Abstract
Objectives
To determine if genicular artery embolization (GAE) is more effective in relieving pain symptoms after 12 months compared to a sham procedure in patients with mild to moderate knee osteoarthritis. A secondary aim was to investigate changes in synovitis observed at 4 months and the correlation with other clinical outcomes.
Materials and methods
A randomized controlled trial included patients with mild to moderate knee osteoarthritis unresponsive to conservative treatment, who were randomly assigned to receive either GAE or sham treatment. Follow-up was at 1, 4, 8, and 12 months. Pain was assessed using the Knee Injury and Osteoarthritis Outcome Score, and synovitis was imaged with contrast-enhanced MRI at baseline, 1, and 4 months. Outcomes were compared using generalized estimating equations.
Results
Fifty-eight patients were included (GAE group: 29, sham group: 29). Pain scores improved from 44.44 (95% CI: 38.87–50.02) at baseline to 65.61 (95% CI: 57.16–74.06) after 12 months in the GAE group and from 42.34 (95% CI: 36.45–48.22) to 58.15 (95% CI: 48.7–67.6) in the sham group. The between-group difference after 12 months was not significant (7.46; 95% CI: −13.63 to −28.56; p = 0.25). There were no significant within-group or between-group changes for synovitis.
Conclusion
Patients undergoing GAE and sham GAE both demonstrate an equal sustained pain reduction at 12 months follow-up, with no differences in synovitis reduction or other clinical outcomes. These results suggest a sustained placebo effect in the long term and do not support the clinical implementation of GAE as a treatment for KOA patients.
Key Points
Question Is GAE more effective than a sham procedure for mild-to-moderate knee osteoarthritis in the long term, and how does it affect synovitis?
Findings There was no significant difference in pain reduction or synovitis between the two groups.
Clinical relevance Pain reduction after GAE appears mainly placebo-driven. The placebo effect is sustained up to 12 months. Furthermore, synovitis was unaffected by the procedure. Therefore, our findings do not support clinical implementation.
Graphical Abstract
Keywords: Embolization, Genicular artery embolization, Knee osteoarthritis, Randomized controlled trial, Synovitis
Introduction
Knee osteoarthritis (KOA) is the most prevalent degenerative joint disorder, resulting in pain, reduced mobility, and diminished quality of life [1]. Conventional treatment options range from conservative approaches such as lifestyle changes, physical therapy, and pharmacological interventions to more invasive procedures, such as knee arthroplasty, in the end stage of the disease [2]. However, there are significant limitations of these treatments, including incomplete pain relief and risks associated with surgery, underscoring the need for alternative therapeutic strategies.
Genicular artery embolization (GAE) is a minimally invasive treatment for KOA that targets peri-genicular hypervascularity, which supplies blood to the synovium, a key contributor to KOA pain [3]. By reducing hypervascularity through endovascular embolization of specific genicular arteries, both synovial inflammation and, as a result, pain are thought to be alleviated [4]. Additionally, perivascular sensory neurons may regress, further reducing pain [3].
Early uncontrolled studies of GAE for KOA showed promising improvements in pain and synovitis, but subsequent sham-controlled randomized clinical trials (RCTs) mostly failed to demonstrate a sustained pain benefit over placebo [5–8]. Overall, the evidence remains inconclusive, with one small trial reporting only a short-term VAS pain advantage at 1 month before sham patients crossed over [9].
We hypothesized that, if there was a true and lasting effect of GAE, the pain reduction in the GAE group would be sustained on longer-term follow-up of our RCT, while the pain in the sham group would relapse. Therefore, the primary objective of this long-term study was to assess whether there was a difference between GAE in reducing pain compared to a sham procedure after 12 months. A secondary aim was to investigate the changes in synovitis observed on magnetic resonance imaging (MRI) and the correlation with other clinical outcomes, providing insight into the working mechanism of GAE.
Methods
The trial protocol, registered as NCT03884049 (ClinicalTrials.gov), was approved by the Erasmus University Medical Centre ethics committee (MEC identifier: MEC-2018-081). This prospective randomized sham-controlled trial was conducted at Erasmus MC, University Medical Centre Rotterdam, The Netherlands.
Participants
Patients were recruited at the orthopedic surgery outpatient clinics of Erasmus MC and four regional hospitals (IJsselland Hospital, Maasstad Hospital, Ikazia Hospital, and Spijkenisse Medical Centre). All participants provided written informed consent.
The most important inclusion criteria were: age ≥ 18 years, knee pain for ≥ 6 months, knee pain (NRS ≥ 4 to ≤ 8) on at least half of the days in the preceding month, insufficient response to conservative treatment for ≥ 6 months determined by an orthopedic surgeon and radiographic KOA (Kellgren and Lawrence grade 1–3 [10]).
The most important exclusion criteria were: contraindications for MRI or angiography and previous surgical treatment for KOA (except for knee arthroscopy). The full list of inclusion and exclusion criteria was published previously [11].
Randomization and blinding
Patients were randomly assigned to GAE treatment or sham treatment, in a 1:1 ratio with random blocks of 4 and 6, using a web-based randomization tool (www.aleaclinical.eu). We did not employ stratified randomization. An email with the allocation was sent to the interventional radiologist performing the procedure. He was the only study team member aware of treatment allocation for the entire study period.
