Usefulness of Power Doppler Ultrasound for Prediction of Re-Therapy With Rituximab in Rheumatoid Arthritis: A Prospective Study of Longstanding Rheumatoid Arthritis Patients

Authors


Abstract

Objective

To assess the value of gray-scale (GS) and power Doppler (PD) ultrasound (US) in detecting inflammatory/destructive changes and for prediction of necessity of re-therapy with rituximab (RTX) in patients with rheumatoid arthritis (RA) over 1 year of followup.

Methods

GSUS and PDUS were performed to assess synovitis, tenosynovitis, and erosions on the clinically dominant hand and forefoot of 20 patients with RA before and after therapy with RTX. US parameters were compared with clinical (Disease Activity Score in 28 joints, tender/swollen joint counts, and patients' visual analog scale of disease activity) and laboratory parameters (C-reactive protein level and erythrocyte sedimentation rate). Results were compared for patients with and without re-therapy with RTX.

Results

Significant decreases in clinical and laboratory parameters were observed after 6 and 12 months. US synovitis scores significantly decreased after 6 and 12 months (P < 0.05 for each). Regarding patients who received re-therapy between 6 and 9 months after the start of therapy (n = 9), a fair therapy response was still detectable before re-therapy. In these patients, PD-positive synovitis was the only parameter that increased up to the 6-month examination. All patients negative for rheumatoid factor and anti–cyclic citrullinated peptide (n = 4) were in the group of patients receiving a second course of treatment. Seropositive patients showed a better response to treatment with less need for re-therapy.

Conclusion

Response to therapy was measurable by clinical and laboratory parameters as well as by US. Since PDUS was able to detect the onset of disease activity before worsening of clinical symptoms occurred, PDUS is most helpful in evaluating disease activity and making earlier therapy decisions.

INTRODUCTION

Rheumatoid arthritis (RA) is an inflammatory systemic disease characterized by persisting synovitis, typically appearing symmetric in numerous joints, especially in the small finger and toe joints. The course of the disease is variable and fluctuating, but mostly progressive. Without adequate therapy, synovial inflammation can rapidly lead to severe destruction of cartilage and bone to the point of joint destruction. To prevent such a progression of disease, remission and low disease activity are the treatment goals, which can be achieved by early diagnosis and initiation of treatment using highly effective disease-modifying antirheumatic drugs (DMARDs) ([1-6]). For patients with an inadequate response to DMARDs, biologic agents such as tumor necrosis factor α (TNFα) inhibitors are the next step in therapy escalation. Clinical trials have shown that early highly effective treatment can limit the destructive joint process and even achieve remission ([7-9]). Rituximab (RTX), a monoclonal anti-CD20 antibody, resembles an established therapy option in case of nonresponse to TNFα inhibitors. Randomized controlled phase II and III studies have proven efficacy of RTX in patients with an inadequate response to methotrexate (MTX) and/or anti-TNFα therapy ([10-13]). After the first course of RTX, further improvement of signs and symptoms of RA has been described for patients with re-therapy after 6 months ([13-15]). In addition, seropositivity for rheumatoid factor (RF) and/or anti–cyclic citrullinated peptide (anti-CCP) has been shown to be of predictive value for improved outcome ([13, 16-19]). However, the optimal time point for re-therapy with RTX has not been sufficiently defined so far, and in a real-life situation, re-therapy is often administered only after worsening of symptoms or flare of disease. Therefore, more sensitive methods are required to identify patients with requirement for re-therapy earlier.

For evaluation of disease activity in patients with RA, clinical data (patient's history, clinical findings, and Disease Activity Score in 28 joints [DAS28]) as well as laboratory parameters (erythrocyte sedimentation rate [ESR] and C-reactive protein [CRP] level) are used. ESR and CRP level as markers of systemic inflammation are elevated in almost all RA patients and therefore are part of the American College of Rheumatology (ACR)/European League Against Rheumatism (EULAR) classification criteria for the disease ([20]). A correlation between ESR and CRP level and joint damage has been shown ([21-25]).

Musculoskeletal ultrasound (US) is able to detect both soft tissue lesions and bone lesions earlier than conventional radiography ([26-30]). Inflammatory processes cannot be judged by radiographs. US allows evaluation of inflammatory changes of joints, as well as their differentiation in exudative or proliferative processes, and the detection of tenosynovitis/paratenonitis. Therefore, US has an inherent part in the diagnostic and followup assessment of arthritic diseases, and is even more sensitive than clinical evaluation ([26-28, 31-34]). The use of power Doppler (PD) US is especially helpful in further differentiation of inflammatory disease activity ([35-40]). Early detection of the inflammatory activity, i.e., by US, is of major relevance because it influences further diagnostic and therapy decisions ([41]).

In this study, we evaluated inflammatory and destructive changes under RTX therapy by gray-scale (GS) US and PDUS over 12 months and compared US findings with clinical and laboratory data. We compared clinical and US findings of patients who received re-therapy over the evaluation period with those patients without re-therapy.

Box 1. Significance & Innovations

  • This is the first study to show usefulness of power Doppler ultrasound (US) in the prediction of rituximab re-therapy.
  • In comparison to clinical and laboratory parameters, gray-scale and power Doppler US are valuable tools to measure therapy response to rituximab in patients with rheumatoid arthritis.
  • All included patients negative for rheumatoid factor and anti–cyclic citrullinated peptide (n = 4) required re-therapy, whereas seropositive patients showed a better response to treatment with less need for re-therapy.
  • Sum scores of all included joints had the same informative value as the 7-joint US scores.

