Objectives: To evaluate contrast-enhanced ultrasound (CE-US) as a monitoring tool to assess hypervascularisation of synovial processes in knee osteoarthritis (OA) treated with intra-articular injections of the bradykinin-receptor 2 antagonist icatibant compared to contrast-enhanced magnetic resonance imaging (CE-MRI).
Patients and methods: In a randomised, double-blind, placebo-controlled trial, 41 patients with painful knee OA underwent US (12.5 MHz for B-mode and 3–8 MHz for CE-US), and 36 of the patients underwent additional MRI (0.2T) at baseline and after 3 injections of the study drug (after a mean of 22.2 days). A total of 15 patients received placebo (group A), 12 patients 500 μg icatibant (group B) and 14 patients 2000 μg icatibant (group C). Pain and the synovial process (B-mode, power Doppler US (PD-US), CE-US, CE-MRI) were assessed at both time points.
Results: At baseline, the placebo group showed more activity in terms of effusion in the superior and lateral recess in ultrasound as well as in PD-US in the lateral recess. Pain improved significantly in all subgroups. Effect sizes were 0.43 (pain at rest) and 0.52 (pain during activity) in group B vs 0.48 and 1.11 in group C. There was no change of US and MRI parameters. We found moderate to good correlation (r) and kappa values (κ) for effusion in the superior recess (r = 0.591, k = 0.453), effusion in the lateral recess (r = 0.304, k = 0.440) and contrast enhancement (r = 0.601, k = 0.242) between US and MRI.
Conclusions: Our results show that CE-US and CE-MRI have good agreement in assessing inflammatory changes in knee OA. For the 41 patients with OA, an analgesic effect of icatibant could clearly be shown, especially for pain during activity in the high dose icatibant group. However, we could not find an anti-inflammatory effect of icatibant by CE-US compared to CE-MRI.
Statistics from Altmetric.com
The synovial process in knee joint synovitis (KJS) of patients with rheumatoid arthritis (RA) can be visualised by ultrasound (US) and MRI, with good agreement between both modalities.1 In knee osteoarthritis (OA), the degenerative process may be accompanied by non-destructive synovial proliferation, joint effusion and popliteal cysts.2 3 Synovitis has been confirmed in knee OA by arthroscopy and histological studies,4–6 especially in patients with early disease.7
The mechanism of painful knee OA is not clearly established.8 9 Among the key mediators of pain in knee OA are bradykinins (BKs) that act directly by activation of BK2 receptors and indirectly by increasing the release of prostaglandins, cytokines and histamine.10–13 Elevated BK levels have been demonstrated in the synovial fluid of patients with OA.14 Furthermore, BK and leukotriene B4 concentrations were found to correlate with the severity of synovitis, and it was proposed to use BK in synovial fluid as an index of the degree of synovitis.15 Icatibant, a BK receptor 2 antagonist, has been reported to have an analgesic effect in patients with knee OA; however its role in inflammatory processes is not clear.16
In a comparative study with contrast-enhanced (CE)-MRI we have recently shown that the higher vascularisation of synovial processes in knee OA can be demonstrated by CE ultrasound (CE-US) using the second generation contrast medium SonoVue (Bracco, Milan, Italy) with higher sensitivity than B-mode or power Doppler ultrasound (PD-US) in comparison to CE-MRI.17
Only symptomatic treatment is currently available for knee OA including intra-articular steroid injection in RA knees.18 New treatment modalities such as disease-modifying drugs are urgently needed.
To the best of our knowledge, this is the first randomised, double-blind, placebo-controlled study comparing CE-US using SonoVue to MRI in assessing the anti-inflammatory and analgesic effects of icatibant.
The primary objective of this study was to evaluate CE-US as a tool to assess and monitor the degree of hypervascularisation of synovial processes in patients with knee OA after intra-articular treatment with two different doses of icatibant (500 μg and 2000 μg) using CE-MRI as a reference standard. Secondary objectives included the diagnostic sensitivity of B-mode US in detecting effusion and synovial hypertrophy as well as PD-US and CE-US in assessing vascularity of synovitis in knee OA as well as the efficacy and safety of icatibant.