The effectiveness of blinding was assessed using James’ blinding index [12]. Three hours post-intervention, participants indicated their perceived group assignment by selecting one of three options: ‘treatment group’, ‘placebo group’, or ‘I do not know’. Responses were used to compute the blinding index, ranging from 0 (no blinding) to 1 (perfect blinding). Blinding was considered successful if the lower bound of the 95% confidence interval (CI) exceeded 0.5 [12].
Procedures
Patients in the intervention group received GAE. A detailed description of the intervention is provided in the published protocol [11]. In short, after local anesthesia, a groin incision was made for vascular access via the common femoral artery. Selective angiography of the genicular arteries was performed using a microcatheter to identify areas of hyperemic blush. Upon visualization, these regions were embolized using 75 or 100 µm Embozene microspheres (Varian). Embolization continued until near-stasis of flow was achieved within the blush, while preserving the surrounding genicular vasculature (Fig. 1). Ice packs were applied to the knee to promote vasoconstriction of small cutaneous arteries and thereby minimize the risk of non-target embolization. All procedures were performed by a single interventional radiologist (A.M.) with 15 years of experience in interventional radiology and embolization. He was not involved in data collection. Sham-group patients underwent a sham GAE procedure. Local anesthesia was administered, and a similar groin incision was made as in the normal procedure. Steps taken during a normal procedure were mimicked, like gaining vascular access, maneuvering the C-arm and table, acquiring images, applying ice packs to the outside of the knee, and preparing and injecting the embolization material. Furthermore, communication between team members was consistent with that in a normal procedure. The sham procedure lasted between 1 h and 2 h, similar to the actual GAE procedure. After the procedure was finished, manual compression was given in the groin area, and a pressure bandage was applied, followed by 3 h rest in the supine position [11].
Fig. 1.
Patient inclusion flow chart
Clinical outcomes
Follow-up measurements were collected during in-person visits at Erasmus MC after 1, 4, 8, and 12 months by the same assessor (T.A.v.Z.), blinded to allocation. If patients were unable to attend a follow-up visit, questionnaires were mailed. KOA symptoms were assessed with the validated Dutch Knee Injury and Osteoarthritis Outcome Score (KOOS) questionnaire [13]. The KOOS consists of five subscales: symptoms, pain, daily living, sports and recreational activities, and quality of life. The subscales are summarized in a score ranging from 0 (no complaints) to 100 (worst complaints). Pain, stiffness, and the feeling of knee swelling were evaluated using a Visual Analogue Scale (VAS) ranging from 0 (no pain, stiffness, or swelling) to 100 (worst pain, stiffness, or swelling). To distinguish between intermittent and constant OA pain, the Intermittent and Constant Osteoarthritis Pain (ICOAP) questionnaire was used [14]. The presence of a neuropathic pain component was assessed using the painDETECT questionnaire, with a score of ≤ 12 indicating unlikely neuropathic pain [15]. The EQ-5D-5L questionnaire assessed quality of life [16].
After 12 months, patients were asked to complete the global perceived effect (GPE) questionnaire [17]. This questionnaire is a 7-point Likert scale asking: “To what extent are your complaints changed when compared with the situation just before you started treatment?” Answer options ranged from “worse than ever” to “completely recovered”.
Patients were asked not to start any new therapies during the trial, but doing so would not result in exclusion from the trial. Patients were allowed to stop any ongoing therapies at baseline. Concurrent therapies were documented during all follow-up visits.
MRI acquisition
MRIs were acquired at baseline, 1, and 4 months with a 3-T MR system (SIGNA Premier, General Electric HealthCare) using a dedicated 18-channel transmit-receiver knee coil (Quality Electrodynamics). Besides morphological pulse sequences to assess structural features of knee OA, a contrast-enhanced (CE) T1-weighted 3D water excitation spoiled gradient-echo sequence (weSPGR) was acquired 6 min after administration of 0.1 mmol/kg gadobutrol (Gadovist, Bayer AG). The scan parameters for this sequence were TR/TE = 10.7/5.4 ms, flip angle = 20°, FOV = 20 × 20 cm, slice thickness = 0.6 mm, and a 512 × 512 mm matrix. Additionally, 2D axial and sagittal fat-suppressed T2-weighted fast spin echo images with an in-plane resolution of 0.4 × 0.6 mm and 3 mm slice thickness were obtained.
MRI analysis
Synovitis was assessed on CE-MRI using the semiquantitative scoring system of Guermazi et al [18]. Synovial thickness was scored (grade 0: < 2 mm, grade 1: 2–4 mm, and grade 2: > 4 mm) at 11 sites. The sum across all 11 sites yielded the total synovitis score (0–4: normal, 5–8: mild, 9–12: moderate, and ≥ 13: severe). Additionally, effusion synovitis and Hoffa synovitis were scored on the non-contrast fat-suppressed T2-weighted images as described in the MRI Osteoarthritis Knee Score (MOAKS) [19]. Effusion synovitis, defined as hyperintensity within the articular cavity (Grade 0 = none, 1 = small, 2 = medium, and 3 = large), was scored on the axial T2-weighted images. Hoffa synovitis, defined as hyperintensity in Hoffa’s fat pad (0 = normal, 1 = mild, 2 = moderate, and 3 = severe), was scored on the sagittal T2-weighted images. Semiquantitative scoring was performed blinded to group allocation and timepoint. This aimed to minimize any observer bias [20]. All scoring was performed by a single reader (T.A.v.Z.) with 7 years’ experience in musculoskeletal imaging. This reader was trained by a musculoskeletal radiologist (E.H.G.O.) with 20 years’ experience on a separate set of cases. The radiologist was available for consultation on uncertain cases.