PATIENTS AND METHODS

Patients

All patients participating in this study gave their written consent prior to evaluation; the ethics committee gave its approval for this study (Tuebingen, Germany; 199/2007BO2). Twenty patients (14 women, mean ± SD age 56.0 ± 14.5 years) with longstanding RA (according to the ACR classification criteria [42]; mean ± SD disease duration 10.1 ± 7.3 years) and moderate to high disease activity (mean ± SD DAS28 at baseline 5.3 ± 1.0) were treated, according to protocol, with 2 infusions of 1,000 mg of RTX plus comedication at baseline and after 14 days. Fifteen of the patients were RF positive. Fifteen patients were positive for anti-CCP, of which 7 were highly positive (anti-CCP levels >100 units/ml). Four patients were seronegative for RF and anti-CCP.

Eighteen patients had received ≥1 anti-TNFα inhibitor previously; 2 patients had only been treated with DMARDs before receiving RTX (due to RA/systemic lupus erythematosus overlap and latent tuberculosis). Five patients received MTX comedication once a week during RTX therapy. Other concomitant therapies accompanying RTX treatment were leflunomide (3 patients), antimalarial drugs (2 patients), and azathioprine (1 patient). Five patients received combination therapy with 2 DMARDs (sulfasalazine and antimalarial drug, MTX and leflunomide, MTX and sulfasalazine, MTX and antimalarial drug, and sulfasalazine and cyclosporin A). Four patients received RTX as monotherapy. Eighteen of the patients were also treated with prednisolone (mean ± SD dosage 8.6 ± 5.0 mg/day) accompanying RTX therapy. During the observation period, none of the joints of interest was allowed to be treated with intraarticular corticosteroid injections. Six patients received nonsteroidal antiinflammatory drugs during the observation period.

Of all 20 patients, 14 received an additional course of RTX treatment during the observation period. The decision for re-therapy was made by the physicians at the Rheumatology Research Centre with regard to common recommendations in concordance with the “treatment on demand” approach by Smolen et al ([43]). The mean ± SD time before re-therapy was 7.0 ± 2.2 months after the start of therapy. Nine of these patients received the second course of treatment in the time interval of 6 and 9 months after the beginning of therapy, with the checkup 6 months after baseline being the last before the second course of treatment.

US technique

GSUS and PDUS were performed by EULAR guidelines ([44]) and according to the Outcome Measures in Rheumatology definition ([45]) on the clinically dominant hand and forefoot (wrist [radial, median, and ulnar from palmar/dorsal sides], metacarpophalangeal [MCP] joints 2–5 [palmar/dorsal sides; MCP joints 2 and 5 also from radial/ulnar sides], proximal interphalangeal [PIP] joints 2–5 [palmar/dorsal sides], and metatarsophalangeal [MTP] joints 2–5 [plantar/dorsal sides; MTP joint 5 additionally from fibular side]) using an Esaote MyLab70 US machine with a high-resolution 6–18-MHz linear array transducer. Within the region of interest, the color gain box was kept as small as possible, wall filter and persistency were kept low, pulse repetition frequency was set to 750 MHz, and color gain was adjusted as proposed by published recommendations ([46]). Subluxated, luxated, or mutilated joints inaccessible to the transducer were not taken into account. Examinations took place at baseline (before the first infusion of RTX), at week 2 (before the second RTX infusion), and at week 4, as well as after 3, 6, 9, and 12 months. The 2 examiners (SO, MB) were qualified and well experienced in US. The US examinations were performed in a center different from the clinical and laboratory examinations, meaning that the results of the US examinations did not have an influence on the physicians' decisions for re-therapy.

Parameters evaluated by US were signs of synovitis on GSUS and signs of vascularization on PDUS, graded by semiquantitative scales from 0–3. Signs of synovitis (synovial hypertrophy and/or effusion) seen on GSUS were developed from the methods of Backhaus et al and Scheel et al ([31, 33]), where grade 0 = no effusion or hypertrophy; grade 1 = minimal synovitis; grade 2 = moderate synovitis, joint capsule elevation parallel to the joint area; and grade 3 = extensive synovitis, strong distension of the joint capsule. Signs of vascularization in the joint area seen on PDUS were modeled according to Szkudlarek et al ([34]) based on intraarticular Doppler signals (underneath the joint capsule), where grade 0 = no color pixels/no flow, grade 1 = 2 single vessels and 1 confluent vessel or up to 3 single vessels, grade 2 = signals in <50% of the intraarticular joint area, and grade 3 = signals almost completely filling the intraarticular joint area (≥50%).

Different semiquantitative synovitis sum scores were then generated by adding the results of semiquantitative GSUS and PDUS grading, separately for the dorsal and palmar/plantar sides as well as combined, where score Ia = synovitis score of the hand (dorsal and palmar sides; wrist [median, radial, and ulnar] and MCP and PIP joints 2–5 [range 0–66]); score IIa = synovitis score of the foot (MTP joints 2–5 [dorsal and plantar sides; range 0–24]); and score IIIa = combined synovitis scores of the hand and foot (combined scores of synovitis of the hand and foot [range 0–90]). For evaluation of tenosynovitis, GSUS was used to detect the presence or absence of inflammatory signs (0/1) in the tendon sheath; PD activity was graded within the pathologic fluid, as explained above (range 0–3).