PATIENTS AND METHODS
A total of 266 patients were enrolled in a randomised, double-blind, placebo-controlled, 3-arm, parallel-group, multicentre study. All patients had knee OA of Kellgren and Lawrence19 grade II or III with a confirmed diagnosis of at least 6 months prior to study start based on the 1986 American College of Rheumatology (ACR) diagnostic criteria.20 Further inclusion criteria were an indication for intra-articular injection therapy and age ⩾45 years. Mean knee joint pain at rest and during activity on the affected side had to be ⩾40 mm on a visual analogue scale (VAS) from 0 to 100 mm. The study was approved by the local ethics committee. Informed consent was obtained from all patients enrolled in the study.
Patients were treated by three intra-articular injections of either 500 μg or 2000 μg icatibant or matching placebo (5.0 ml 0.9% sodium chloride solution) into the affected knee joint. The injections were performed at weekly intervals within 2 weeks, and patients were followed-up (without any additional injection) for another 10 weeks. Patients were allowed to take acetaminophen (up to 500 mg four times a day) as a rescue medication.
Knee OA imaging substudy
Of the 266 patients, 43 were seen at our study site between February 2004 and May 2004. Only 2 patients dropped out after randomisation for medical reasons, giving a total of 41 patients finally included in the imaging substudy: 15 patients receiving placebo (group A), 12 the smaller dose of 500 μg icatibant (group B) and 14 the higher dose of 2000 μg icatibant (group C) (table 1). All 41 patients underwent US, and 36 of the 41 patients (mean age 64.3, range 45–77, SD 6.7; 21 female, 15 male) underwent additional MRI at baseline (1–5 days before the first intra-articular injection) and after the third injection (1–4 days after the last injection of the study drug). Five patients did not undergo MRI due to a high body mass index, anxiety or due to technical problems. US and MRI were performed on the same day prior to injection. The mean time between pre and post-therapeutic imaging was 22.2 days (range 10–34 days, SD 4.5 days).
Imaging and assessment of pain (for pain at rest and pain during activity) were carried out as described previously.17
Pain at rest and during activity was assessed on a VAS (0–100 mm) and graded as 0/absent VAS (0 mm), 1/mild (<40 mm), 2/medium (VAS = 40–60 mm), or 3/severe (VAS>60 mm).
B-mode and PD-US of the superior and lateral knee recesses were performed according to European League Against Rheumatism (EULAR) recommendations21 using a 12.5 MHz linear transducer (Technos MPX, Esaote Bioimaging, Genoa, Italy). CE-US of the superior recess was performed using a 3–8 MHz linear transducer. Effusion, synovial hypertrophy, PD-US and CE-US parameters were scored and graded from 0 to 3. For grading of effusion in the superior knee recess the cut-offs were <5 mm (grade 0), 5–<8 mm (grade 1), 8–<11 mm (grade 2) and ⩾11 mm (grade 3). For grading of synovial hypertrophy in the superior recess the cut-offs were 0 mm (grade 0), >0–<4 mm (grade 1), 4–<8 mm (grade 2) and ⩾8 mm (grade 3). In addition, the presence of synovial hypertrophy and degree of effusion in the lateral knee recess were measured but not graded. The lateral knee recess was examined using parallel scanning with a knee flexed (less than 90 degrees). Furthermore each knee was evaluated with PD-US using a Doppler frequency of 6.3 MHz, pulse repetition frequency of 500 Hz, gain of 103 and low wall filter. PD-US findings were graded on a semiquantitative scale from 0–3 (0: no PD-US signal; 1: mild PD-US signal, single colour marked vessel; 2: medium PD-US signal; 3: strong PD-US signal, large confluent marked vessels >50% of examined area). Contrast enhancement on US was graded in the same way using slope values (0: no enhancement; 1: mild enhancement; 2: medium enhancement; 3: strong enhancement). Slope values were obtained by the so-called time–intensity analysis, which is a dedicated US technique (CnTI, Esaote) operating at a low mechanical index (MI = 0.089–0.118; most examinations were performed with an MI of 0.098) after intravenous bolus injection of 4.8 ml of SonoVue The time–intensity analysis, which has already been used for different indications,22–24 provides a signal intensity averaged trend, measured on the B-mode (CnTI images for a period of 90 s (example shown below)). The abscissa gives the time in s (TIME-sec), while the ordinate shows averaged intensity values at each second/frame (INTENSITY). Slope values were calculated after measuring the intensity data then by the following formula:
Where int peak: intensity at peak/maximum, int min: intensity at minimum and time to peak: time at endpoint minus time at start of contrast medium enhancement measurement.