Statistical analysis
An intention-to-treat analysis was conducted. Baseline characteristics were analyzed using descriptive statistics. The primary outcome was the KOOS pain subscale (0-100) after 12 months. To evaluate differences in pain scores between groups over time, a generalized estimating equations (GEE) model was used, accounting for the correlation of repeated measures within individuals. Pain scores at different timepoints were modeled with group, time, and their interaction as predictors. The interaction effect between time point and group determined whether there was a significant difference between groups. Secondary outcomes included synovial enhancement, the other KOOS dimensions (symptoms, daily living, sport and recreation, and quality of life), VAS pain, ICOAP, painDETECT, and the EQ-5D-5L, and were analyzed with the same method. GPE was compared using an independent samples T-test. For comparison of proportions, a chi-squared test was used. The association between synovitis at baseline and a decrease in pain 12 months after treatment in the GAE group was tested using Spearman’s rank. A p < 0.05 was considered statistically significant.
Results
Details about patient inclusion and drop out are presented in Fig. 2. Between June 2019 and December 2021, 58 patients were included out of 166 screened. One patient dropped out between the 8-month and 12-month follow-up points. Two patients did not undergo MRI at the 4-month follow-up point, but had all other outcomes available. Baseline characteristics are presented in Table 1. James’s blinding index was 0.70 (95% CI: 0.61–1.00), indicating successful blinding.
Fig. 2.
KOOS pain score over time
Table 1.
Baseline characteristics
| GAE (n = 29) | Sham (n = 29) | ||
|---|---|---|---|
| Age | 59.01 (8.64) | 57.28 (7.95) | |
| Sex | f | 19 (65.52) | 16 (55.17) |
| m | 10 (34.5) | 13 (44.83) | |
| BMI (kg/m2) | 29.04 (4.62) | 30.99 (4.79) | |
| Laterality | Left | 14 (48.28) | 16 (55.17) |
| Right | 15 (51.72) | 13 (44.83) | |
| KL grade | 1 | 3 (10.34) | 1 (3.45) |
| 2 | 9 (31.03) | 12 (41.38) | |
| 3 | 17 (58.62) | 16 (55.17) | |
| Duration of complaints in years | 7.79 (7.37) | 8.17 (8.26) | |
| KOOS pain | 44.44 (15.59) | 42.34 (16.46) | |
| KOOS symptom | 50.25 (17.15) | 50.99 (20.95) | |
| KOOS activity | 53.1 (15.88) | 50.05 (17.98) | |
| KOOS sport | 17.68 (16.69) | 18.62 (16.79) | |
| KOOS QoL | 27.16 (17.38) | 27.16 (14.3) | |
| ICOAP constant | 44.14 (20.09) | 43.45 (20.92) | |
| ICOAP intermittent | 49.28 (14.94) | 47.84 (20.31) | |
| painDETECT | 9.83 (5.22) | 9.72 (6.11) | |
| EQ 5D 5 L | 0.65 (0.21) | 0.61 (0.28) | |
| VAS stiffness | 59.48 (24.92) | 59.48 (30.95) | |
| VAS swelling | 41.21 (29.48) | 34.9 (26.66) | |
| Procedure time (min) | 107.22 (24.43) | 94.39 (16.15) | |
| Contrast agent used (mL) | 156.52 (213.45) | - | |
| Radiation (mGy/cm2) | 6988.59 (7499.71) | - | |
| No. of arteries embolized | 3.79 (1.67) | - |
Ordinal variables are given as count (%)
BMI body mass index, KL grade Kellgren and Lawrence grade, KOOS knee injury and osteoarthritis outcome score, ICOAP intermittent and constant osteoarthritis pain, VAS visual analogue scale continuous variables are given as mean (SD)
Primary outcome
The primary outcome, KOOS pain at 12 months, did not differ significantly between the GAE and sham groups (between-group difference 7.46 points, 95% CI: –13.63 to 28.56; Table 2). KOOS pain also did not differ between groups at any earlier time point. KOOS pain nevertheless improved substantially in both groups: from 44.44 (95% CI: 38.87–50.02) to 65.61 (95% CI: 57.16–74.06) in the GAE group and from 42.34 (95% CI: 36.45–48.22) to 58.15 (95% CI: 48.70–67.60) in the sham group at 12 months (Fig. 3). At all earlier follow-up visits, both groups also showed significant within-group improvements in KOOS pain compared with baseline (p < 0.001 for both groups).
Table 2.