Erosions were evaluated by GSUS. The presence or absence of erosions (visible in 2 perpendicular planes) was noted on each joint, as well as the size (mm) measured in the longitudinal plane. Erosions were then graded according to Sommier et al (where grade 0 = no erosion, grade 1 = small erosion <2 mm, grade 2 = erosion size 2–3 mm, and grade 3 = erosion >3 mm or multiple [>1] erosions) to create erosion scores ([47]).

Different sum scores were then generated, separately for the dorsal and palmar/plantar sides as well as combined, where score Ib = erosion score of the hand (number of erosions and semiquantitative erosion score of the hand: wrist [median, radial, and ulnar], MCP and PIP joints 2–5 [dorsal and palmar sides], and MCP joints 2 and 5 [also from radial/ulnar sides]; range 0–24 and 0–72, respectively); score IIb = erosion score of the foot (number of erosions and semiquantitative erosion score of the foot: MTP joints 2–5 [dorsal and plantar sides] and MTP joint 5 [also from fibular side]; range 0–9 and 0–27, respectively); and score IIIb = combined erosion score of the hand and foot (number of erosions and semiquantitative erosion score: combined scores of erosions of the hand and foot; range 0–33 and 0–99, respectively).

Additionally, we calculated a novel 7-joint US (US7) score ([31]). The US7 score screens the most commonly involved joints in RA (wrist, MCP/PIP joints 2 and 3, and MTP joints 2/5) of the clinically more affected hand and forefoot to quickly assess disease activity. For the US7 GS synovitis score (range 0–27), we took into account the semiquantitative grades of the wrist (dorsomedian, ulnar, and palmar sides), MCP and PIP joints 2 and 3 (palmar side), as well as MTP joints 2 and 5 (dorsal side). For the US7 PD synovitis score (range 0–39), we included the wrist (dorsomedian, ulnar, and palmar sides), MCP and PIP joints 2 and 3 (dorsal and palmar sides), and MTP joints 2 and 5 (dorsal side). Tenosynovitis on GSUS (range 0–5) was scored by its presence (yes = 1, no = 0) on the wrist (dorsomedian, ulnar, and palmar sides) and on MCP joints 2 and 3 (palmar side; paratenonitis on MCP joints 2 and 3 from dorsal side as a rare finding was not taken into account). For PDUS (range 0–15), the semiquantitative grades of the same joint areas were used.

The US7 erosion score (range 0–14) is composed of the number of joints presenting an erosion. The score takes into account MCP/PIP joints 2 and 3 (dorsal and palmar sides), MCP joint 2 (radial side), MTP joints 2 and 5 (dorsal and plantar sides), and MTP joint 5 (fibular side). Additionally, the wrist joint (dorsomedian, ulnar, and palmar sides) was included (range 0–17).

Clinical parameters

At the time of each US assessment, the joints taken into account for DAS28 calculation were examined for tenderness and swelling. The patients' assessment of disease activity using a 100-mm visual analog scale (VAS) was documented. Furthermore, the DAS28 was calculated.

Laboratory data

At each checkup, blood was taken from the patient to determine the ESR (normal value <20 mm/hour) and CRP level (normal value <0.5 mg/dl or <5 mg/liter). Also, each patient's levels of IgM-RF (normal value <24 IU) and anti-CCP (anti–mutated and citrullinated vimentin antibodies, normal value <20 units/ml; Orgentec) were tested once before the start of therapy.

Statistical analysis

For statistical analysis and calculation, collected data underwent descriptive methods; correlation coefficients were calculated by the 2-sided Spearman's rho test. Significance of differences between 2-tailed groups was calculated by Wilcoxon's test of the median. The level of significance was defined as an alpha level less than 0.05. Statistical analysis was carried out using SPSS, version 19.0.

RESULTS

All patients (n = 20).

In comparison to baseline, all clinical and laboratory parameters decreased after 6 months (P < 0.05). After 12 months, the DAS28, number of swollen joints, ESR, and CRP level still showed a significant decrease (P < 0.05) compared to baseline (Figure 1). The patients' VAS as well as the number of tender joints had further decreased up to this point, although the change compared to baseline was not significant.

Figure 1.

Mean changes in clinical and laboratory parameters over 1-year followup of the total cohort (n = 20). DAS28 = Disease Activity Score in 28 joints; tender28 = 28 tender joint count; swollen28 = 28 swollen joint count; ESR = erythrocyte sedimentation rate; CRP = C-reactive protein. Color figure can be viewed in the online issue, which is available at http:onlinelibrary.wiley.com/doi/10.1002/acr.22103/abstract.

Various overall US parameters (combined scores of the hand and forefoot joints) showed a significant decrease. The GSUS synovitis sum score (IIIa) decreased after 6 and 12 months, as did the number and size (semiquantitative score) of erosions after 12 months (IIIb) (Table 1). Synovitis by PDUS and tenosynovitis in GS and PD mode showed slight numerical increases between 6 and 12 months after the start of treatment.

