Objectives To investigate whether children with juvenile idiopathic arthritis (JIA) in clinical remission have subclinical synovial disease on ultrasound, and whether ultrasound abnormalities predict an early flare of synovitis.
Methods Thirty-nine consecutive children who had clinically defined inactive disease (ID) for a minimum of 3 months underwent ultrasound assessment of 52 joints. All joints were scanned for synovial hyperplasia, joint effusion, power Doppler (PD) signal and tenosynovitis. Patients were then followed clinically for up to 2 years until a flare of synovitis occurred in one or more joints, or until the 2-year visit if the disease remained in clinical remission.
Results Synovial hyperplasia, joint effusion, PD signal and tenosynovitis in at least one joint were detected in 76.9%, 66.7%, 33.3% and 15.4% of patients, respectively. During the 2-year follow-up, 24 patients (61.5%) experienced sustained ID, whereas 15 patients (38.5%) had a flare of synovitis in a total of 45 joints after a median of 10.6 months (range 6.3–13.7 months). At study entry, the rate of synovial hyperplasia, joint effusion and tenosynovitis was comparable between patients with persistent ID and patients with synovitis flare, whereas patients with persistent ID had a greater frequency of PD signal than patients with synovitis flare. Only 17 of the 45 flared joints had ultrasound abnormalities at study entry.
Conclusion The authors found that ultrasound-detected synovial abnormalities are common in children with JIA in clinical remission. However, the presence of ultrasound pathology did not predict an early flare of synovitis in the affected joints.
Statistics from Altmetric.com
The achievement of clinical remission is the ultimate aim for paediatric rheumatologists treating children with juvenile idiopathic arthritis (JIA).1,–,3 With the recent shift towards early aggressive interventions, and the development of new therapeutic agents,4,–,6 many children with JIA are likely to experience disease remission. These advances in therapeutic effectiveness make it important to ensure that the methods used for assessing disease activity are accurate.
A reliable definition of inactive disease is critical in determining when antirheumatic therapy can be stopped. In addition, a high frequency of early flares after discontinuation of methotrexate or biological medications for clinical remission has been reported in children with JIA.7,–,10 This suggests that residual inflammation may persist in the presence of clinically undetectable disease activity. Although recent data obtained with the use of immunological biomarkers are promising,10 so far no consistent predictors of disease flare in patients discontinued from drug therapy after the achievement of clinical remission are available.
The current method to assess clinical remission in JIA is based on the so-called Wallace criteria,1 ,11 which are composed of clinical and laboratory variables. However, these measures have the limitation of not directly measuring inflammation at the primary site of pathology, and may be subject to confounding influences. Imaging techniques, such as ultrasound and MRI, are capable of directly visualising and objectively quantifying synovial inflammation (Lanni et al, unpublished data).12 There are increasing data to support their validity as disease assessment tools, and their superior sensitivity compared with clinical evaluation for detecting inflammation. Recent studies in children with JIA have shown a high frequency of ultrasound-detected subclinical synovitis in clinically normal joints.13,–,19 Furthermore, evidence of ongoing synovial disease on both ultrasound and MRI was found in a sizable proportion of patients in clinical remission.20 ,21 However, it is unknown whether imaging-detected changes in individual joints predict an early relapse of clinically overt synovitis in the same joints.
The primary aim of the present study was twofold. First, to investigate the prevalence of ultrasound-detected subclinical synovitis in JIA patients with clinically defined inactive disease, and second, to evaluate whether the presence of ultrasound abnormalities predicts an early flare of synovitis. A secondary aim was to examine ultrasound findings in healthy children.
Patients and methods
All consecutive patients who met the International League of Associations for Rheumatology (ILAR) criteria for JIA22 attended the rheumatology outpatient clinic of the Department of Pediatrics of the Fondazione IRCCS Policlinico San Matteo of Pavia, Italy, between January 2009 and April 2010, and who had inactive disease for a minimum of 3 months were included in the study. Informed consent was obtained from all children, parents or guardians, as appropriate. The study protocol was approved by the local institutional review board.
Clinical and laboratory assessment
In each patient, the state of inactive disease was defined according to Wallace criteria,1 as the presence of no joints with active arthritis, no fever, rash, serositis, splenomegaly or generalised lymphoadenopathy attributable to JIA, no active uveitis, normal acute phase reactants, and a physician global assessment indicating no disease activity.
Joint assessment was performed by two paediatric rheumatologists (SMM and CV) who had 10 and 4 years of clinical rheumatology practice, respectively, were blinded to ultrasound findings, and reached consensus on the presence/absence of articular signs in each joint. Seventy-three joints were assessed for swelling, tenderness/pain on motion and restricted motion according to a standard technique.23 A joint with active arthritis was defined as a joint with swelling or, if no swelling was present, with tenderness/pain on motion and restricted motion.