Comparison of the KOOS score and synovial enhancement between the two groups
| GAE | Sham | Between-group difference | p-value | ||
|---|---|---|---|---|---|
| KOOS pain | Baseline | 44.44 (38.87–50.02) | 42.34 (36.45–48.22) | 2.11 (−11.38 to 15.6) | 0.61 |
| 1 month | 67.53 (60.97–74.09) | 61.78 (55.07–68.49) | 5.75 (−9.86 to 21.36) | 0.23 | |
| 4 months | 65.8 (58.22–73.39) | 60.73 (52.93–68.53) | 5.08 (−13.03 to 23.18) | 0.36 | |
| 8 months | 61.97 (52.85–71.1) | 58.24 (49.7–66.78) | 3.74 (−17.06 to 24.53) | 0.56 | |
| 12 months | 65.61 (57.16–74.06) | 58.15 (48.7–67.6) | 7.46 (−13.63 to 28.56) | 0.25 | |
| Synovitis score | Baseline | 5.79 (4.45–7.12) | 5.45 (4.19–6.7) | 0.34 (−2.41 to 3.08) | 0.72 |
| 1 month | 5.75 (4.21–7.29) | 5.48 (4.39–6.58) | 0.27 (−2.57 to 3.1) | 0.78 | |
| 4 months | 5.57 (4.04–7.11) | 5.89 (4.56–7.22) | −0.32 (−3.36 to 2.72) | 0.76 |
Scores are given as mean (95% CI)
Fig. 3.
Pre- (left) and post (right) embolization angiography of the lateral inferior genicular artery. Notice the reduction of the hyperemic blush on the postembolization image while the lateral inferior genicular artery was preserved
Secondary outcomes
Synovitis scores remained essentially unchanged over time in both groups. For the GAE group, mean synovitis score was 5.79 (95% CI: 4.45–7.12) at baseline, 5.75 (95% CI: 4.21–7.29) at 1 month, and 5.57 (95% CI: 4.04–7.11) at 4 months, with no statistically significant within-group change (baseline to 1 month two-sided p = 0.94; baseline to 4 months two-sided p = 0.59; Fig. 4). In the sham group, synovitis was 5.45 (95% CI: 4.19–6.70), 5.48 (95% CI: 4.39–6.58), and 5.89 (95% CI: 4.56–7.22) at baseline, 1 month, and 4 months, respectively, again without a statistically significant within-group change (baseline to 1 month two-sided p = 0.93; baseline to 4 months two-sided p = 0.92). No significant between-group differences were observed at any time point (Table 2). There was no significant association between (degree of) synovitis at baseline and decrease of pain 12 months after treatment in the GAE group (ρ: −0.06; p = 0.69).
Fig. 4.
Contrast-enhanced T1 weSPGR images of the knee of a patient from the GAE group at the level of the suprapatellar recess. The baseline image (left) shows a marked synovial enhancement. There was no change 1 month after treatment (middle image) and 4 months after treatment (right image)
The mean GPE 12 months after treatment for the GAE group (mean 2.79; SD 1.60) and the sham group (mean 3.29; SD 1.61) was not significantly different (p = 0.25). There was no statistically beneficial effect of GAE over sham for any KOOS symptom subscale, on intermittent or constant pain, on a neuropathic pain component, or on quality of life. All secondary outcomes are described in Table 3. The usage of co-interventions throughout the study is presented in Table 4. At no point was there a difference in co-intervention usage between the two groups.
Table 3.
Comparison of questionnaire outcomes between broth groups for the KOOS subscales, a neuropathic pain component (painDETECT), Intermittent and constant pain (ICOAP), quality of life (EQ-5D-5L), and the VAS (Visual analogue scale) for pain
| GAE | Sham | Between-group difference | p-value | ||
|---|---|---|---|---|---|
| KOOS symptoms | Baseline | 50.25 (44.11–56.38) | 50.99 (43.49–58.48) | 0.74 (–15.37 to 16.85) | 0.88 |
| 1 month | 70.44 (64.86–76.03) | 67.12 (59.76–74.47) | −3.33 (−18.69 to 12.04) | 0.48 | |
| 4 months | 69.58 (62.77–76.39) | 67.98 (60.1–75.86) | −1.6 (−18.93 to 15.73) | 0.76 | |
| 8 months | 67.36 (60.33–74.4) | 65.52 (56.9–74.13) | −1.85 (−20.35 to 16.66) | 0.74 | |
| 12 months | 70.32 (63.93–76.71) | 67.07 (58.49–75.65) | −3.25 (−21.05 to 14.55) | 0.55 | |