Table 1. Group 1 (total cohort): changes in US parameters over 1 year in all patients (n = 20)*
 Score Ia/b, hand jointsScore IIa/b, forefoot jointsScore IIIa/b, all joints and US7 score
BaselineAfter 12 monthsBaselineAfter 12 monthsBaselineAfter 6 monthsAfter 12 months
  1. US = ultrasound; GSUS = gray-scale US; PDUS = power Doppler US; US7 = 7-joint US.
  2. aSum scores of semiquantitative grading (range 0–3).
  3. bP < 0.05 (2-sided exact significance by Wilcoxon's test) for synovitis scores (a) and erosion scores (b).
  4. cSum scores of grading (where 0 = absent and 1 = present).
GSUS synovitis sum scorea       
Mean ± SD28.8 ± 13.322.6 ± 11.96.6 ± 3.63.8 ± 2.335.4 ± 14.128.4 ± 13.426.3 ± 12.2
Range10–586–510–140–810–647–628–53
P vs. baseline 0.009b 0.002b 0.028b0.001b
PDUS synovitis sum scorea       
Mean ± SD9.4 ± 8.28.4 ± 7.91.0 ± 1.90.5 ± 0.910.4 ± 8.67.8 ± 6.48.9 ± 8.5
Range0–261–290–70–30–280–201–31
P vs. baseline 0.268 0.352 0.2910.133
Erosionsc       
Mean ± SD12.6 ± 6.18.7 ± 4.44.1 ± 2.13.4 ± 1.516.7 ± 7.213.2 ± 6.812.2 ± 5.1
Range0–231–190–71–61–282–254–23
P vs. baseline 0.003b 0.192 0.0730.007b
Semiquantitative erosion sum scorea       
Mean ± SD17.5 ± 9.913.4 ± 10.57.4 ± 4.45.3 ± 2.824.8 ± 12.720.8 ± 12.419.1 ± 12.5
Range0–391–460–151–101–512–466–54
P vs. baseline 0.03b 0.013b 0.1560.011b
US7 GS synovitis sum scorea       
Mean ± SD    14.0 ± 4.510.6 ± 4.410.2 ± 4.4
Range    5–205–225–19
P vs. baseline     0.010b0.002b
US7 PD synovitis sum scorea       
Mean ± SD    5.9 ± 4.54.2 ± 3.74.8 ± 4.5
Range    0–140–110–16
P vs. baseline     0.0680.101
US7 GS tenosynovitis sum scorec       
Mean ± SD    1.0 ± 1.20.8 ± 1.01.0 ± 1.0
Range    0–30–40–3
P vs. baseline     0.5560.912
US7 PD tenosynovitis sum scorea       
Mean ± SD    0.8 ± 1.30.5 ± 1.20.9 ± 1.8
Range    0–40–50–6
P vs. baseline     0.3130.797
US7 erosionsc       
Mean ± SD    7.4 ± 3.86.0 ± 3.25.3 ± 2.5
Range    0–131–132–11
P vs. baseline     0.0640.040b
US7 erosions, including the wristc       
Mean ± SD    9.6 ± 4.48.2 ± 3.87.4 ± 2.9
Range    0–152–163–14
P vs. baseline     0.1620.039b
GSUS synovitis sum score, dorsala       
Mean ± SD13.6 ± 7.710.7 ± 6.05.4 ± 3.43.0 ± 1.8   
Range4–292–260–120–6   
P vs. baseline 0.042b 0.001b   
PDUS synovitis sum score, dorsala       
Mean ± SD6.3 ± 5.25.3 ± 4.30.9 ± 1.90.4 ± 0.7   
Range0–150–140–70–2   
P vs. baseline 0.267 0.375   
Erosions, dorsalc       
Mean ± SD5.8 ± 3.44.4 ± 2.51.6 ± 1.21.3 ± 1.1   
Range0–100–90–40–4   
P vs. baseline 0.071 0.477   
Semiquantitative erosion sum score, dorsala       
Mean ± SD7.6 ± 5.26.8 ± 5.32.5 ± 2.11.5 ± 1.3   
Range0–200–230–90–4   
P vs. baseline 0.310 0.105   
GSUS synovitis sum score, palmar/plantara       
Mean ± SD15.3 ± 6.311.9 ± 6.51.2 ± 1.40.8 ± 1.2   
Range6–294–270–50–4   
P vs. baseline 0.026b 0.760   
PDUS synovitis sum score, palmar/plantara       
Mean ± SD3.2 ± 3.83.1 ± 4.00.1 ± 0.50.1 ± 0.3   
Range0–130–150–20–1   
P vs. baseline 0.609 1.000   
Erosions, palmar/ plantarc       
Mean ± SD5.3 ± 2.83.1 ± 2.41.6 ± 1.51.2 ± 1.1   
Range0–110–80–40–3   
P vs. baseline 0.007b 0.187   
Semiquantitative erosion sum score, palmar/plantara       
Mean ± SD7.1 ± 4.34.6 ± 5.03.0 ± 3.02.0 ± 1.8   
Range0–160–180–90–5   
P vs. baseline 0.045b 0.045b  

As for the US7 score, synovitis on GSUS decreased significantly from mean ± SD 14.0 ± 4.5 at baseline to 10.6 ± 4.4 after 6 months (P = 0.01) and to 10.2 ± 4.4 after 12 months (P = 0.002 compared to baseline). Another US7 parameter that significantly decreased after 12 months was the erosion score, which diminished from mean ± SD 7.4 ± 3.8 to 5.3 ± 2.5 (P = 0.04, without wrist inclusion) and from 9.6 ± 4.4 to 7.4 ± 2.9 (P = 0.039) when including the wrist joints.

