SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. REFERENCES
  8. Supporting Information

Objective

The 2010 American College of Rheumatology/European League Against Rheumatism (ACR/EULAR) classification criteria for rheumatoid arthritis (RA) refer to a possible use of ultrasound “for confirmation of the clinical findings.” We undertook this study to determine the optimized definition of ultrasound-detected synovitis for the 2010 ACR/EULAR criteria and to assess the impact of its use on the accuracy of RA classification.

Methods

One hundred nine patients with musculoskeletal symptoms for ≤3 years were enrolled in the study. Patients underwent clinical, laboratory, radiographic, and comprehensive ultrasonographic assessments at baseline and received routine management from expert rheumatologists who were blinded to the ultrasound findings.

Results

Sensitivity and specificity of the 2010 ACR/EULAR criteria using different definitions of synovitis to identify patients who developed a disease requiring methotrexate (MTX) treatment within 1 year were 58.5% and 79.4%, respectively, for clinical synovitis (tenderness or swelling), 78.0% and 79.4%, respectively, for ultrasound-detected synovitis with a gray-scale (GS) imaging score ≥1 (GS ≥1 ultrasound synovitis), and 56.1% and 93.7%, respectively, for GS ≥2 ultrasound synovitis or a synovial power Doppler (PD) signal score ≥1 (GS ≥2/PD ≥1 ultrasound synovitis). Receiver operating characteristic curve analysis for the criteria scores revealed the largest area under the curve with GS ≥2/PD ≥1 ultrasound synovitis.

Conclusion

Ultrasound assessment improves the accuracy of the 2010 ACR/EULAR criteria for identifying patients with a disease requiring MTX treatment. Our data provide preliminary but vital information for the methodology to confirm the presence of synovitis using ultrasound in the 2010 ACR/EULAR criteria.

Rheumatoid arthritis (RA) is a chronic inflammatory disease characterized by joint swelling, joint tenderness, and destruction of synovial joints, leading to severe disability and premature death (1, 2). Over the last decade, the use of disease-modifying antirheumatic drugs, in particular, methotrexate (MTX) and biologic agents, has been shown to improve the clinical outcome of RA, especially when the treatment is implemented at an early stage, before joint damage accumulates (3).

The new classification criteria for RA were developed and published in 2010 by the joint working group of the American College of Rheumatology (ACR) and the European League Against Rheumatism (EULAR) (4–6). To identify patients who would benefit from early effective intervention, factors associated with the subsequent decision by a physician to start MTX treatment within a year after presentation were extracted from cohorts of real-world patients with early arthritis (5). As a result, the final criteria consist of clinical and laboratory features characteristic of early RA and enable rheumatologists to classify a patient with early RA without typical radiographic changes much easier than by using the 1987 revised ACR criteria (7). However, the accuracy of the new criteria for identifying patients with a disease that required MTX treatment was not consistently very high among the cohorts used for validation (area under the curve [AUC] 0.66–0.82 for receiver operating characteristic [ROC] curve analysis). One possible reason for this varying accuracy of the criteria is poor reproducibility of clinical joint examination findings (i.e., swelling and tenderness) (8–10).

Ultrasound assessment is more sensitive than clinical joint examination in detecting inflammation in synovial tissues (8, 11–13). In fact, the 2010 ACR/EULAR criteria refer to a possible use of imaging techniques such as ultrasound “for confirmation of the clinical findings” (4). However, there have been no reports of validated methodology for translating ultrasound findings of synovitis on gray-scale (GS) images and power Doppler (PD) signal in the synovium into a dichotomous variable for the presence of synovitis. This study aimed to determine the optimized definition of synovitis findings on ultrasound for the 2010 ACR/EULAR criteria and to assess the impact of its use on the accuracy of RA classification.

PATIENTS AND METHODS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. REFERENCES
  8. Supporting Information

Patients.

Patients with musculoskeletal symptoms lasting for ≤3 years who were referred to the Department of Allergy and Clinical Immunology at Chiba University Hospital from January 2010 to December 2010 because of a possible diagnosis of RA were consecutively recruited to the study. Patients whose musculoskeletal symptoms were obviously explained by other diseases or whose radiographs of the hands and feet showed erosions typical of RA were excluded according to the eligibility for applying the 2010 ACR/EULAR criteria (4), whereas those with no clinically swollen joints were not excluded from enrollment in order to include patients with subclinical synovitis that could be detected by ultrasound. One hundred nine patients were enrolled and underwent clinical, laboratory, radiographic, and ultrasonographic assessments at baseline and received routine clinical management for 1 year from expert rheumatologists who were blinded to the ultrasound findings. The study design was approved by the Ethics Committee of Chiba University, and subjects' written informed consent was obtained according to the Declaration of Helsinki.

Clinical, laboratory, and radiographic assessments and management of RA.

Patient recruitment, clinical evaluation, and disease management were performed by rheumatologists certified by the Japan College of Rheumatology. Each rheumatologist made an independent decision on the management of RA according to his or her expertise. Findings of conventional radiography of the hands and feet were recorded, and the presence of erosions was assessed by a single trained rheumatologist (DN).

Ultrasound examination.