| KOOS activity | Baseline | 52.26 (46.38–58.14) | 49.98 (43.33–56.62) | −2.28 (−17.04 to 12.48) | 0.61 |
| 1 month | 74.95 (68.19–81.71) | 69.22 (61.43–77.01) | −5.73 (−22.89 to 11.43) | 0.28 | |
| 4 months | 73.53 (65.38–81.67) | 65.58 (57.86–73.29) | −7.95 (−26.61 to 10.72) | 0.16 | |
| 8 months | 70.18 (60.64–79.72) | 64.15 (55.95–72.34) | −6.03 (−26.96 to 14.89) | 0.35 | |
| 12 months | 72.01 (63.36–80.65) | 65.53 (56.55–74.5) | −6.48 (−27.21 to 14.25) | 0.31 | |
| KOOS sport | Baseline | 16.63 (10.36–22.9) | 18.62 (12.62–24.63) | 1.99 (−12.45 to 16.43) | 0.65 |
| 1 month | 35.17 (27.08–43.25) | 31.57 (22.12–41.03) | −3.59 (−24.29–17.1) | 0.57 | |
| 4 months | 38.95 (30.23–47.67) | 28.87 (19.3–38.44) | −10.09 (−31.63 to 11.46) | 0.13 | |
| 8 months | 33.1 (24.52–41.68) | 27.41 (17.85–36.98) | −5.69 (−27.07 to 15.69) | 0.39 | |
| 12 months | 38.37 (26.83–49.9) | 33.68 (21.88–45.49) | −4.69 (−32.15 to 22.78) | 0.58 | |
| KOOS QOL | Baseline | 27.16 (20.94–33.37) | 27.16 (22.04–32.27) | 0 (− 13.39 to 13.39) | 1 |
| 1 month | 42.03 (35.91–48.14) | 39.22 (33.79–44.66) | −2.8 (−16.42 to 10.82) | 0.5 | |
| 4 months | 45.47 (37.97–52.98) | 44.83 (37.63–52.02) | −0.65 (−17.95 to 16.65) | 0.9 | |
| 8 months | 46.34 (37.23–55.45) | 40.09 (31.44–48.73) | −6.25 (−27.15 to 14.65) | 0.33 | |
| 12 months | 48.28 (39.42–57.13) | 40.99 (32.47–49.5) | −7.29 (−27.73 to 13.15) | 0.24 | |
| painDETECT | Baseline | 9.83 (7.96–11.7) | 9.72 (7.54–11.91) | −0.1 (−4.88 to 4.68) | 0.94 |
| 1 month | 8.66 (6.68–10.63) | 6.66 (4.83–8.48) | −2 (−6.47 to 2.47) | 0.14 | |
| 4 months | 7.16 (5.1–9.22) | 6.07 (4.56–7.57) | −1.09 (−5.34 to 3.16) | 0.4 | |
| 8 months | 8.72 (6.42–11.03) | 7 (5.27–8.73) | −1.72 (−6.51 to 3.06) | 0.24 | |
| 12 months | 7.1 (4.72–9.49) | 7.03 (5–9.07) | −0.07 (−5.29 to 5.15) | 0.97 | |
| ICOAP constant | Baseline | 44.14 (36.95–51.32) | 43.45 (35.97–50.93) | −0.69 (−17.95 to 16.57) | 0.9 |
| 1 month | 18.45 (11.88–25.02) | 25.17 (17.76–32.58) | 6.72 (−9.75 to 23.2) | 0.18 | |
| 4 months | 22.07 (13.46–30.68) | 25.86 (18.9–32.82) | 3.79 (−14.62 to 22.21) | 0.5 | |
| 8 months | 27.76 (18.03–37.49) | 28.62 (20.03–37.21) | 0.86 (−20.73 to 22.45) | 0.9 | |
| 12 months | 23.97 (15.53–32.4) | 28.06 (19.58–36.54) | 4.09 (−15.81 to 24) | 0.5 | |
| ICOAP intermittent | Baseline | 49.28 (43.94–54.63) | 47.84 (40.58–55.11) | −1.44 (−16.44 to 13.57) | 0.75 |
| 1 month | 27.44 (20.56–34.32) | 30.32 (22.59–38.04) | 2.87 (−14.33 to 20.08) | 0.59 | |
| 4 months | 29.6 (21.1–38.1) | 31.18 (23.49–38.87) | 1.58 (−17.48 to 20.65) | 0.79 | |
| 8 months | 30.46 (21.72–39.2) | 35.49 (26.88–44.1) | 5.03 (−15.39 to 25.45) | 0.42 | |
| 12 months | 29.6 (21.21–37.98) | 31.04 (22.52–39.56) | 1.44 (−18.45 to 21.33) | 0.81 | |
| EQ5D5L | Baseline | 0.65 (0.57–0.72) | 0.61 (0.51–0.71) | −0.04 (−0.24 to 0.17) | 0.56 |
| 1 month | 0.8 (0.74–0.86) | 0.75 (0.7–0.81) | −0.05 (−0.19 to 0.1) | 0.29 | |
| 4 months | 0.74 (0.64–0.84) | 0.73 (0.66–0.8) | −0.01 (−0.21 to 0.19) | 0.85 | |
| 8 months | 0.73 (0.64–0.82) | 0.69 (0.61–0.77) | −0.04 (−0.24 to 0.17) | 0.55 | |
| 12 months | 0.69 (0.57–0.81) | 0.65 (0.54–0.76) | −0.03 (−0.31 to 0.24) | 0.68 | |
| VAS pain | Baseline | 56.14 (49.9–62.37) | 52.69 (45.54–59.84) | −3.45 (−19.22 to 12.33) | 0.48 |
| 1 month | 33.62 (24.71–42.53) | 34.31 (25.21–43.41) | 0.69 (−20.5 to 21.88) | 0.92 | |
| 4 months | 33.24 (22.83–43.65) | 38.86 (30.15–47.58) | 5.62 (−16.96 to 28.21) | 0.42 | |
| 8 months | 37.69 (27.25–48.13) | 42.24 (32.2–52.29) | 4.55 (−19.55 to 28.65) | 0.54 | |
| 12 months | 34.66 (24.55–44.76) | 37.41 (26.82–48.01) | 2.76 (−21.61 to 27.12) | 0.71 |
Scores are given as mean (95% CI)
Table 4.