We found a positive correlation coefficient between reduction of ESR and CRP level (r = 0.77, P = 0.001). Positive correlation coefficients were also found between reduction of the DAS28 and GSUS synovitis score of the hand (Ia; r = 0.67, P = 0.004) and the synovitis score of the palmar side of the hand alone (r = 0.58, P = 0.018), as well as the US7 GS synovitis score (r = 0.57, P = 0.022) over 1 year. A decrease in the DAS28 also correlated with that of tenosynovitis on GSUS (r = 0.56, P = 0.023) and CRP level (r = 0.56, P = 0.028). A decrease in the number of tender joints also correlated with a reduction of synovitis on GSUS overall (IIIa; r = 0.72, P = 0.001), for the US7 score (r = 0.75, P = 0.001), and for the hand (Ia; r = 0.75, P = 0.001), as well as for the palmar (r = 0.69, P = 0.002) and dorsal (r = 0.54, P = 0.024) sides alone.

For the hand alone, after 1 year compared to baseline, significant decreases were found for the GSUS synovitis score of all joints (Ia; P = 0.009) as well as for the GSUS synovitis score evaluated from the dorsal (P = 0.042) and palmar (P = 0.026) sides alone. We also found a significant decrease in the number and size (semiquantitative score) of erosions of all examined hand joints (Ib) as well as for the palmar side alone (P < 0.05 for each). Regarding the dorsal side alone, the number and size of erosions decreased, although not significantly (P = 0.071 and P = 0.31, respectively).

For the forefoot alone, we detected a significant decrease of the GSUS synovitis score of all examined joints (IIa; P = 0.002) as well as of the joints on just the dorsal side (P = 0.001) after 1 year compared to baseline. Concerning erosions, the overall semiquantitative score (IIb) as well as on the plantar side decreased significantly (P < 0.05 for each) (Table 1). An example of reduction of synovitis by GSUS and PDUS is shown in Figures 2 and 3.

Figure 2.

Therapy response from baseline until 6 months later in the dorsal/radial wrist region. A and B, Reduction of gray-scale ultrasound (US) synovitis score from grade 3 to grade 1, respectively. C, Reduction of power Doppler US synovitis score from grade 2 to grade 0 (no US image available for grade 0). + = elevation of the joint capsule.

Figure 3.

Therapy response from baseline until 6 months later in the dorsal/midline wrist region. A and B, Reduction of gray-scale ultrasound (US) synovitis score from grade 3 to grade 2, respectively. C and D, Reduction of power Doppler US synovitis score from grade 2 to grade 1, respectively. + = elevation of the joint capsule.

Patients who received RTX re-therapy between 6 and 9 months after the beginning of therapy, with the checkup 6 months after baseline being the last before the second course of treatment (n = 9).

Compared to baseline, 6 months after the start of treatment, CRP level was the only parameter to have decreased significantly in this patient group (P = 0.016). The other clinical/laboratory parameters had decreased only numerically up to this point.

Three months after the second course of treatment (9-month assessment time point), the number of tender and swollen joints, DAS28, and VAS had decreased significantly compared to baseline and numerically compared to the status before re-therapy. ESR and CRP level also numerically decreased compared to baseline, although compared to 3 months earlier (before the second course of RTX), the laboratory parameters had actually increased.

After 12 months compared to baseline, the number of swollen joints was the only clinical parameter to show a significant decrease (P = 0.031). The other values decreased numerically compared to baseline, except for CRP level, which showed an increase. Comparing clinical status at the final examination time point with status before re-therapy, an increase in the DAS28 and laboratory parameters was seen. Conversely, tender and swollen joint counts as well as the VAS were reduced.

Regarding overall GSUS and PDUS synovitis as well as erosion scores, no significant decrease at the 6-, 9-, and 12-month assessment time points was observed compared to baseline. All values had decreased numerically up to the 6-month followup, except for PD activity in synovitis of all joints (IIIa; increase not significant). One year after the start of therapy, no significant changes were seen compared to assessment prior to re-therapy (Table 2). As for the US7 scores, no significant decreases after 6 and 12 months were observed.

