Article Text
PDF
Abstract
Objective To assess the responsiveness of ultrasound (US) inflammatory findings in the shoulder and hip of patients with polymyalgia rheumatica (PMR) who started treatment with corticosteroids.
Methods Fifty-three patients with active PMR who started treatment with prednisone in six Spanish centres were prospectively studied. The patients underwent clinical, laboratory and US assessment at baseline, 4 and 12 weeks. The US investigation consisted of detection and quantification of inflammatory findings in the shoulder and hip. The responsiveness of clinical, laboratory and US parameters was tested by the standardised response mean. Intraobserver and interobserver reliability between US investigators was assessed.
Results At baseline, 34 patients (69%) had inflammation in at least one bilateral site. During the follow-up period, clinical, laboratory and US variables showed a parallel decrease. A significant decrease in US inflammatory parameters was found at week 4 (p<0.001). After 4 and 12 weeks of treatment with corticosteroids, US inflammatory findings showed similar or better sensitivity to change than clinical and laboratory markers of PMR activity. Intraobserver and interobserver intraclass correlation coefficients were 0.96 and 0.99, respectively (p<0.05).
Conclusion US may be a responsive additional tool in monitoring the response to corticosteroids in patients with PMR in daily practice and multicentre trials.
Statistics from Altmetric.com
Introduction
Polymyalgia rheumatica (PMR) is a common disease in older people characterised by proximal pain and morning stiffness, acute phase laboratory markers and a rapid response to systemic corticosteroids.1 2 Although the cause of these symptoms has remained unclear, a number of studies on imaging techniques have shown synovial intra-articular and periarticular inflammation in the proximal joints.3,–,11
MRI and ultrasonography (US) have shown bilateral subacromial-subdeltoid (SAD) bursitis as an imaging hallmark of PMR.3 6 In addition, long head biceps (LHB) tenosynovitis and glenohumeral (GH) and coxofemoral (CF) synovitis have been described in 60–85% of patients with untreated PMR.5,–,8
US is a bedside, patient-friendly, non-invasive and relatively low-cost imaging modality which has been incorporated into rheumatological practice12 in evaluating and monitoring inflammatory diseases.13 Previous studies have demonstrated a high sensitivity and specificity of US for detecting PMR inflammatory findings compared with MRI as gold standard.6 8 Some authors have therefore proposed that US findings should be used as potential diagnostic and/or classifying criteria for PMR.6 14 15 However, to our knowledge, no studies have assessed the responsiveness of US in monitoring the response to treatment in PMR.
This study aimed to assess the sensitivity to change of US inflammatory findings in the shoulder and hip joints of patients with a clinical diagnosis of PMR who started treatment with corticosteroids.
Patients and methods
From October 2005 to October 2007, 59 patients (66% women; mean±SD age 74±7 years, range 60–86) with active PMR who started treatment with corticosteroids were prospectively recruited from six Spanish rheumatology centres.
The diagnosis of PMR was made clinically by experienced rheumatologists and was based on persistent inflammatory pain and morning stiffness (lasting at least 1 month) in the neck and shoulder girdle (with or without pelvic girdle involvement) before inclusion in the study and a rapid therapeutic response to 10–20 mg/day prednisone after the baseline assessment. All patients initially included in the study were over 50 years of age and had no other clinically evident musculoskeletal disease.
The patients underwent clinical, laboratory and US assessment at baseline (within 1 day of starting corticosteroid therapy) and at 4 and 12 weeks.
Clinical and laboratory assessment
The following clinical data were collected: visual analogue scale for pain, morning stiffness duration (min), erythrocyte sedimentation rate (ESR) (normal <20 mm/h) and C-reactive protein (CRP) (normal 0–5 mg/l). We classified the ESR value as normal, slightly elevated (<40 mm/h) and highly elevated (≥40 mm/h).