Ultrasound was performed in a temperature-controlled room on the same day as the clinical evaluation by 1 of the 6 rheumatologists trained in musculoskeletal ultrasound (DN, KI, AO, TI, YS, or KT), all of whom were blinded to the clinical information and laboratory data. A systematic multiplanar GS and PD ultrasound examination of 38 joints (see Supplementary Table 1, available on the Arthritis & Rheumatism web site at http://onlinelibrary.wiley.com/doi/10.1002/art.37848/abstract) was performed using a LOGIQ 7 Pro or a LOGIQ E9 (GE Healthcare), a Viamo or an Aplio XG (Toshiba Medical Systems Corporation), or a HI VISION Avius or a HI VISION Preirus (Hitachi Medical Corporation) instrument, depending on availability. For PD ultrasound, pulse repetition frequency was adjusted to the lowest possible value for the anatomic area scanned and for the machine used, and low wall filters were used. Color gain was set just below the level at which color noise appeared.

Ultrasound findings of synovitis on GS imaging and positive synovial PD signal were defined according to consensus definitions (14–17). We used the scoring system reported by Naredo et al (18, 19). Synovitis on GS imaging was graded semiquantitatively on a scale of 0–3 (0 = absent, and 1 = mild, 2 = moderate, and grade 3 = marked) for synovial hypertrophy of articular recess, tendon sheath, and bursa. Intraarticular, tenosynovial, and intrabursal PD signals were graded on a semiquantitative scale of 0–3 (0 = absent [no synovial flow], 1 = mild [≤3 isolated signals], 2 = moderate [>3 isolated signals or confluent signal in less than half of the synovial area], and 3 = marked [signals in more than half of the synovial area]). The maximum grades for synovitis on GS imaging and the synovial PD signal obtained from multiple synovial sites within a joint region (e.g., right second metacarpophalangeal [MCP] joint or left wrist) were recorded as GS and PD scores for the joint region. Each patient's total GS and PD scores were calculated by summing the corresponding scores of all joints. Before implementing the study, an exercise to standardize the scanning and grading methods was conducted with static images and a healthy volunteer.

Intraobserver reliability of ultrasound assessment was evaluated by randomly selecting 4 images for each joint from stored images of baseline ultrasound examination. Seventy-six images per sonographer were graded again for synovitis on GS imaging and synovial PD signal by the same sonographer under blinded conditions at the end of the study period. Interobserver reliability between sonographers was evaluated with the same sets of images (a total of 456 images) graded by the other 5 sonographers at the end of the study period.

Statistical analysis.

Statistical analysis was performed using SPSS software, version 16.0J (IBM Japan). Normally distributed continuous data were summarized with means and either SDs or 95% confidence intervals (95% CIs) and were analyzed using parametric tests (Student's t-test). Non-normally distributed data were summarized with medians and interquartile ranges (IQRs) and were analyzed using nonparametric tests (Mann-Whitney U test or Wilcoxon's signed rank test). Categorical data were summarized with percentages and were analyzed using a chi-square test, Fisher's exact test, or McNemar's test. Bonferroni correction was applied for multiple testing of partially correlated measurements. A multivariate logistic regression model was constructed using a forward stepwise method. Unless otherwise specified, P values less than 0.05 were considered significant.

RESULTS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. REFERENCES
  8. Supporting Information

Patient demographics and baseline data.

The patient demographics and clinical, laboratory, and radiographic data at baseline are shown in Table 1. Forty-eight patients (44.0%) in our study cohort did not present with any swollen joints and were not eligible for the 2010 ACR/EULAR criteria without ultrasound assessment despite other symptoms suggesting inflammatory arthritis.

Table 1. Patient demographics and clinical and laboratory data at baseline, and comparison between patients who subsequently required MTX treatment and those who did not*
Variable at baselineTotal (n = 109)No MTX treatment (n = 63)MTX treatment (n = 41)
  • *

    IQR = interquartile range; RF = rheumatoid factor; ACPA = anti–citrullinated protein antibody; ESR = erythrocyte sedimentation rate; CRP = C-reactive protein; DAS28-ESR = Disease Activity Score in 28 joints using the ESR; DAS28-CRP = DAS28 using the CRP level; HAQ DI = Health Assessment Questionnaire disability index.

  • During the 1-year followup period, 5 patients dropped out.

  • P < 0.0021 versus no methotrexate (MTX) treatment, by t-test, Mann-Whitney U test, chi-square test, or Fisher's exact test with Bonferroni correction.

  • §

    P < 0.0004 versus no MTX treatment, by t-test, Mann-Whitney U test, chi-square test, or Fisher's exact test with Bonferroni correction.