Table 2. Group 2: changes in clinical/laboratory and US parameters over 1 year in patients who received re-therapy between 6 and 9 months after the beginning of therapy (n = 9)*
 BaselineAfter 3 months (before re-therapy)After 6 months (before re-therapy)After 9 months (after re-therapy)After 12 months
  1. US = ultrasound; DAS28 = Disease Activity Score in 28 joints; VAS = visual analog scale; ESR = erythrocyte sedimentation rate; CRP = C-reactive protein; GSUS = gray-scale US; PDUS = power Doppler US; US7 = 7-joint US.
  2. aP < 0.05 (2-sided exact significance by Wilcoxon's test).
  3. bSum scores of semiquantitative grading (range 0–3).
DAS28     
Mean ± SD5.4 ± 1.24.0 ± 0.84.2 ± 0.73.9 ± 1.54.4 ± 0.8
Range4.2–7.03.0–5.32.7–4.81.3–5.53.3–5.5
P vs. baseline 0.004a0.0780.047a0.063
Tender joints     
Mean ± SD10.7 ± 9.53.9 ± 4.76.5 ± 5.25.4 ± 5.45.5 ± 7.0
Range0–260–140–140–160–18
P vs. baseline 0.008a0.1880.031a1.000
Swollen joints     
Mean ± SD9.0 ± 7.44.1 ± 3.63.9 ± 4.22.9 ± 2.82.0 ± 2.5
Range2–221–121–140–80–6
P vs. baseline 0.008a0.0940.016a0.031a
VAS (0–100 mm)     
Mean ± SD55.8 ± 18.048 ± 14.254.8 ± 20.235.4 ± 24.840.3 ± 32.3
Range23–7621–6420–757–700–80
P vs. baseline 0.2110.1800.047a0.563
ESR, mm/hour     
Mean ± SD32.2 ± 26.722.2 ± 16.714.9 ± 7.124.9 ± 21.830.0 ± 18.3
Range7–804–522–262–5010–54
P vs. baseline 0.1600.0630.2970.313
CRP, mg/dl     
Mean ± SD1.4 ± 1.31.2 ± 1.10.6 ± 0.40.8 ± 0.81.6 ± 1.9
Range0.2–4.40.1–3.60.1–1.40.0–2.00.1–4.9
P vs. baseline 0.6410.016a0.0780.625
IIIa, GSUS synovitis sum scoreb     
Mean ± SD34.7 ± 16.935.9 ± 21.027.1 ± 17.030.7 ± 13.931.4 ± 16.5
Range16–6415–827–627–4912–53
P vs. baseline 0.6370.1480.1760.219
IIIa, PDUS synovitis sum scoreb     
Mean ± SD7.2 ± 7.810.6 ± 13.88.4 ± 7.88.0 ± 8.09.7 ± 11.2
Range1–230–450–200–231–31
P vs. baseline 0.5040.5310.7190.672
US7 GSUS synovitis sum scoreb     
Mean ± SD13.7 ± 5.113.1 ± 6.311.2 ± 5.712.2 ± 4.811.9 ± 5.8
Range5–205–255–224–185–19
P vs. baseline 0.9840.3130.4380.219
US7 PDUS synovitis sum scoreb     
Mean ± SD4.8 ± 4.65.2 ± 7.14.6 ± 4.64.7 ± 4.64.9 ± 6.1
Range0–120–230–110–120–16
P vs. baseline 0.9920.8751.0000.594

For the hand, the only significant decrease at the 6-month checkup compared to baseline was the GSUS synovitis score of just the dorsal side (P = 0.023). No significant changes were detected after 1 year compared to baseline.

For the forefoot alone, no significant decrease in US parameters was detected before re-therapy was given. After 1 year, the dorsal GSUS synovitis score was the only US parameter to show a significant decrease (P = 0.031).

All patients who received RTX re-therapy (n = 14).

The mean ± SD disease duration of patients who received re-therapy was 9.4 ± 6.6 years. Three patients received re-therapy earlier than 6 months after the start of treatment and 2 patients received re-therapy later than 9 months after the start of treatment. All patients negative for RF and anti-CCP were in the group of patients who were given a second course of treatment during the observation period.

After 6 months, compared to baseline, the number of swollen joints as well as laboratory parameters significantly decreased (P < 0.05). No significant changes in overall US scores were found. As for the US7 scores, no significant decrease after 6 months was observed.

For the hand alone, the overall number of erosions (Ib; P = 0.031) as well as the number of erosions on just the palmar side (P = 0.035) showed a decrease. Semiquantitative grading of synovitis by PDUS of the palmar side alone showed a significant increase from baseline up to 6 months later (P = 0.033).

For the forefoot, semiquantitative grading of synovitis on GSUS of just the dorsal side showed a significant decrease (P = 0.016). After 12 months, compared to baseline, the DAS28 and the number of swollen joints significantly decreased in this group of patients. Other clinical parameters also decreased, although not significantly.

As for the combined US parameters, a significant decrease was observed for the overall (IIIa) and US7 GSUS synovitis scores after 1 year. For the hand, significant changes were detected only for the number of erosions overall (Ib; P = 0.025) and for the palmar side alone (P = 0.027). For the forefoot, there was a significant decrease in overall synovitis on GSUS (IIa; P = 0.024), synovitis on GSUS of the dorsal side alone (P = 0.018), and the plantar erosion score (P = 0.043) at the end of the observation period of 12 months.

Patients who did not receive re-therapy (n = 6).

After 6 months, all clinical parameters had decreased in this group of patients. The DAS28, tender joints, VAS, and ESR had significantly decreased (P < 0.05). The mean DAS28 dropped from 5.5 to 3.1 (P = 0.031).

After 12 months, the clinical values were still lower compared to baseline, but no significant difference was detected at this time point. Between 6 and 12 months after therapy, all clinical parameters had increased again, with the mean DAS28 increasing from 3.1 to 4.1. CRP showed a fluctuation over time, with mean levels of 1.8 mg/dl at the start of therapy, 0.8 mg/dl after 6 months, 1.4 mg/dl after 9 months, and 1.3 mg/dl after 1 year.