Cointervention usage
| Co intervention | Timepoint | GAE | Sham | p-value |
|---|---|---|---|---|
| Physiotherapy | Baseline | 5 (17.2%) | 4 (13.8%) | 0.72 |
| 1 month | 3 (10.3%) | 2 (6.9%) | 0.64 | |
| 4 months | 3 (10.3%) | 3 (10.3%) | 1.00 | |
| 8 months | 5 (17.2%) | 6 (20.7%) | 0.74 | |
| 12 months | 3 (10.3%) | 3 (10.3%) | 1.00 | |
| Brace usage | Baseline | 2 (6.9%) | 3 (10.3%) | 0.64 |
| 1 month | 3 (10.3%) | 1 (3.4%) | 0.30 | |
| 4 months | 2 (6.9%) | 3 (10.3%) | 0.64 | |
| 8 months | 1 (3.4%) | 1 (3.4%) | 1.00 | |
| 12 months | 1 (3.4%) | 0 (0%) | 0.31 | |
| Acetaminophen | Baseline | 11 (37.9%) | 14 (48.3%) | 0.43 |
| 1 month | 7 (24.1%) | 12 (41.4%) | 0.16 | |
| 4 months | 6 (20.7%) | 12 (41.4%) | 0.09 | |
| 8 months | 11 (37.9%) | 12 (41.4%) | 0.79 | |
| 12 months | 10 (34.5%) | 10 (34.5%) | 1.00 | |
| NSAID oral | Baseline | 8 (27.6%) | 11 (37.9%) | 0.40 |
| 1 month | 7 (24.1%) | 5 (17.2%) | 0.51 | |
| 4 months | 7 (24.1%) | 3 (10.3%) | 0.16 | |
| 8 months | 7 (24.1%) | 6 (20.7%) | 0.75 | |
| 12 months | 6 (20.7%) | 2 (6.9%) | 0.13 | |
| NSAID topical | Baseline | 2 (6.9%) | 1 (3.4%) | 0.55 |
| 1 month | 1 (3.4%) | 0 (0%) | 0.31 | |
| 4 months | 0 (0%) | 0 (0%) | - | |
| 8 months | 0 (0%) | 0 (0%) | - | |
| 12 months | 0 (0%) | 0 (0%) | - | |
| Opioids | Baseline | 0 (0%) | 2 (6.9%) | 0.15 |
| 1 month | 0 (0%) | 3 (10.3%) | 0.08 | |
| 4 months | 0 ((0%) | 3 (10.3%) | 0.08 | |
| 8 months | 0 ((0%) | 1 (3.4%) | 0.31 | |
| 12 months | 1 (3.4%) | 2 (6.9%) | 0.56 |
Proportions were compared using a chi-squared test
NSAID nonsteroidal anti-inflammatory drug
Adverse events
Adverse events up to 4 months were reported previously [5]. No additional adverse events were registered in the period between 4- and 12-month follow-up. Two patients in the GAE group reported paresthesia at the lateral ankle after the procedure. For one patient, this resolved spontaneously at the 8-month follow-up time point. For the other patient, this persisted throughout the entire study duration.
Discussion
In this RCT, we compared pain reduction in patients treated with GAE to sham GAE. After 12 months of follow-up, we found a statistically significant decrease in pain within both groups, but contrary to our hypothesis, no significant difference between the two groups. Furthermore, we found no difference in synovitis scores using CE-MRI between the two groups at 4 months after the treatment. Other secondary outcomes (VAS pain score, ICOAP, painDETECT, EQ-5D-5L, and the GPE) did not show a significant difference either. Our results suggest that the pain reduction observed in patients treated with GAE is mainly due to a placebo effect, and that this placebo effect is sustained after 12 months.
The results of this study, demonstrating that GAE offers no greater clinical benefit than a sham procedure, are of interest to all physicians treating KOA patients. These findings suggest that the therapeutic effects attributed to GAE may be primarily driven by placebo mechanisms rather than by the embolization itself. Remarkably, both groups showed clinically relevant improvement. Although this effect is clearly present in multiple studies without a control group, our study does not provide evidence for an underlying mechanism of action [7, 8, 21, 22].
The placebo effect is a well-documented phenomenon in the treatment of KOA, particularly when assessing pain relief [23]. Interestingly, the invasiveness of a procedure is positively correlated with the magnitude of the placebo effect [24]. More invasive treatments, such as intra-articular injections, tend to elicit stronger placebo responses than less invasive treatments. The relatively invasive nature of GAE could be an explanation for the large placebo effect.
Our findings are similar to a study by Landers et al, also comparing GAE to a sham procedure, in which the median KOOS pain score in the GAE group (n = 29) improved from 47.2 at baseline to 66.7 after 12 months, while the sham group (n = 30) improved from 47.2 to 61.1 [6]. This difference was not statistically significant, and they concluded that GAE produced no significant benefit over a placebo procedure.
In a recent meta-analysis, the weighted mean difference for VAS pain was −37 (95% CI: −44 to −30) after 1 month and −36 (95% CI: −51 to −22) after 12 months [25]. This is a greater effect than was achieved in the GAE group of our current study: −22.5 after 1 month and −21.5 after 12 months. The possibility exists that in our study, no difference was found between the treatment and control group because the effect achieved in the treatment group was smaller than expected.