Slope values represent the change in contrast medium intensity over time (see formula for calculation of slope value) and might better represent hypervascularisation of synovial processes compared to maximum intensity, which represents the amount of blood volume. In addition, slope values are very stable parameters as they eliminate systematic errors by avoiding variance, which is in contrast to using maximum intensity values only that are susceptible to different interfering factors such as instrument settings (including gain), random measuring errors and degree of intensity. Furthermore, expressing the results of CE-US in terms of slope values allows direct comparison with CE-MRI, which also uses slope values.
Of the 41 patients, 36 underwent MRI at baseline and follow-up. Knees were imaged on a 0.2T permanent magnet C-scan system (Esaote), using a dedicated dual phased-array coil. The MR imaging protocol comprised a sagittal short tau inversion recovery (STIR) sequence, a coronal T1-weighted spin echo (SE) sequence, an axial T2-weighted turbo spin echo (TSE) sequence and a sagittal T2-weighted multi echo (spin echo and proton density) sequence, all using a slice thickness of 5.0 mm. This was followed by a three-dimensional T1-weighted gradient-recalled echo (GRE) sequence in sagittal orientation with 0.8 mm slice thickness followed by a dynamic series of 13 fast sagittal T1-weighted SE sequences containing 3–8 mm slices (9 s per sequence). After the first dynamic sequence gadolinium diethylenetriaminepentaacetic acid (Gd-DTPA; Magnevist, Schering AG, Berlin, Germany) was intravenously injected at a dose of 0.2 mmol/kg body weight.25 These fast dynamic sequences were followed by a second T1-weighted sagittal GRE sequence with 0.8 mm slice thickness. The post contrast GRE sequences were used for three-dimensional reconstruction in transverse and coronal orientations.
Joint effusion was defined as high signal intensity on T2-weighted sequences and low signal intensity on T1-weighted sequences that does not enhance after contrast media injection. The extent of an effusion was assessed on the axial and sagittal T2-weighted images and on the post-contrast T1 images by measuring the maximum anteroposterior diameter of the suprapatellar recess in the sagittal images and the maximum side-to-side diameter of the lateral recess in the axial images.
Effusion and contrast enhancement in the superior recess were scored and graded. Effusion in the superior recess was graded using the same cut-offs as for the US assessment. Contrast enhancement was measured as slope values and scored as the following: 0: no contrast enhancement; 1: mild enhancement; 2: medium enhancement; 3: strong enhancement.
Parameters not normally distributed were analysed with non-parametric statistical methods using the statistical package SPSS V. 13.0 (SPSS, Chicago, Illinois, USA) with a level of significance of 0.05 (Mann–Whitney U test). For parameters normally distributed the Student t test was performed. Agreement between sonographic and MRI findings and clinical data was analysed by the Spearman rank correlation test. κ Analysis was performed. Follow-up data were compared with the paired t test if values were normally distributed and with the Wilcoxon test if data were not normally distributed. For all treatment groups, effect sizes (ES) were calculated compared to the placebo group using the pooled baseline standard deviation according to the following formulas:29 30
Where n1: number of patients in treatment group, n2: number of patients in placebo group, SD1: standard deviation at baseline in treatment group and SD2: standard deviation at baseline in placebo group.