Age, mean ± SD years51.6 ± 14.850.6 ± 15.851.9 ± 13.8
Female, no. (%)85 (78)50 (79)31 (76)
At least 1 swollen joint present, no. (%)61 (56)27 (43)31 (76)
Swollen joint count, median (IQR)1 (0–4)0 (0–2)2 (0.5–5)
 Small joints, median001
 Large joints, median000
Tender joint count, median (IQR)1 (0–4.5)1 (0–3)2 (0–5)
 Small joints, median002
 Large joints, median000
RF positive, no. (%)50 (46)16 (25)30 (73)§
 Low11 (10)5 (8)5 (12)
 High39 (36)11 (18)25 (61)§
ACPA positive, no. (%)33 (30)6 (10)24 (59)§
 Low1 (1)0 (0)1 (2)
 High32 (29)6 (10)23 (56)§
Abnormal ESR, no. (%)59 (54)25 (43)32 (87)§
ESR, median (IQR) mm/hour17.5 (7–28)12.5 (5.75–26)23 (18.5–39.5)§
Abnormal CRP level, no. (%)48 (44)22 (35)23 (56)
CRP level, median (IQR) mg/dl1 (0–6)1 (0–6)3 (1–7.5)
Duration of symptoms ≥6 weeks, no. (%)106 (97)60 (96)41 (100)
Duration of symptoms, median (IQR) weeks24 (12–40)24 (12–44)20 (10–38)
DAS28-ESR, mean ± SD3.60 ± 1.383.21 ± 1.374.19 ± 1.23
DAS28-CRP, mean ± SD3.08 ± 1.262.82 ± 1.223.47 ± 1.28
HAQ DI score (0–3), median (IQR)0.5 (0.13–0.97)0.5 (0.03–0.84)0.63 (0.25–1.13)

Baseline ultrasound findings and 2 provisional definitions for ultrasound-detected synovitis.

A total of 4,142 joints were assessed by ultrasound at baseline. The findings are summarized in Table 2. The number of patients with at least 1 joint with a positive GS score was significantly larger than the number of patients with at least 1 swollen joint (88 [80.7%] versus 61 [56.0%]; P < 0.001), showing that patients without any swollen joints may have ultrasound-detected synovitis. Similarly, the number of joints with a positive GS score was larger than the number of joints with swelling or tenderness (median 4 [IQR 1–7] versus 2 [IQR 0–6], P = 0.241), although the difference did not reach statistical significance.

Table 2. Ultrasound findings at baseline, and comparison between patients who subsequently required MTX treatment and those who did not*
Variable at baselineTotal (n = 109)No MTX treatment (n = 63)MTX treatment (n = 41)
  • *

    GS = gray-scale; PD = power Doppler (see Table 1 for other definitions).

  • During the 1-year followup period, 5 patients dropped out.

  • P < 0.0005 versus no MTX treatment, by Mann-Whitney U test or chi-square test with Bonferroni correction.

  • §

    P < 0.0023 versus no MTX treatment, by Mann-Whitney U test or chi-square test with Bonferroni correction.

At least 1 joint with GS score ≥1, no. (%)88 (81)43 (68)40 (98)
At least 1 joint with GS score ≥2, no. (%)54 (50)19 (30)32 (78)
At least 1 joint with PD score ≥1, no. (%)54 (50)21 (33)29 (71)
At least 1 joint with PD score ≥2, no. (%)36 (33)12 (19)23 (56)
Joints with GS score ≥1, median (IQR)4 (1–7)2 (0–4)5 (3.5–11)§
 Small joints, median214§
 Large joints, median102§
Joints with GS score ≥2, median (IQR)0 (0–2)0 (0–1)2 (1–4.5)
 Small joints, median001
 Large joints, median000
Joints with PD score ≥1, median (IQR)0 (0–2)0 (0–1)1 (0–4)§
 Small joints, median001§
 Large joints, median000
Joints with PD score ≥2, median (IQR)0 (0–1)0 (0–0)1 (0–2)§
 Small joints, median001§
 Large joints, median000
Total GS score, median (IQR)5 (1–8.5)3 (0–6)7 (6–18.5)
 Small joints, median315
 Large joints, median102§
Total PD score, median (IQR)0 (0–3)0 (0–1)3 (0–7.5)§
 Small joints, median002§
 Large joints, median000

As shown in Table 2, a GS score ≥1 was identified in the majority of patients (88 [80.7%]), whereas a GS score ≥2 was identified in approximately one-half of the patients (54 [49.5%]). On the other hand, a PD score ≥1 was identified in approximately one-half of the patients (54 [49.5%]), whereas a PD score ≥2 was identified in fewer patients (36 [33.0%]). These data suggested that mild synovitis on GS imaging might not be specific enough to be used for confirming synovitis, while mild synovial PD signals seemed specific enough for that purpose in accordance with previous reports (20–22). Therefore, to compare ultrasound findings with clinical joint involvement, we tentatively chose 2 different definitions for ultrasound-detected synovitis. Namely, joints with a GS score ≥1 were defined as having “GS ≥1 ultrasound synovitis,” and joints with a GS score ≥2 or a PD score ≥1 were defined as having “GS ≥2/PD ≥1 ultrasound synovitis.” For a simple description, joints with swelling or tenderness (the definition of an “involved” joint in the 2010 ACR/EULAR criteria [4]) were defined as having “clinical synovitis.”