US assessment
Each patient underwent US assessments throughout the follow-up period by the same rheumatologist ultrasonographer at each centre who was unaware of the clinical and laboratory findings and was not involved in the treatment decisions. The US examinations were carried out with the same real-time scanner in all centres (GE Logiq 5 Pro; General Electric Healthcare, Kyunggi-do, Korea) using multifrequency linear array transducers (7–12 MHz).
The US investigation included detection and quantification of intra-articular synovitis, tenosynovitis and bursitis in the shoulder and intra-articular synovitis in the hip according to the standardised scanning method16 17 and published reference values.18 The maximal thickness of the hypoechoic material within the synovial recesses and periarticular synovial sites was measured. SAD bursitis was considered if the maximal thickness of hypoechoic material within the bursa was >1.1 mm. GH synovitis was detected if the distance between the humeral neck convexity and the joint capsule at the axillary recess was >3.5 mm. LHB tenosynovitis was diagnosed if the distance between the tendon and the bicipital groove was >2.5 mm. Hip synovitis was considered if the distance between the femoral neck convexity and the joint capsule was >8 mm.
US reliability
Intraobserver and interobserver reliability between US investigators was assessed by two blind evaluations of 20 recorded images (without visible measurements) of the studied synovial sites from patients included in the study, randomly chosen by the principal investigator (MJP).
Statistical analysis
Statistical analyses were performed using the software Stata 10 (Stata Corporation, College Station, Texas, USA) and the SPSS statistical software Version 13.0 (SPSS, Chicago, Illinois, USA). Baseline and 4-week results were compared using the non-parametric Mann–Whitney test. Correlation between US and clinical changes throughout the follow-up period was assessed using the Spearman linear correlation coefficient. Statistical significance was accepted at p<0.05.
Intraobserver and interobserver reliability for the US variables was evaluated by calculating the intraclass correlation coefficient. Sensitivity to change of clinical, laboratory and US parameters was assessed by estimating the standardised response mean.
Results
At 12 weeks, data from 53 patients with PMR were available. Six patients were excluded from the analysis because the diagnosis was changed during the follow-up period (to rheumatoid arthritis in two patients and to pyrophosphate crystal disease in one patient) or they missed the 4-week visit (three patients).
The baseline clinical characteristics of the patients are summarised in table 1. At baseline, 34 patients (69%) had US inflammation in at least one bilateral site. The most frequent US bilateral findings were SAD bursitis (65%), LHB tenosynovitis (45%), CF synovitis (30%) and GH synovitis (18%). Figures 1–5 show representative images of the US inflammatory findings in our PMR population.
Clinical characteristics at baseline
Longitudinal scan of subacromial-subdeltoid bursitis in a patient with polymyalgia rheumatica and normal erythrocyte sedimentation rate. b, bursitis; gt, greater tuberosity; hh, humeral head; hn, humeral anatomic neck.
Transverse scan of biceps tenosynovitis in a patient with polymyalgia rheumatica. bg, bicipital groove; ts, tenosynovitis.
Longitudinal scan of biceps tenosynovitis in a patient with polymyalgia rheumatica. bg, bicipital groove.
Longitudinal scan of glenohumeral synovitis (axillary recess) in a patient with polymyalgia rheumatica. hh, humeral head; hn, humeral neck; s, synovitis.
Longitudinal scan of hip synovitis in a patient with polymyalgia rheumatica. fh, femoral head; fn, femoral neck; s, synovitis.
Twenty-one (40%) patients had ESR <40 mm/h and all but two had an increased CRP level. US bilateral inflammation was detected in 16 patients (76%) with ESR <40 mm/h. No differences in the US inflammatory findings were found between patients with normal or slightly elevated ESR and those with highly elevated ESR (data not shown).
During the follow-up period, clinical, laboratory and US parameters showed a parallel decrease. However, there was no correlation between these variables.