After 6 months, the overall GSUS synovitis score (IIIa) had significantly decreased from mean ± SD 43.2 ± 8.4 to 28.3 ± 9.8 (P = 0.031). After 12 months, it further decreased to 24.2 ± 7.4 (P = 0.031 compared to baseline). PDUS showed a decrease from baseline (18.3 ± 6.0) to 6 months later (8.0 ± 5.2; P = 0.094), but increased again at the final examination (11.7 ± 6.8; P = 0.188 compared to baseline).

As for the US7 score, after 6 months, the GSUS synovitis score had significantly decreased from mean ± SD 15.8 ± 2.6 to 9.7 ± 1.6 (P = 0.031); the mean GSUS synovitis score was still 9.7 one year after the start of therapy. The PDUS sum score had also significantly decreased from baseline (mean ± SD 9.5 ± 2.9) until 6 months later (4.5 ± 2.1; P = 0.031), but increased again at the final examination (6.3 ± 3.7).

For the hand, after 6 months, a significant decrease of synovitis on GSUS was found for all joints (Ia; P = 0.031) and the palmar side alone (P = 0.031). After 12 months, a significant decrease was still seen from the palmar side alone (P = 0.031). For the forefoot, after 6 months, no significant change was observed. After 12 months, synovitis on GSUS overall (IIa; P = 0.031) and of just the dorsal side had significantly decreased.

The clinical, laboratory, and US score (IIIa) changes of the different groups, divided dependent on RTX re-therapy initiation, are shown in Figure 4.

Figure 4.

Changes of clinical/laboratory and ultrasound (US) parameters over 1-year followup. DAS28 = Disease Activity Score in 28 joints; CRP = C-reactive protein; GSUS = gray-scale US; PDUS = power Doppler US. Color figure can be viewed in the online issue, which is available at http:onlinelibrary.wiley.com/doi/10.1002/acr.22103/abstract.

DISCUSSION

The aim of the present study was to evaluate inflammatory as well as structural changes by US in RA patients receiving therapy with RTX over 1 year. Sonographic observations were compared with clinical and laboratory parameters to see if a therapy response could be detected, and if so, by which parameters this could best be done.

Regarding all patients receiving treatment with RTX, after the observation time of 1 year, all clinical parameters had decreased. A significant reduction of the DAS28 was observed, from 5.3 initially to 4.3, indicating a moderate therapy response to RTX corresponding to the EULAR criteria ([48]). A significant decrease was also detected for the number of swollen joints, as well as for the laboratory parameters ESR and CRP level. The decrease in disease activity could also be visualized by GSUS, where a significant reduction of the synovitis score of all joints as well as of the US7 synovitis score was seen over the observation period. A decrease in the US7 GSUS synovitis score correlated with that of the DAS28 and with the number of tender joints. The DAS28 is the main instrument to evaluate disease activity in RA; therefore, this indicates GSUS as a helpful additional tool to detect changes in disease activity.

When looking at the patient group that received the second course of treatment between 6 and 9 months after the start of therapy, the 6-month checkup marks the time point where the decision for another treatment cycle was made. At this point, CRP level was the only clinical parameter to have decreased significantly compared to baseline (although the others showed a decrease as well). Three months later, having received the second course of RTX, all clinical parameters had decreased, and changes were even significant compared to baseline. The laboratory parameters were lower compared to baseline, but compared to 3 months earlier (before re-therapy), they had actually increased. In the overall and US7 scores, no significant reduction in synovitis or erosions was observed before or after re-therapy. In contrast, a numerical increase of PDUS synovitis was seen 6 months after the start of treatment. Clinical and laboratory parameters are usually taken into account when assessing disease activity. They showed a decrease up to the 6-month evaluation, and therefore might have indicated a positive therapy response. Still, the treatment effect was considered insufficient and re-therapy was decided on (after physicians' decisions in accordance with common recommendations such as those by Smolen et al [[6, 43]] and/or if the DAS28 did not reach low disease activity or remission). However, US did not show a therapy response at this time point, and therefore might be more sensitive than clinical and laboratory parameters. This seems to be the case, especially for PDUS, which showed an increase in activity in overall synovitis. PDUS detects hypervascularization as a sign of active inflammation. The presence of PD activity implicates ongoing or newly inflamed disease activity that is not represented by clinical/laboratory parameters as yet, and therefore might be used to consider therapy escalation when one would not due to a rather inconspicuous clinical presentation.

When looking at all patients who received re-therapy, the value of PDUS becomes clear when noticing that semiquantitative grading of synovitis by PDUS of the palmar side of the hand showed a significant increase from baseline up to 6 months later (P = 0.033). All clinical parameters had decreased up to this point, laboratory parameters even significantly. No significant changes in overall US scores were found up to then. Since this was the group of patients receiving re-therapy, US, especially PDUS, seems to be very sensitive in detecting ongoing disease activity with need for re-therapy.

Regarding the group of patients who received RTX re-therapy, one aspect was also noticeable: all RF-negative and anti-CCP–negative patients were given re-therapy. Several studies have predicted a better response of seropositive patients to RTX ([13, 16-19]). In our patient group, this was confirmed by the fact that the response to the first circle of RTX infusion was considered insufficient in patients seronegative for both RF and anti-CCP. These patients presented ongoing disease activity and need for re-therapy at the 6-month evaluation time point, although showing at least a moderate response (DAS28 decreasing by a mean of 1.1 points) 6 months after the first treatment.