Figure 1 illustrates the difference in the proportion of joint involvement between clinical synovitis and ultrasound synovitis for each joint. As expected, the prevalence of GS ≥1 ultrasound synovitis was higher than that of clinical synovitis in most joints (Figure 1A). As shown in Figure 1B, the joints with the highest prevalence of ultrasound-only synovitis were the wrist and the knee (22.5% and 20.6%, respectively). Surprisingly, the prevalence of GS ≥1 ultrasound synovitis was significantly lower than that of clinical synovitis in the proximal interphalangeal (PIP) joints (8.8% versus 16.4%; P < 0.001) (Figure 1A), and 10.7% of PIP joints were only positive for swelling or tenderness (Figure 1B). In contrast, the prevalence of GS ≥2/PD ≥1 ultrasound synovitis was lower than that of clinical synovitis in most joints (Figure 1A). These results demonstrate that the degree and the direction of discrepancy between clinical synovitis and ultrasound synovitis could vary depending on the joint region and the definition of ultrasound synovitis.

thumbnail image

Figure 1. Prevalence of clinical and ultrasound-detected synovitis in each joint. A, Proportions of joints with clinical and ultrasound-detected synovitis. Ultrasound findings of synovitis were defined either as a gray-scale (GS) imaging score of ≥1 (GS ≥1 ultrasound synovitis; shaded bars) or as GS ≥2 ultrasound synovitis or a synovial power Doppler signal score of ≥1 (solid bars). Open bars represent clinical synovitis. B, Proportions of joints with synovitis detected only with clinical or ultrasound (GS score ≥1) assessment. ∗∗ = P < 0.00125; ∗∗∗ = P < 0.00013 by McNemar's test with Bonferroni correction. PIP (IP) = proximal interphalangeal (interphalangeal); MCP = metacarpophalangeal; MTP = metatarsophalangeal.

Download figure to PowerPoint

Intra- and interobserver reliability of ultrasound assessment.

Intra- and interobserver reliability of GS and PD scores were evaluated by calculating intraclass correlation coefficients (ICCs). As summarized in Supplementary Table 2 (available on the Arthritis & Rheumatism web site at http://onlinelibrary.wiley.com/doi/10.1002/art.37848/abstract), the ICCs for intraobserver reliability of GS and PD scores were very high (0.94 and 0.99, respectively). The ICCs for interobserver reliability of GS and PD scores were also high (0.86 and 0.89, respectively), suggesting that the ultrasound data used for analyses are reproducible and reliable. There was no significant difference in ICC between ultrasound machines.

Impact of ultrasound on the fulfillment of the 2010 ACR/EULAR RA classification criteria.

We next investigated how ultrasound-detected synovitis could change the assessment at an individual patient level. For that purpose, we analyzed in 2 steps how individual patients in our cohort were reclassified with the 2010 ACR/EULAR criteria when ultrasound-detected synovitis was used. First, we replaced the requirement for joint swelling in the classification tree by a requirement for ultrasound-detected synovitis. Second, we determined the joint count in the criteria by the presence of ultrasound-detected synovitis instead of by the presence of swelling or tenderness. Accordingly, median criteria scores were 5 (IQR 3–7) with either clinically defined synovitis or GS ≥1 ultrasound synovitis and 4 (IQR 2–6) with GS ≥2/PD ≥1 ultrasound synovitis. As shown in Table 3, a total of 85 patients (52 + 33 [78.0%]) had the same classification result at baseline whether the definition of GS ≥1 ultrasound synovitis was used or not. However, 17 patients (15.6%) who did not fulfill the criteria without ultrasound were reclassified as having RA with ultrasound. Thirteen of these 17 patients did not present with any swollen joints and had been excluded according to the clinical criteria. Interestingly, using the definition of GS ≥1 ultrasound synovitis, 7 patients (6.4%) were reclassified in the opposite direction from having RA to not having RA. When we applied the definition of GS ≥2/PD ≥1 ultrasound synovitis, even more patients (17 [15.6%]) were reclassified from having RA to not having RA. The total number of patients reclassified by using either definition of ultrasound was 24 (either 17 + 7 or 7 + 17 [22.0%]) (Table 3).

Table 3. Numbers of patients who fulfilled the 2010 American College of Rheumatology/European League Against Rheumatism RA classification criteria according to clinical or ultrasound assessment*
 Clinical synovitis
Not RARATotal
  • *

    Values are the number (%) of patients. RA = rheumatoid arthritis.

  • Defined as swelling or tenderness.

  • Ultrasound findings of synovitis were defined either as a gray-scale (GS) imaging score of ≥1 (GS ≥1 ultrasound synovitis) or as GS ≥2 ultrasound synovitis or a synovial power Doppler (PD) signal score of ≥1 (GS ≥2/PD ≥1 ultrasound synovitis).

GS ≥1 ultrasound synovitis  
 Not RA52 (48)7 (6)59 (54)
 RA17 (16)33 (30)50 (46)
 Total69 (64)40 (36)109 (100)
GS ≥2/PD ≥1 ultrasound synovitis  
 Not RA62 (57)17 (16)79 (73)
 RA7 (6)23 (21)30 (27)
 Total69 (63)40 (37)109 (100)

Clinical course and treatment.

During the 1-year followup period, 5 patients dropped out. Forty-one of the remaining 104 patients (39.4%) received MTX for the treatment of RA during the study period. The median period from baseline to MTX prescription was 15 days (IQR 0.5–142 days), and the median initial dose was 6 mg weekly (IQR 4–7 mg weekly). The time to MTX prescription was significantly shorter in patients who clinically fulfilled the 2010 ACR/EULAR classification criteria at baseline (27 [65.9%], median 6 days [IQR 0–22.75]) than in those who did not (14 [34.1%], median 150 days [IQR 13.5–304.5]) (P = 0.002). At the time when MTX was initiated, an increased number of patients (34 [82.9%]) fulfilled the criteria. One RA patient with interstitial lung disease and elevated liver enzymes was treated with sulfasalazine and additional bucillamine. Another patient who developed RA symptoms shortly after giving birth to her second child was treated with prednisolone (10 mg daily) alone because she was breastfeeding. Three patients received prednisolone (range 10–20 mg daily) for non-RA conditions (polymyalgia rheumatica, remitting seronegative symmetrical synovitis with pitting edema, and undifferentiated arthritis).

