was established, regardless of the
ESR, in 201 consecutive Caucasian patients fulfilling all the following criteria: (a) Age
50 years or older, (b) severe proximal pain for more than 1 month in at least 2 out of the
3 following areas: neck, shoulder and/or pelvic girdles, and (c) a rapid resolution of the
syndrome (<7 days) with low dose of prednisone (10 mg/day in most of the patients) .
The possibility of GCA was excluded either by a negative temporal artery biopsy or by
absence of manifestations of GCA after a long-term follow-up (at least 18 months).
Patients already treated with corticosteroid prior to the diagnosis of PMR were not
included in the study. We then compared the demographic and clinical features of patients
with classical PMR with those patients in which the ESR at the time of diagnosis was less
than 40 mm/h.
An ESR lower than 40 mm/h was found in 41 of the 201 patients (20.4%). Almost 75% of the
patients with an ESR < 40 mm/h had previously been treated with non steroid
antiinflammatory drugs (NSAIDs)4 without significant clinical improvement. A low ESR was
significantly more frequent in males (58.5% versus 35.0%; P=.006) of younger age (66.0 +
7.5 versus 69.3 ± 8.4 years; P=.022). Furthermore, the group with an ESR < 40 mm/h had
a lower frequency of fever (5.0% versus 27.0%, P=.003), anemia (0.0 versus 19.4; P=.002),
and abnormal protein electrophoresis (17.9 versus 58.0; P<.001). In addition, there was
a trend to a lower proportion of patients with weight loss (35.9% versus 52.5%; P=.67).
Despite this somewhat less severe clinical picture PMR, the mean cumulative dose of
corticosteroids and the frequency of relapses were similar in both groups. In our series
the ESR value was unrelated to the duration of the illness prior to diagnosis. In support
of our data are those of Helfgott and Kieval, who in their series of 117 patients with
PMR3, reported a similar frequency and clinical features in cases with a low ESR.
Besides our findings, we like most authors believe that an ESR > 40 mm/h is an
important objective criterium for the diagnosis of PMR and we share the concern of some
authors5 on the possible over diagnosis of PMR in patients with a normal ESR. Therefore,
we think that in patients with a low ESR the diagnosis of PMR has to be made following
very strict guidelines. Therefore we propose that a diagnosis of PMR with a normal or only
slightly raised ESR should be made only if the patient meets the following 4 criteria:
1. Age at disease onset > 50 years
2. Bilateral moderate or severe pain during >1 month involving at least 2 of the
following areas: neck, shoulder and pelvic girdles
3. Absence of arthritis in small joints
4. Complete resolution of the syndrome in less than 7 days, with 5 mg of prednisone twice
daily, or equivalent
In summary, we think that the recognition of PMR without significantly increased ESR, can
be useful to the clinician, thus avoiding unnecessary suffering to the patient and the
high risk of ineffective NSAIDs treatment in this elderly population. However, its
diagnosis should be made following strict criteria.
Vicente Rodríguez-Valverde, MD
Professor of Medicine and Chief
Rheumatology Division
Hospital Universitado "M. Valdecilla"
Av. Valdecilla s/n
39008 Santander, Spain