The following data were recorded for each patient through clinical chart review: sex, age at disease onset and study entry, disease duration, ILAR category, antinuclear antibody status, course of joint disease, JIA outcome measures assessed as reported23,–,26 at study entry, and medications received in the 6 months prior to study entry and at study entry.
At study entry, all patients underwent an ultrasound assessment, which was performed separately, immediately after the clinical evaluation, by an experienced rheumatologist (SR) with over 9 years of experience in musculoskeletal ultrasonography. The ultrasound examiner was blinded to clinical findings. A total of 52 joints (2 elbows, 2 wrists, 10 metacarpophalangeal, 10 proximal interphalangeal, 2 knees, 2 ankles, 2 subtalar, 2 tarsal, 10 metatarsophalangeal and 10 foot interphalangeal) were scanned for the presence of synovial hyperplasia, joint effusion, power Doppler (PD) signal and tenosynovitis. The ultrasound assessment required 25–40 min, depending on the child's age and cooperation. The ultrasound examination was performed with Logiq 9 (General Electric Medical System, Milwaukee, Wisconsin, USA) equipped with an 8–15 MHz volumetric probe (4D16L) and linear probe (9L).
Synovial hyperplasia was defined as an abnormally hypoechoic joint space, distinct from the intra-articular fat pad and non-compressible with the transducer. Joint effusion was detected as the presence of an abnormally anechoic space within the joint, which was compressible. PD signal was evaluated only inside the area of synovitis and was considered positive in the presence of vessel dots on PD images. Tenosynovitis was defined as the presence of hypoechoic or anechoic thickened tissue with or without fluid within the tendon sheath, which was seen in two perpendicular planes and could exhibit PD signal. In each joint, synovial hyperplasia, joint effusion, PD signal and tenosynovitis were recorded as present or absent. The first three features were also graded on a 0–3 scale, as previously reported.15 The threshold used for synovial abnormalities was 1. All ultrasound findings were interpreted using both longitudinal and transverse planes. The ultrasound examination technique as well as definitions and scoring of ultrasound features were based on published guidelines, particularly those provided by the OMERACT.27,–,30
Information regarding assessment of inter-observer and intra-observer reproducibility of ultrasound evaluation has been reported previously.15
Assessment of disease course over time
After the clinical and ultrasonographic assessments, all patients were followed-up for up to 2 years with periodic clinical evaluations every 3 months. Based on the course of joint disease over time, patients were classified in two groups: (1) patients who had persistently inactive disease until the 2-year visit; and (2) patients who experienced a flare of synovitis. A flare of synovitis was defined as a recurrence of active arthritis, defined as above, in one or more joints. The ultrasound was not repeated in patients who had a clinical flare.
Thirty-nine healthy children, 23 boys and 16 girls, aged 3.5–19 years (median 10.2 years, IQR 8.7–14 years), who did not have musculoskeletal complaints, underwent an ultrasound assessment of the same joints scanned in children with JIA. The ultrasound was performed by the same ultrasonographer, who followed the same methodology used in children with JIA. The ultrasonographer was not blinded to the fact that these children were healthy controls.
Descriptive statistics were reported as medians and IQR for continuous variables and as absolute frequencies and percentages for categorical variables. Comparison of demographic, clinical and ultrasound features between patients groups was made by the Mann–Whitney U test in case of quantitative variables, and by the χ2 or Fischer exact test, as appropriate, in case of qualitative variables. The statistical package used was Stata 10 (StataCorp, College Station, Texas, USA).
A total of 39 patients, 4 boys and 35 girls, were included in the study. Eighteen patients had persistent oligoarthritis, 12 had extended oligoarthritis, 4 had rheumatoid-factor negative polyarthritis, 2 had psoriatic arthritis, 2 had enthesitis-related arthritis and 1 had systemic arthritis. The median age at disease onset was 3.5 years (IQR 1.9–6.6 years), the median disease duration and age at study entry were 4.3 years (IQR 2.3–8.4 years) and 11.9 years (IQR 7.3–15.1 years), respectively. At study entry, 10 and 13 patients had clinical remission on or off medications, respectively, according to the Wallace criteria.1 ,11 The total number of JIA patients followed at the study centre was 82. Around 40% of them had inactive disease at the last follow-up visit.
In the 6 months before enrolment, 14 patients had received no therapy, whereas patients numbering 17, 12, 4 and 1 were given disease-modifying antirheumatic drugs (DMARDs), non-steroidal anti-inflammatory drugs (NSAIDs), intra-articular corticosteroids, or etanercept, respectively. At study entry, 22 patients were receiving no therapy, whereas patients numbering 15, 3 and 1 were receiving DMARDs, NSAIDs or etanercept, respectively. The main clinical features of the study patients are presented in table 1.