Comparisons of baseline variables between patients who received MTX and those who did not are summarized in Tables 1 and 2. Most of the ultrasound variables, particularly for small joints, showed a statistically significant difference between the 2 groups even after applying the Bonferroni correction for multiple comparisons (Table 2).

Improved accuracy of the 2010 ACR/EULAR RA classification criteria to predict development of a disease requiring MTX treatment by using ultrasound-detected synovitis.

To determine the benefit of using ultrasound-detected synovitis, we used the same gold standard used in the process of developing the 2010 ACR/EULAR criteria (5): we analyzed the association between criteria fulfillment at baseline and the proportion of patients who required MTX treatment within the first year of enrollment. As shown in Figure 2A, the proportion of patients who developed a disease requiring MTX treatment within 1 year was significantly higher in those who fulfilled the clinical criteria than in those who did not (64.9% versus 25.4%; P < 0.001). However, the difference was even larger when ultrasound-detected synovitis definitions were applied (GS ≥1 ultrasound synovitis, 71.1% versus 15.3%; P < 0.001) (GS ≥2/PD ≥1 ultrasound synovitis, 85.2% versus 23.4%; P < 0.001). It is worth noting that the incidence of developing a disease requiring MTX treatment within 1 year was higher in the patients reclassified from not having RA to having RA than in those reclassified from having RA to not having RA according to the GS ≥1 ultrasound synovitis definition (60.0% versus 14.3%; P = 0.074 by Fisher's exact test) (Figure 2B) and according to the GS ≥2/PD ≥1 ultrasound synovitis definition (100% versus 43.8%; P = 0.046) (Figure 2C).

thumbnail image

Figure 2. Fulfillment of criteria for rheumatoid arthritis (RA) at baseline and proportions of patients who required methotrexate (MTX) treatment within 1 year. A, Comparison among different definitions of synovitis used in the criteria. Clinical synovitis was defined as swelling or tenderness. Ultrasound (US) findings of synovitis were defined either as a gray-scale (GS) imaging score of ≥1 (GS ≥1 ultrasound synovitis) or as GS ≥2 ultrasound synovitis or a synovial power Doppler (PD) signal score of ≥1 (GS ≥2/PD ≥1 ultrasound synovitis). ∗∗ = P < 0.00333; ∗∗∗ = P < 0.00033 by chi-square test or Fisher's exact test with Bonferroni correction. B and C, Comparison among subgroups with different combinations of criteria fulfillment using clinical synovitis and either GS ≥1 ultrasound synovitis (B) or GS ≥2/PD ≥1 ultrasound synovitis (C). Fisher's exact tests with Bonferroni correction did not yield statistically significant results.

Download figure to PowerPoint

The sensitivity and specificity of the 2010 ACR/EULAR RA classification criteria without ultrasound in predicting a requirement for MTX treatment within the first year of classification were 58.5% and 79.4%, respectively. When the GS ≥1 ultrasound synovitis definition was applied, the sensitivity increased to 78.0%, whereas the specificity remained the same at 79.4%. Using the GS ≥2/PD ≥1 ultrasound synovitis definition, the sensitivity decreased slightly to 56.1%; however, the specificity increased markedly to 93.7%. The high specificity with the GS ≥2/PD ≥1 ultrasound synovitis definition was attributable mostly to applying the definition to the joint involvement score, whereas the high sensitivity with the GS ≥1 ultrasound synovitis definition was likely due to the synergistic effect of ultrasound application in the 2 steps (see Supplementary Table 3, available on the Arthritis & Rheumatism web site at http://onlinelibrary.wiley.com/doi/10.1002/art.37848/abstract). ROC curve analysis demonstrated the largest AUC for the criteria score with GS ≥2/PD ≥1 ultrasound synovitis (0.893 [95% CI 0.832–0.954]) followed by that with GS ≥1 ultrasound synovitis (0.868 [95% CI 0.799–0.936]) and that without ultrasound (0.844 [95% CI 0.736–0.896]) (Figure 3A). These results show that ultrasound assessment of synovitis with either definition improves the accuracy of the 2010 ACR/EULAR criteria to predict the requirement for MTX treatment and that both definitions can be useful depending on the purpose.

thumbnail image

Figure 3. Receiver operating characteristic (ROC) curves for 2010 American College of Rheumatology/European League Against Rheumatism (ACR/EULAR) RA classification criteria with clinical synovitis and ultrasound-detected synovitis to identify patients requiring MTX treatment. Shown are comparisons between the ACR/EULAR criteria with clinical synovitis and those with ultrasound-detected synovitis in all patients (n = 109) (A) and in patients whose clinical criteria scores had a range of 3–7 (n = 49) (B). See Figure 2 for other definitions. ROC analyses on other definitions are described in Supplementary Table 4, available on the Arthritis & Rheumatism web site at http://onlinelibrary.wiley.com/doi/10.1002/art.37848/abstract.