In the longitudinal US assessment, a significant decrease in mean US measures at the synovial sites was found at week 4 (table 2). Seventeen (50%) of the 34 patients who had baseline US bilateral inflammation became normal on US at week 4 and remained normal at week 12.
Comparison of ultrasonographic measurements at shoulder and hip synovial sites at baseline and 4 weeks
Table 3 shows the standardised response mean for the clinical, laboratory and US variables in patients with baseline US inflammation. After 4 weeks of corticosteroid therapy, LHB tenosynovitis and SAD bursitis showed similar sensitivity to change to ESR and CRP, respectively. The responsiveness of GH and CF synovitis was better than for the laboratory parameters. Sensitivity to change of GH and CF synovitis was similar to sensitivity to change of clinical variables. At 12 weeks the US inflammatory findings showed similar or better sensitivity to change than clinical and laboratory variables.
Standardised response mean for clinical, laboratory and ultrasonographic variables
Intraobserver and interobserver intraclass correlation coefficients were 0.96 and 0.99, respectively (p<0.05).
Discussion
In our study, US detected bilateral inflammation in 69% of untreated patients with PMR. In keeping with the findings of Cantini et al,10 patients with PMR with normal or slightly elevated and highly elevated ESR showed similar US inflammatory findings. Thus, US can be an extremely useful monitoring aid in patients with PMR with an unclear clinical response or clinical relapse and low ESR.
Despite the parallel evolution of clinical, laboratory and US parameters throughout the study, no correlation was found between changes in these variables as has been shown in previous studies.7 US findings seem to be a measurement of PMR activity independent of standard clinical and laboratory variables.
Salvarani et al 3 19 have previously described normalisation of PMR inflammation on MRI after local corticosteroid treatment. Koski5 reported normalisation of US joint inflammation after treatment with systemic corticosteroids in patients with PMR. We have demonstrated a significant improvement in US inflammatory findings in our PMR population after 4 weeks of corticosteroid therapy, which has been shown to be an effective treatment for this disease.20 In addition, the responsiveness of US findings was similar or even better than the responsiveness of conventional clinical and laboratory markers of PMR activity.
Our study has some limitations. We did not study a control group, and this may have introduced bias into the study. To avoid as much bias as possible, the US investigators were blinded to the clinical and laboratory findings at each visit.
Interobserver and intraobserver reliability were assessed on recorded images instead of on real-time US scanning. The absence of data on US image acquisition reliability may have overestimated our good results. However, the different geographical location of the US investigators made the performance of a real-time interobserver reliability exercise on a statistically sufficient number of patients with active PMR unfeasible. For ethical reasons, the non-treatment of patients for a period in order to re-examine them for testing intraobserver reliability was not considered.
We did not use the Doppler technique in the US assessment. In addition, we did not consider other inflammatory findings that have been described in patients with PMR such as trochanteric bursitis.8
In conclusion, we suggest that US may be a responsive additional tool in the assessment of the response to corticosteroids in patients with PMR in daily practice and multicentre trials.
Acknowledgments
The authors thank the following clinical investigators who enrolled patients into the study and recorded the clinical data. Carlos Isasi and Rodrigo Ghiglino, Department of Rheumatology Hospital Puerta de Hierro, Madrid; Ana Cruz, Department of Rheumatology Hospital Severo Ochoa, Madrid; Pilar Fernández-Dapica and Beatriz Joven, Department of Rheumatology Hospital 12 de octubre, Madrid; Virginia Villaverde and Cruz Fernández-Espartero, Department of Rheumatology Hospital de Móstoles, Madrid; Antonio Naranjo, Department of Rheumatology, Hospital Dr. Negrin, Las Palmas de Gran Canarias, Spain.
References
Footnotes
-
Ethics approval This study was conducted with the approval of the ethics committee, Hospital Universitario Puerta de Hierro Majadahonda, Madrid, Spain and written consent was obtained from all patients.
-
Provenance and peer review Not commissioned; externally peer reviewed.