The group of patients who were not given a second treatment with RTX showed a better therapy response after 6 months: the DAS28 significantly decreased from 5.5 to 3.1, indicating a good therapy response and inactivity/low disease activity according to the EULAR criteria ([48]). Tender joint count, patients' VAS, and ESR also significantly decreased. We also observed positive therapy effects by US, with decreases in the overall GSUS (P < 0.05) and PDUS synovitis scores as well as significant decreases in the US7 (GSUS and PDUS) synovitis scores at the 6-month examination. In the overall synovitis scores, a significant decrease up to the 6-month examination could only be observed for GSUS. PDUS showed a decrease, although not significant. Again, in this patient group, US7 showed its sufficiency, since it was able to detect inflammatory changes like the overall synovitis score. Furthermore, it showed significant decreases in GSUS as well as in PDUS 6 months after the start of therapy, although regarding the overall sum scores, significance at this time point could only be detected for reduction of GSUS. Therefore, the US7 seems to indicate therapy response even better than the overall sum score. Interestingly, 12 months after the start of therapy, increases of clinical parameters (all but CRP) and PDUS (overall and US7) were observed. PDUS might indicate insufficient therapy at this time point, since it detects inflammation in the joints, marking the restart of disease activity. This also indicates PDUS as a possible predictor of need for re-therapy in RA patients receiving RTX therapy.

Interestingly, besides improvement of clinical inflammatory activity, less synovitis by US, and lowered acute-phase parameters, we also made the observation of a significant decrease in the number (all joints and US7 score) and size (as represented by the semiquantitative scores) of erosions after 1 year of treatment. This is possibly due to healing processes under therapy or may be a fault due to less synovitis working as an enhancer for the US waves to detect bone irregularities. Performing radiographs was not part of the study outline, and therefore was not involved prior to or after treatment with RTX at certain times that would have been favorable to compare US and radiographic data. In this study, we examined all wrist, MCP, PIP, and MTP joint regions of the clinically more affected hand and forefoot from the dorsal, palmar/plantar, and lateral sides (only MCP joints 2 and 5 and MTP joint 5). This examination proved to be rather time consuming (approximately 40–50 minutes, depending on the amount of pathologies to be documented). We evaluated parameters from all angles, since specific pathologies appear most often on certain sides. For example, the highest number of erosions is seen laterally on MCP joint 2 and MTP joint 5 ([31, 49]).

We found significant decreases in the synovitis score for the dorsal and palmar sides alone (all joints) as well as for both sites combined; from this, no conclusion can be made to only scan one or the other site (palmar/plantar versus dorsal). Most studies did not include sonographic analysis of the foot. However, van der Heijde has shown that in early arthritis, more joints are affected by structural damage in the feet than in the hands on radiographs ([50]). Another study examined the feet of RA patients for inflammatory and erosive changes and showed synovitis to be present most frequently in the second MTP joint, whereas erosions were most frequently imaged in the fifth MTP joint, which was implied by the US7 score ([31, 51]). In our study, US of the forefoot was able to reflect therapy response in concordance with the overall synovitis scoring results, showing a good response for all patients (decrease in synovitis and erosion size) and a less good response for patients in need of re-therapy (no significant decrease in US parameters after 6 months in the patient group receiving re-therapy after that time point). From our results, we recommend the use of the US7 scoring system instead of scanning all joints of the hand and forefoot, since it is less time consuming (only approximately 10–15 minutes) and seems to generate the same findings (significant decreases seen for overall synovitis/erosions as well as for the US7 score alone). Monitoring therapy response and detecting continuous arthritis has great significance, since insufficient treatment will lead to future structural damages ([7-9]). Since US has been shown to be even more sensitive than clinical examination in detecting joint inflammation, the inclusion of the US7 score when evaluating disease activity on followup visits is recommended ([27, 28, 31, 33, 34, 40, 52]).

In summary, the RA patients in our group showed a generally good response to therapy with RTX. Therapy effects could be visualized by clinical parameters and laboratory parameters, especially up to 6 months after the start of therapy. US seems to be more sensitive in representing therapy response than clinical and laboratory parameters. Specifically, PDUS was the best indicator for requirement of re-therapy, since it best demonstrated ongoing or rise in disease activity when other parameters did not.

However, to better assess the potential of US in evaluating the therapy effect of and the need for re-treatment with RTX in patients with RA, more studies with larger patient numbers, to be able to compare sufficiently sized groups of patients with different times of re-therapy, should be conducted. In addition, clearer definitions as to when re-treatment is administered (at flare and/or ongoing disease and/or nonresponse) and parameters on which determination of re-therapy is based on should be given. Also, radiographs should be performed in these future studies to compare radiographic and sonographic data to clarify whether RTX indeed has healing effects on erosions, as our findings indicate.

AUTHOR CONTRIBUTIONS

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. Backhaus 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. Reiche, Ohrndorf, Burmester, Backhaus.

Acquisition of data. Reiche, Ohrndorf, Feist, Messerschmidt, Backhaus.

Analysis and interpretation of data. Reiche, Ohrndorf, Burmester, Backhaus.

ROLE OF THE STUDY SPONSOR

Roche had no role in the study design, data collection, data analysis, or writing of the manuscript, as well as approval of the content of the submitted manuscript. Publication of this article was not contingent on the approval of Roche.

Ancillary