Download figure to PowerPoint

When the diagnostic performance of criteria was compared between different definitions of synovitis, no definitions provided as good a combination of sensitivity and specificity as did either GS ≥1 ultrasound synovitis or GS ≥2/PD ≥1 ultrasound synovitis (see Supplementary Table 4, available on the Arthritis & Rheumatism web site at http://onlinelibrary.wiley.com/doi/10.1002/art.37848/abstract). The AUC analysis indicates that more strict definitions (e.g., GS score ≥2 or PD score ≥2) could enable even more accurate classification if the cutoff points were lower than the original one (i.e., ≥6/10 is needed for classification). However, such criteria with a different cutoff point would not be interchangeable with the original criteria, and therefore, we only used the definitions that provided the most improved performance under the same conditions as the 2010 ACR/EULAR criteria.

We were also interested in the benefit of ultrasound in patients with equivocal arthritis, and we performed the same analyses in the 49 patients whose clinical criteria scores had a range of 3–7, which was the IQR in total patients. In this group, the difference in AUC between clinic-based and ultrasound-based criteria was more apparent (with clinical assessment, 0.457 [95% CI 0.292–0.622]; with the GS ≥1 ultrasound synovitis definition, 0.736 [95% CI 0.595–0.878]; and with the GS ≥2/PD ≥1 ultrasound synovitis definition, 0.800 [95% CI 0.673–0.927]) (Figure 3B). This result demonstrates that the clinical criteria are less informative in equivocal cases and that confirming synovitis with ultrasound is especially beneficial in such cases.

Regression models.

To determine the relative importance of ultrasound findings for each joint independent of clinical information in predicting a disease requiring MTX treatment, the presence or absence of GS ≥1, GS ≥2, PD ≥1, or PD ≥2 ultrasound synovitis for each joint was first compared between patients who received MTX treatment and those who did not. The variables with P < 0.05 in this univariate analysis in addition to the 2010 ACR/EULAR criteria score were entered into a multivariate logistic regression model as exploratory variables. As summarized in Supplementary Table 5 (available on the Arthritis & Rheumatism web site at http://onlinelibrary.wiley.com/doi/10.1002/art.37848/abstract), the exploratory variables that remained in the model in addition to the criteria score were GS ≥1 ultrasound synovitis and PD ≥2 ultrasound synovitis in the wrist. The same procedure for equivocal cases did not retain any variables, probably due to the smaller sample size.

DISCUSSION

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. REFERENCES
  8. Supporting Information

In the present study, we showed that confirming the presence of synovitis with ultrasound changed not only the prevalence of synovitis at the joint level, but also the fulfillment of the 2010 ACR/EULAR RA classification criteria in a certain proportion of patients. Furthermore, we demonstrated that criteria based on ultrasound findings may classify more accurately patients who subsequently require MTX treatment than do criteria not based on ultrasound findings. In our study, 2 provisional definitions of ultrasound-detected synovitis provided either superior sensitivity or superior specificity when used in the 2010 ACR/EULAR criteria.

The 2010 ACR/EULAR criteria, which were developed by the joint working group through a data-driven approach followed by consensus methodology, underscore the importance of synovitis, requiring at least 1 clinically swollen joint and giving the highest weight to the joint involvement (4). However, joint involvement is the most skill-dependent and irreproducible domain as compared with others, especially when joint manifestation is modest. The results of our study, which used the same gold standard that was used in the development of the original criteria, indicate that the presence of mild synovitis on GS imaging could be used to confirm the presence of synovitis for improved sensitivity of RA classification; further, the presence of moderate (but not mild) synovitis on GS imaging or abnormal synovial PD signals could be used for improved specificity. Given that rheumatologists who made the clinical decision to prescribe MTX already knew the 2010 ACR/EULAR criteria and had access to all clinical data except ultrasound findings, the superiority of ultrasound-detected synovitis over clinically detected synovitis in predicting the requirement for MTX treatment is significant.

One important finding of our study is that ultrasound is not only very sensitive for detecting synovitis, which has been demonstrated by a number of studies (8, 11, 12), but it also provides very specific assessment of synovitis as compared to clinical joint examination and can therefore exclude false-positive results, both at the joint level and at the level of the whole patient. A certain proportion of joints with apparent swelling or tenderness did not have any ultrasound-detected synovitis, depending on the joint (5.8–10.9%) (Figure 1B). Furthermore, a certain proportion of patients who fulfilled the 2010 ACR/EULAR criteria were reclassified as not having RA according to ultrasound detection of synovitis (6.4–15.6%) (Table 3), and these patients were less likely to develop disease requiring MTX treatment, even though the ultrasound information was not available to the rheumatologists (Figures 2B and C). These data suggest that joints with clinical synovitis alone are frequently false positive, and patients who fulfill the criteria with only those false-positive joints do not necessarily need immediate MTX treatment.

Filer et al (21) performed comprehensive ultrasound assessment in 58 patients with very early arthritis and showed that the ultrasound joint count (GS score ≥1) was greater than the clinical joint count in every joint region and that the ultrasound joint count improved the sensitivity of the 2010 ACR/EULAR criteria for predicting development of RA at the cost of specificity. In addition, they demonstrated by regression analysis that involvement of wrist and MCP joints on GS imaging and synovial PD signal involvement of metatarsophalangeal joints provided independent information predictive of RA development.