At baseline there were no significant differences between the three subgroups treated with placebo (group A), 500 μg icatibant (group B) or 2000 μg icatibant (group C) in terms of pain at rest and during activity, synovial hypertrophy in the lateral recess, CE-US (slope values), PD-US findings in the superior recess and effusion in the superior and lateral recess on MRI or CE-MRI (table 1).
However, sonographically the placebo group showed statistically significantly higher values for effusion in the superior and lateral recess compared to group B (p = 0.029 and p = 0.012, respectively) and synovial hypertrophy in the superior recess (p = 0.024 compared to group B and p = 0.021 compared to group C). Additionally, more patients in the placebo group had positive PD-US findings in the lateral recess compared to group C (p = 0.025).
Follow-up data after 3-weekly intra-articular injections of the study drug
Table 2 summarises the follow-up data for pain, US and MRI parameters. There was no statistically significant difference in intake of acetaminophen between the subgroups. A statistically significant improvement could be observed for synovial hypertrophy in the superior recess in the placebo group and for effusion in the lateral recess on MRI in the 2000 μg icatibant group. No statistically significant differences were found for any of the other US and MRI parameters including PD-US, CE-US and CE-MRI. This is also illustrated in fig 1A–D. With regard to safety, no adverse events occurred during the study in our patients.
Pain at rest and during activity as assessed by VAS improved significantly in all three subgroups.
The percentage of patients achieving 20% and 50% improvement for pain at rest and during activity is shown in table 3.
Comparison of the placebo group (group A) with the icatibant groups (group B, C, D) revealed no statistically significant differences for VAS20- or VAS50- improvement for pain at rest and during activity. However, more patients achieved 50% improvement for pain at rest (42.3% in verum group vs 33.3% in placebo group) and pain during activity (46.2% in verum group vs 20.0% in placebo group) in the verum group compared to placebo.
VAS data for pain at rest and pain during activity as expressed by the effect size (ES) showed the best improvement for the 2000 μg icatibant group (compared to placebo) with an ES of 1.11. All other effect sizes were below 0.60 (fig 2).
Correlation between pain, US and MRI parameters
At follow-up, pain assessed by VAS showed a moderate to good correlation with effusion in the superior and lateral recess, synovial hypertrophy in the lateral recess and with PD-US in the lateral recess (table 4). Correlation between US and MRI was 0.591 (p<0.001) in terms of effusion the superior recess and 0.601 (p<0.001) in terms of contrast enhancement. κ Analysis showed values of 0.453 (effusion in the superior recess) and 0.242 (contrast enhancement). Correlation between CE-US and PD-US was 0.520 (p<0.001).
In a previous study comparing CE-US and CE-MRI in our 41 patients with knee OA and healthy controls we have shown that CE-US is a valid method for assessing the degree of hypervascularisation in the setting of synovitis in knee OA.17
SonoVue is a second generation ultrasound contrast medium that has already been successfully used for different indications.31–40 While the first generation ultrasound contrast medium Levovist (Schering, Berlin, Germany) has been established for the detection of vascularity in joints of patients with RA in several studies,18 41–48 only few studies are available on the role of SonoVue in assessing KJS in RA and these are either uncontrolled49 or non-prospective.50 Some longitudinal prospective US studies have shown that contrast medium is suitable to monitor the therapeutic effect of TNFα inhibitors on knee or different hand joints in RA or psoriatic arthritis.51–55
Here, we now present follow-up data to the first prospective, randomised, double-blind, placebo-controlled, three-arm, parallel-group study to assess the potential of CE-US for monitoring the efficacy of an intra-articular injection of the BK-receptor antagonist icatibant in knee OA. The patients underwent follow-up CE-US after three intra-articular injections of either 500 μg or 2000 μg icatibant or matching placebo.