Ultrasound findings at study entry
The frequency of ultrasound abnormalities in at least one joint among the 39 patients enrolled was 76.9% for synovial hyperplasia, 66.7% for joint effusion, 33.3% for PD signal and 15.4% for tenosynovitis. Of the 2028 scanned joints, 131 (6.5%) had synovial hyperplasia, 79 (3.9%) had joint effusion, 30 (1.5%) had PD signal and 10 (0.5%) had tenosynovitis; 139 joints (6.9%) had one or more synovial abnormality. Overall, ultrasound abnormalities were seen most frequently in the knees, followed by the wrist, ankle, metatarsophalangeal and tarsal joints (table 2).
A distinctive high frequency of PD signal was seen in the wrists, which accounted for 43.3% of the 30 joints with this ultrasound abnormality. Tenosynovitis was detected only in the wrist, proximal interphalangeal, ankle and foot interphalangeal joints.
Evaluation of disease course over time
During the 2-year follow-up, 24 patients (61.5%) had sustained inactive disease, whereas 15 patients (38.5%) experienced a flare of synovitis in a total of 45 joints after a median of 10.6 months (range 6.3–13.7 months).
Comparison of demographic and clinical features between patients with persistently inactive disease and patients with synovitis flare
As shown in table 1, patients with persistently inactive disease were younger at disease onset and study entry and had received or were receiving DMARDs or biological medications less frequently in the 6 months before enrolment and at study entry, respectively, than did patients with synovitis flare. The two patient groups were comparable for sex ratio, disease duration at study entry, distribution of ILAR categories, antinuclear antibody status, articular course and values of JIA outcome measures at study entry.
Ultrasound findings at study entry in flared joints
Only 17 (37.8%) of the 45 joints which flared during follow-up had ultrasound abnormalities at study entry. The most common abnormalities were joint effusion (88.2%) and synovial hyperplasia (76.5%), whereas PD signal was detected in only one joint, while no joint had tenosynovitis (Table 3).
Of the 45 relapsed joints, 37 were previously affected clinically, whereas 8 were newly affected.
Comparison of ultrasound findings between patients with persistently inactive disease and patients with synovitis flare
The frequency of ultrasound abnormalities at study entry in patients with persistently inactive disease and synovitis flare is presented in figure 1. The frequency of synovial hyperplasia, joint effusion and tenosynovitis was comparable between the two groups, whereas patients with persistent ID had a greater frequency of PD signal than patients with synovitis flare. figure 2 shows the frequency of ultrasound abnormalities in specific joints in patients with persistent inactive disease and synovitis flare. Overall, there was a greater frequency of ultrasound abnormalities in patients with persistent ID than in those with synovitis flare, with the sole exception of a greater frequency of synovial hyperplasia and joint effusion in the knee in patients with synovitis flare than in those with persistent inactive disease.
Results of ultrasound assessment in joints previously affected clinically
Before the achievement of inactive disease, the 39 patients enrolled had a total of 282 joints involved clinically: 56 (19.9%) of these joints had ultrasound abnormalities. At study entry, a total of 139 joints had ultrasound abnormalities: 60 (43.2%) of these joints were previously affected clinically, whereas 79 (56.8%) were not.
Frequency of ultrasound abnormalities in healthy children
Fourteen of the 39 healthy children (35.9%) were found to have ultrasound abnormalities in at least one joint, which were joint effusion in 10 children, synovial hyperplasia in 5 children, and tenosynovitis in 1 child; PD signal was not detected. The involved joints were the knee (n=9), first metatarso-phalangeal (n=8), ankle (n=1) and tarsal (n=1) joints. All ultrasound abnormalities in healthy children were graded as 1.
Our study showed that a sizable proportion of children with JIA who were classified as having inactive disease clinically had ultrasound evidence of ongoing synovial pathology in one or more joints. The most common ultrasound-detected abnormalities were synovial hyperplasia and joint effusion, which were seen in 76.9% and 66.7% of patients, respectively. One-third of patients displayed positive PD signal and 15.4% had tenosynovitis. Overall, synovial changes were most common in the knee, ankle and wrist joints. As found in a previous study,15 the frequency of positive PD signal was distinctly greater in the wrist joint.