We found no significant changes in synovitis scores in patients after GAE. This contrasts with a recent study by Dablan et al, which found a significant decrease in synovial enhancement (from 5.1 to 2.9) on CE-MRI using the same whole-joint semi-quantitative scoring method [26]. Pre-interventional synovial enhancement scores were similar between their study (5.1) and ours (5.8 in the GAE group and 5.5 in the sham group), but post GAE synovial enhancement was markedly different for their study (2.9) compared to ours (5.8). Furthermore, they reported a moderate correlation between pre-procedural synovial enhancement and reduction in pain scores. Patients with more severe pre-procedural synovial enhancement had a greater pain reduction. This correlation is of importance since it could hypothetically be used for patient selection. However, this correlation was not observed in the GAE group of our study. In our study, image analysis was performed blinded for patient information and timepoint (i.e., pre- or postprocedural) to minimize any observer bias. It is unclear whether the image analysis was done blinded to the time point in the study by Dablan et al, and this could be a potential source of bias that explains the discrepant results.
Our results challenge current assumptions about the mechanism of action underlying GAE and raise important questions about its clinical utility. Given the invasive nature, potential risks, and substantial cost of the procedure—estimated at approximately $3095.75 by a US based study and € 3432.37 by a Spanish based study—these results underscore the need for caution before broadly adopting GAE in clinical practice [27, 28] The high cost further amplifies the importance of confirming true therapeutic benefit, especially in resource-constrained healthcare systems. This study highlights the essential role of placebo-intervention controlled trials in validating the efficacy of interventional treatments and supports a more evidence-based, cost-conscious approach to managing KOA.
One potential limitation of the study is the small sample size. While the power calculation – based on the effect of GAE reported in an initial study by Okuno et al—resulted in the required sample size, the treatment group showed a smaller improvement and the control group a larger improvement than anticipated [7]. As a result, it is possible that a smaller effect of GAE may have been obscured due to insufficient statistical power, as the study was not designed to detect such a small effect. A second potential limitation is patient selection based on KL grade rather than specific markers of synovitis, like synovitis on MRI. One could argue that for a procedure to reduce perigenicular synovitis, patient selection should be based on the presence and severity of synovitis. However, since we found no significant correlation between synovitis at baseline and clinical outcome, we strongly believe this is unlikely to have influenced the observed outcomes. Larger, sham-controlled trials with similar methodology are required to establish with greater certainty that there is no difference between GAE and a sham procedure in terms of clinical benefit. In conclusion, GAE lacks long-term efficacy for mild-to-moderate KOA. At 12 months, the GAE and sham GAE groups showed equivalent sustained pain reduction and no difference in objective measures like synovitis or other clinical outcomes. This suggests a robust placebo effect and does not support the clinical implementation of GAE for KOA patients due to a lack of demonstrable benefit over the sham procedure.
Acknowledgements
We extend our sincere gratitude to all the individuals who volunteered their time and efforts to participate in this randomized controlled trial.
Abbreviations
- CI
Confidence interval
- GAE
Genicular artery embolization
- KOA
Knee osteoarthritis
- KOOS
Knee injury and osteoarthritis outcome score
- OA
Osteoarthritis
- RCT
Randomized clinical trial
- VAS
Visual analog scale
Funding
Funding was provided by Stichting Coolsingel (grant number: 562), Cook Medical (grant number: EUFRB-22-04-2016-IR-NL), Erasmus MC (grant number: 2017-17202), and Boston Scientific (no grant number). The funders had no role in study design, data collection, data analysis, data interpretation, or writing of the report.
Compliance with ethical standards
Guarantor
The scientific guarantor of this publication is Edwin Oei.
Conflict of interest
The authors of this manuscript declare no relationships with any companies, whose products or services may be related to the subject matter of the article.
Statistics and biometry
One of the authors has significant statistical expertise.
Informed consent
Written informed consent was obtained from all subjects (patients) in this study.
Ethical approval
Institutional Review Board approval was obtained.
Study subjects or cohorts overlap
Some study subjects or cohorts have been previously reported in: van Zadelhoff et al [5]; 10.1136/bmjopen-2024-087047.
Methodology
Prospective
Randomized controlled trial
Performed at one institution
Footnotes
Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Cross M, Smith E, Hoy D et al (2014) The global burden of hip and knee osteoarthritis: estimates from the Global Burden of Disease 2010 study. Ann Rheum Dis 73:1323–1330 [DOI] [PubMed] [Google Scholar]
- 2.Katz JN, Arant KR, Loeser RF (2021) Diagnosis and treatment of hip and knee osteoarthritis: a review. JAMA 325:568–578 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Mapp PI, Walsh DA (2012) Mechanisms and targets of angiogenesis and nerve growth in osteoarthritis. Nat Rev Rheumatol 8:390–398 [DOI] [PubMed] [Google Scholar]
- 4.Sterbis E, Casadaban L (2023) Genicular artery embolization technique. Tech Vasc Interv Radiol 26:100878 [DOI] [PubMed] [Google Scholar]