Although their study and ours have some results in common, there are 2 major differences. First, we considered the 2010 ACR/EULAR criteria as the most established criteria at the moment and aimed to incorporate ultrasound assessment into the criteria, whereas Filer et al (21) determined the ultrasound components that correlated with subsequent fulfillment of the 1987 revised ACR criteria, using the 2010 ACR/EULAR criteria as one model for comparison. Second, the study subjects were different. While they studied patients with symptoms lasting ≤3 months, we included patients with symptom duration up to 3 years, which was the same criterion used in the data-driven process for developing the 2010 ACR/EULAR criteria. However, we did not exclude patients with no swollen joints at presentation based on the hypothesis that such patients could also have significant synovitis that is detected only by ultrasound. Indeed, the hypothesis was correct, as we actually identified 10 patients with ultrasound-detected synovitis without any swollen joints who subsequently developed disease requiring MTX treatment. These differences in the purpose and the study subjects may account for the difference between the 2 studies in the prevalence of ultrasound-detected synovitis for each joint and in its relative importance.

Although the 2010 ACR/EULAR RA classification criteria were not meant to be “diagnostic criteria,” they have already been used as a diagnostic aid in daily rheumatology practice. Ultrasound has also been used in rheumatology practice, and the benefit of ultrasound assessment combined with the well-accepted criteria is clinically relevant. Nevertheless, although we demonstrated that confirming synovitis with ultrasound is preferentially beneficial in equivocal cases, a comprehensive ultrasound scan of 38 joints in those cases is still time-consuming and not necessarily cost-effective in daily practice. To improve the feasibility of ultrasound-assisted diagnosis, the number of joints needs to be reduced. Thus, identifying the optimal core set of joints with a high predictive value can be the next agenda. However, such analyses to identify every clinically important joint require a much larger sample size than ours, because regional distribution of synovial inflammation is highly variable among patients. Another possible strategy to reduce the number of joints is to scan joints only when the clinical findings are equivocal.

The major limitation of our study is that it was a single-center study with a small sample size. Compared with the early arthritis cohorts used in the data analysis for the 2010 ACR/EULAR criteria (5), a larger proportion of our study cohort were female (78.0% versus 66.8%), mean swollen and tender joint counts were lower (3.3 versus 6.6 and 3.3 versus 8.0, respectively), markers of inflammation were less elevated (mean erythrocyte sedimentation rate 22 mm/hour versus 32 mm/hour and mean C-reactive protein level 0.9 mg/dl versus 2.6 mg/dl), and the mean duration of arthritis symptoms was longer (8.4 months versus 4.9 months). However, most of the means of these variables were within the range of variation among the cohorts used in the phase I process for the 2010 ACR/EULAR criteria (5). The sample size of our study was large enough to validate the additional benefit of ultrasound in the 2010 ACR/EULAR criteria, but a larger number of patients would have allowed more accurate assessment of the less prevalent pathology (e.g., PD score ≥2) and of the joints less frequently involved at an early stage (e.g., shoulder).

Another limitation is the validity and generalizability of the ultrasound method used in our study. Although standardized grading methods and standardized synovial sites and joints for global assessment have not yet been reported, results of a study using such standardized methods in the future will be more readily generalizable. Taken together, a global, multicenter study that uses standardized methods for ultrasound evaluation is warranted to provide a feasible, cost-effective, and validated ultrasound methodology to confirm the presence of synovitis for the 2010 ACR/EULAR RA classification criteria. Our data present a rationale for such a large-scale study and provide preliminary but vital information for the possible definitions of ultrasound-detected synovitis and the possible target population.

AUTHOR CONTRIBUTIONS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. REFERENCES
  8. Supporting Information

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Ikeda had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Ikeda.

Acquisition of data. Nakagomi, Ikeda, Okubo, Iwamoto, Sanayama, Takahashi, Takatori, Suzuki, Takabayashi, Nakajima.

Analysis and interpretation of data. Nakagomi, Ikeda, Yamagata.