Our results are in keeping with those reported by Rebollo-Polo and co-workers20 in 28 children with JIA who had been in clinical remission for a minimum of 3 months. In their series, the frequency of grayscale abnormalities (synovial hypertrophy or joint effusion) in joints previously affected or unaffected clinically was 57.1% and 50%, respectively, for the wrist, and 40% and 12.5%, respectively, for the ankle. The correspondening percentages for positive PD signal were 21.4% and 0% for the wrist, and 6.7% and 0% for the ankle. None of the knees previously affected or unaffected clinically had grayscale pathology or positive PD signal.
Taken together, our findings and those of Rebollo-Polo et al indicate that there is an important discrepancy between clinical and ultrasound findings, suggesting that ultrasound may be too sensitive. However, although ultrasound may not be suitable as ‘blinded’ outcome measure for clinical trials, it should be available in paediatric rheumatology centres as it is very helpful for diagnostic purposes, and to guide intra-articular injections.12 To investigate whether the presence of ultrasound abnormalities have a prognostic value, particularly in predicting the risk of early synovitis flare, we followed-up on all enrolled patients for up to 2 years, censoring their follow-up when a flare of synovitis occurred in one or more joints, or at the 2-year visit if the disease remained in clinical remission. Fifteen (38.5%) patients had a flare of synovitis after a median of 10.6 months, whereas 24 (61.5%) patients experienced continued clinical remission until the 2-year visit. When we compared the frequency of ultrasound abnormalities at study entry between patients with continued remission and patients with flare of synovitis, we found that the rate of synovial hypertrophy, joint effusion and tenosynovitis was comparable in the two groups, and that the rate of positive PD signal was greater in children with continued remission. These findings suggest that the ultrasound demonstration of ongoing synovial pathology in JIA patients with clinically defined inactive disease does not predict an early flare of synovitis. The lack of predictive value of ultrasound was reinforced by the observation that only 17 of the 45 joints which flared during follow-up had ultrasound abnormalities at study entry.
None of the four individual ultrasound features examined in our study revealed any prognostic value. The poor predictive role of greyscale abnormalities is not surprising as it is known that synovial thickening and joint effusion in patients with chronic arthritis may represent residual findings that do not necessarily reflect ongoing active disease.32 Joint effusions have been reported in healthy children who had no synovial inflammation.33 In our healthy sample, joint effusion and synovial hyperplasia were found in 25.6% and 12.8% of subjects, respectively. Notably, detection of synovial hypertrophy in children is more challenging than in adults as the synovial tissue is often difficult to distinguish from the hypoechoic cartilage of epyphyses.34
The lack of predictive value of positive PD signal was unexpected; however, as this ultrasound feature, which is a sign of increased synovial vascularisation, is considered a more reliable indicator of active synovitis than the grayscale technique alone.34,–,37 As opposed to greyscale abnormalities, this ultrasound feature was not detected in healthy children evaluated in our study and in other investigations.15,33 In adult patients with RA, vascularisation detected by PD has been found to predict short-term disease flare after clinical remission and radiographic progression.29 ,31 ,32,–,40 In children with JIA, positive PD signal has been seen most commonly in the wrist joint, which constitutes a distinctly vulnerable site of structural damage in childhood arthritis.41 However, it has been warned that the presence of juxta-articular flow at colour Doppler examination in the growing child may either represent normal flow of the well-vascularised cartilage of the epiphysis or synovial hyperemia indicating inflammation.34 Furthermore, the appraisal of low-grade PD signal in growing children may be confounded by the physiologically enhanced synovial blood flow.42 Notably, patients with sustained remission, who had a greater prevalence of positive PD signal, were younger than patients with synovitis flare. All but two joints that displayed positive PD signal in our sample were assigned the lowest score of 1. This problem underscores the need for defining the threshold grading for PD signal that should be considered pathological in children with chronic arthritis.
There are several limitations to this study, particularly relating to the poor standardisation of musculoskeletal ultrasound in children, and the lack of normative data regarding the impact of different ages, stages of growth and puberty.12 The OMERACT guidelines may not be appropriate as they were developed for adults. We did not validate the presence of synovial abnormalities with other imaging techniques, such as MRI. Furthermore, we did not use age-related reference values for cartilage thickness to distinguish cartilage from synovial fluid, a task known to be difficult in growing children. The heterogeneity of JIA categories may weaken our study as some subtypes are more likely to relapse than others. Only one patient had received biologics, which means that our results do not apply to remission induced by these medications. Another limitation is the lack of assessment of the shoulder, hip and temporomandibular joints. In conclusion, ultrasound-detected synovial abnormalities were common in our JIA patients in clinical remission. However, their presence did not predict an early flare of synovitis in the affected joints. Larger prospective studies are needed to establish the diagnostic and prognostic values of ultrasound assessment in childhood arthritis.
Competing interests None.
Patient Consent Obtained.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.