- 5.van Zadelhoff TA, Bos PK, Moelker A, Bierma-Zeinstra SMA, van der Heijden RA, Oei EHG (2024) Genicular artery embolisation versus sham embolisation for symptomatic osteoarthritis of the knee: a randomised controlled trial. BMJ Open 14:e087047 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Landers S, Hely R, Hely A et al (2023) Genicular artery embolization for early-stage knee osteoarthritis: results from a triple-blind single-centre randomized controlled trial. Bone Jt Open 4:158–167 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Okuno Y, Korchi AM, Shinjo T, Kato S, Kaneko T (2017) Midterm clinical outcomes and MR imaging changes after transcatheter arterial embolization as a treatment for mild to moderate radiographic knee osteoarthritis resistant to conservative treatment. J Vasc Inter Radio 28:995–1002 [DOI] [PubMed] [Google Scholar]
- 8.Padia SA, Genshaft S, Blumstein G et al (2021) Genicular artery embolization for the treatment of symptomatic knee osteoarthritis. JBJS Open Access 6:e21.00085 [DOI] [PMC free article] [PubMed]
- 9.Bagla S, Piechowiak R, Sajan A, Orlando J, Hartman T, Isaacson A (2022) Multicenter randomized sham controlled study of genicular artery embolization for knee pain secondary to osteoarthritis. J Vasc Interv Radiol 33:2–10.e12 [DOI] [PubMed] [Google Scholar]
- 10.Kellgren JH, Lawrence JS (1957) Radiological assessment of osteo-arthrosis. Ann Rheum Dis 16:494–502 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.van Zadelhoff TA, Moelker A, Bierma-Zeinstra SMA, Bos PK, Krestin GP, Oei EHG (2022) Genicular artery embolization as a novel treatment for mild to moderate knee osteoarthritis: protocol design of a randomized sham-controlled clinical trial. Trials 23:24 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.James KE, Bloch DA, Lee KK, Kraemer HC, Fuller RK (1996) An index for assessing blindness in a multi-centre clinical trial: disulfiram for alcohol cessation—a VA cooperative study. Stat Med 15:1421–1434 [DOI] [PubMed] [Google Scholar]
- 13.de Groot IB, Favejee MM, Reijman M, Verhaar JAN, Terwee CB (2008) The Dutch version of the knee injury and osteoarthritis outcome score: a validation study. Health Qual Life Outcomes 6:16 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Maillefert JF, Kloppenburg M, Fernandes L et al (2009) Multi-language translation and cross-cultural adaptation of the OARSI/OMERACT measure of intermittent and constant osteoarthritis pain (ICOAP). Osteoarthritis Cartilage 17:1293–1296 [DOI] [PubMed] [Google Scholar]
- 15.Timmerman H, Wolff AP, Schreyer T et al (2013) Cross-cultural adaptation to the Dutch language of the pain DETECT-questionnaire. Pain Pract 13:206–214 [DOI] [PubMed] [Google Scholar]
- 16.Herdman M, Gudex C, Lloyd A et al (2011) Development and preliminary testing of the new five-level version of EQ-5D (EQ-5D-5L). Qual Life Res 20:1727–1736 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Hudak PL, Wright JG (2000) The characteristics of patient satisfaction measures. Spine (Phila Pa 1976) 25:3167–3177 [DOI] [PubMed]
- 18.Guermazi A, Roemer FW, Hayashi D et al (2011) Assessment of synovitis with contrast-enhanced MRI using a whole-joint semiquantitative scoring system in people with, or at high risk of, knee osteoarthritis: the MOST study. Ann Rheum Dis 70:805–811 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Hunter DJ, Guermazi A, Lo GH et al (2011) Evolution of semi-quantitative whole joint assessment of knee OA: MOAKS (MRI osteoarthritis knee score). Osteoarthritis Cartilage 19:990–1002 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Hóbjartsson A, Thomsen ASS, Emanuelsson F et al (2013) Observer bias in randomized clinical trials with measurement scale outcomes: a systematic review of trials with both blinded and nonblinded assessors. Can Med Assoc J 185:E201 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Okuno Y, Korchi AM, Shinjo T, Kato S (2015) Transcatheter arterial embolization as a treatment for medial knee pain in patients with mild to moderate osteoarthritis. Cardiovasc Interv Radiol 38:336–343 [DOI] [PubMed] [Google Scholar]
- 22.Bagla S, Piechowiak R, Hartman T, Orlando J, Del Gaizo D, Isaacson A (2020) Genicular artery embolization for the treatment of knee pain secondary to osteoarthritis. J Vasc Interv Radiol 31:1096–1102 [DOI] [PubMed] [Google Scholar]
- 23.Zhang W (2019) The powerful placebo effect in osteoarthritis. Clin Exp Rheumatol 37:118–123 [PubMed] [Google Scholar]
- 24.Zhang W, Robertson J, Jones AC, Dieppe PA, Doherty M (2008) The placebo effect and its determinants in osteoarthritis: meta-analysis of randomised controlled trials. Ann Rheum Dis 67:1716–1723 [DOI] [PubMed] [Google Scholar]
- 25.Taslakian B, Miller LE, Mabud TS et al (2023) Genicular artery embolization for treatment of knee osteoarthritis pain: systematic review and meta-analysis. Osteoarthritis Cartilage Open 5:100342 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Dablan A, Erdim Ç, Güzelbey T et al (2024) Effectiveness of genicular artery embolization for reducing synovitis as assessed by contrast-enhanced MR imaging in knee osteoarthritis: a pilot study. J Vasc Interv Radiol 35:1313–1322.e1316 [DOI] [PubMed] [Google Scholar]
- 27.Davies E, Isaacson A (2018) 3:27 PM abstract no. 204 cost analysis of geniculate artery embolization versus conservative therapy for pain secondary to knee osteoarthritis. J Vasc Interv Radiol 29:S89 [Google Scholar]
- 28.Hernández-Yumar A, González-Hernández Y, del Pino-Sedeño T et al (2025) Genicular artery embolization for knee osteoarthritis: a systematic review with meta-analysis and cost-analysis. Gac Sanit 39:102459 [DOI] [PubMed]