REFERENCES

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. REFERENCES
  8. Supporting Information
  • 1
    Pincus T, Callahan LF, Sale WG, Brooks AL, Payne LE, Vaughn WK. Severe functional declines, work disability, and increased mortality in seventy-five rheumatoid arthritis patients studied over nine years. Arthritis Rheum 1984; 27: 86472.
  • 2
    Scott DL, Symmons DP, Coulton BL, Popert AJ. Long-term outcome of treating rheumatoid arthritis: results after 20 years. Lancet 1987; 1: 110811.
  • 3
    Finckh A, Liang MH, van Herckenrode CM, de Pablo P. Long-term impact of early treatment on radiographic progression in rheumatoid arthritis: a meta-analysis. Arthritis Rheum 2006; 55: 86472.
  • 4
    Aletaha D, Neogi T, Silman AJ, Funovits J, Felson DT, Bingham CO III, et al. 2010 rheumatoid arthritis classification criteria: an American College of Rheumatology/European League Against Rheumatism collaborative initiative. Arthritis Rheum 2010; 62: 256981.
  • 5
    Funovits J, Aletaha D, Bykerk V, Combe B, Dougados M, Emery P, et al. The 2010 American College of Rheumatology/European League Against Rheumatism classification criteria for rheumatoid arthritis: methodological report phase I. Ann Rheum Dis 2010; 69: 158995.
  • 6
    Neogi T, Aletaha D, Silman AJ, Naden RL, Felson DT, Aggarwal R, et al. The 2010 American College of Rheumatology/European League Against Rheumatism classification criteria for rheumatoid arthritis: phase 2 methodological report. Arthritis Rheum 2010; 62: 258291.
  • 7
    Arnett FC, Edworthy SM, Bloch DA, McShane DJ, Fries JF, Cooper NS, et al. The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis. Arthritis Rheum 1988; 31: 31524.
  • 8
    Naredo E, Bonilla G, Gamero F, Uson J, Carmona L, Laffon A. Assessment of inflammatory activity in rheumatoid arthritis: a comparative study of clinical evaluation with grey scale and power Doppler ultrasonography. Ann Rheum Dis 2005; 64: 37581.
  • 9
    Salaffi F, Filippucci E, Carotti M, Naredo E, Meenagh G, Ciapetti A, et al. Inter-observer agreement of standard joint counts in early rheumatoid arthritis: a comparison with grey scale ultrasonography—a preliminary study. Rheumatology (Oxford) 2008; 47: 548.
  • 10
    Marhadour T, Jousse-Joulin S, Chales G, Grange L, Hacquard C, Loeuille D, et al. Reproducibility of joint swelling assessments in long-lasting rheumatoid arthritis: influence on Disease Activity Score-28 values (SEA-Repro study part I). J Rheumatol 2010; 37: 9327.
  • 11
    Wakefield RJ, Green MJ, Marzo-Ortega H, Conaghan PG, Gibbon WW, McGonagle D, et al. Should oligoarthritis be reclassified? Ultrasound reveals a high prevalence of subclinical disease. Ann Rheum Dis 2004; 63: 3825.
  • 12
    Brown AK, Quinn MA, Karim Z, Conaghan PG, Peterfy CG, Hensor E, et al. Presence of significant synovitis in rheumatoid arthritis patients with disease-modifying antirheumatic drug– induced clinical remission: evidence from an imaging study may explain structural progression. Arthritis Rheum 2006; 54: 376173.
  • 13
    Brown AK, Conaghan PG, Karim Z, Quinn MA, Ikeda K, Peterfy CG, et al. An explanation for the apparent dissociation between clinical remission and continued structural deterioration in rheumatoid arthritis. Arthritis Rheum 2008; 58: 295867.
  • 14
    Szkudlarek M, Court-Payen M, Jacobsen S, Klarlund M, Thomsen HS, Ostergaard M. Interobserver agreement in ultrasonography of the finger and toe joints in rheumatoid arthritis. Arthritis Rheum 2003; 48: 95562.
  • 15
    Backhaus M, Burmester GR, Gerber T, Grassi W, Machold KP, Swen WA, et al. Guidelines for musculoskeletal ultrasound in rheumatology. Ann Rheum Dis 2001; 60: 6419.
  • 16
    Wakefield RJ, Balint PV, Szkudlarek M, Filippucci E, Backhaus M, D'Agostino MA, et al. Musculoskeletal ultrasound including definitions for ultrasonographic pathology. J Rheumatol 2005; 32: 24857.
  • 17
    Schmidt WA, Schmidt H, Schicke B, Gromnica-Ihle E. Standard reference values for musculoskeletal ultrasonography. Ann Rheum Dis 2004; 63: 98894.
  • 18
    Naredo E, Moller I, Cruz A, Carmona L, Garrido J. Power Doppler ultrasonographic monitoring of response to anti–tumor necrosis factor therapy in patients with rheumatoid arthritis. Arthritis Rheum 2008; 58: 224856.
  • 19
    Naredo E, Rodriguez M, Campos C, Rodriguez-Heredia JM, Medina JA, Giner E, et al, for the Ultrasound Group of the Spanish Society of Rheumatology. Validity, reproducibility, and responsiveness of a twelve-joint simplified power Doppler ultrasonographic assessment of joint inflammation in rheumatoid arthritis. Arthritis Rheum 2008; 59: 51522.
  • 20
    Freeston JE, Wakefield RJ, Conaghan PG, Hensor EM, Stewart SP, Emery P. A diagnostic algorithm for persistence of very early inflammatory arthritis: the utility of power Doppler ultrasound when added to conventional assessment tools. Ann Rheum Dis 2010; 69: 4179.
  • 21
    Filer A, de Pablo P, Allen G, Nightingale P, Jordan A, Jobanputra P, et al. Utility of ultrasound joint counts in the prediction of rheumatoid arthritis in patients with very early synovitis. Ann Rheum Dis 2011; 70: 5007.
  • 22
    Wakefield RJ, D'Agostino MA, Naredo E, Buch MH, Iagnocco A, Terslev L, et al. After treat-to-target: can a targeted ultrasound initiative improve RA outcomes? Ann Rheum Dis 2012; 71: 799803.

Supporting Information

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. REFERENCES
  8. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
ART_37848_sm_SupplTables.doc85KSupplementary Tables

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.