Icatibant is a synthetic decapeptide and a potent, stable, specific and long-acting antagonist of the BK2 receptor.56 It is stable in the synovial fluid and is not metabolised as quickly as BK in the systemic circulation.57 In a randomised placebo-controlled trial including 113 patients with painful knee OA, icatibant has shown a significant analgesic effect.16 Furthermore icatibant was found to be effective in a variety of in vitro56 58 and in vivo inflammation models.59–62 In our study however, we did not identify an anti-inflammatory effect of icatibant on US or MRI parameters such as effusion and synovial hypertrophy on US, PD-US, CE-US or CE-MRI during the observation period of about 3 weeks. Only MRI of the lateral recess demonstrated a statistically significant improvement in the icatibant 2000 μg group but this might be an accidental finding.
At follow-up, the level of pain at rest and during activity showed significant correlations with effusion on US, synovial hypertrophy on US and vascularity demonstrated by PD-US in the lateral knee recess (table 4). This observation indicates that inflammatory changes in the lateral knee recess should be included into the assessment of overall disease activity. Unfortunately, CE-US could not be performed for the lateral recess in this study due to a rather quick elimination of SonoVue: after intravenous injection of SonoVue contrast enhancement in the superior recess reaches a peak roughly within the first 30 s. As it can be seen in fig 3 in the time intensity curves slope values begin to decline within the following 1.5–2 min. Afterwards, SonoVue dissolves in the blood and is finally eliminated by expiration within the next few minutes (mean elimination time 7 min). So, the role of CE-US for the lateral knee recess has to be further assessed in future US studies.
CE-US correlated significantly with PD-US and CE-MRI in the superior knee recess, which shows that CE-US is a valid method for visualising hypervascularisation in the superior knee recess.
As far as the analgesic effect of icatibant is concerned, there was a statistically significant improvement of pain at rest and during activity in all subgroups including the placebo group. Though not statistically significant, markedly more patients achieved a 50% improvement of pain during activity (table 3). This finding is consistent with the very good effect size of 1.11 for pain during activity in patients who received the higher dose of icatibant (2000 μg). Such an effect size is far better than the effect size of topical non-steroidal anti-inflammatory drugs (NSAIDs)63 (pooled effect size of 0.28) and even seems to be comparable to the effect size of cyclo-oxygenase 2 (COX2)-selective inhibitors (effect size 0.76–0.95) found in a recent placebo-controlled study of patients with knee and hip OA.64 The analgesic effect of icatibant can partly be explained by the dose-dependent inhibition of nociceptive neurons (excitation and sensitisation).58 Given the cardiovascular and gastrointestinal side effect profile of NSAIDs and COX2-selective inhibitors, icatibant could be a useful alternative for treating painful knee OA. It might not be surprising that pain during activity improved more markedly than pain at rest as pain during activity might be more typical for patients with OA while pain at rest predominates in rheumatic knee joint synovitis.
The significant improvement of pain in the placebo group can partly be explained by a flushing and dilution effect of intra-articular injection of 5.0 ml 0.9% sodium chloride solution. Similar statistically significant placebo responses were also seen in placebo-controlled studies with hyaluronic acid in patients with knee OA.65 66
Though patients were randomly assigned to one of the three treatment arms there was more activity in terms of effusion in the superior and lateral recess on US and PD-US in the placebo group. This might have influenced the outcome parameters. Another limitation of our study might be that the study population was rather small, which is probably the reason why we did not find a statistically significant difference in the 50% improvement of pain during activity between the icatibant and placebo groups. We think that further studies are needed to assess the anti-inflammatory role of icatibant in OA.
In conclusion, our results show that CE-US and CE-MRI have good agreement in assessing inflammatory changes in knee OA. It needs to be further evaluated whether CE-US is a valid instrument to monitor disease activity and efficacy of anti-inflammatory drugs in knee joint synovitis in patients with knee OA and other rheumatic diseases.
Competing interests: TK is employed by Esoate, Hallbergmoos, Germany. There are no further conflicts of interest.
Funding: This study was sponsored by Sanofi-Aventis, Frankfurt am Main, Germany.
Ethics approval: The study was approved by the local ethics committee.
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.