Knee artery embolization works equally well with an antibiotic or permanent microspheres in a 12-month trial
In a 60-patient randomized trial, blocking knee blood vessels cut osteoarthritis pain by about half for a year, whether done with imipenem/cilastatin or microspheres.
Summary
Knee osteoarthritis pain often persists despite physical therapy and injections, long before someone is ready for knee replacement. Genicular artery embolization (GAE) is a minimally invasive procedure that blocks abnormal, inflammation-driven blood vessels in the knee lining. This randomized trial of 60 patients (82 knees) compared two blocking agents: temporary imipenem/cilastatin and permanent microspheres. Both worked. Average pain scores fell from about 8 to about 4 out of 10 at one year, and the two agents did not differ significantly on pain. Opioid and NSAID use dropped sharply. No major complications occurred, though temporary skin discoloration appeared in nearly half of treated knees. The findings suggest permanent microspheres are a reasonable option where the antibiotic cannot be used off-label.
Detailed Summary
Knee osteoarthritis is a leading cause of lost mobility with age, and many people have pain that persists despite physical therapy and injections but are not yet candidates for knee replacement. Genicular artery embolization (GAE) is a minimally invasive procedure that blocks abnormal, inflammation-driven blood vessels in the knee's synovial lining. It was first done with imipenem/cilastatin (I/C), whose crystals cause temporary occlusion. I/C is not approved as an embolic agent, and off-label use is not permitted in many countries. Permanent microspheres are increasingly used instead, but no prospective controlled trial had compared the two directly.
The GAUCHO trial randomized 60 patients (82 knees) at a single Brazilian center between February 2021 and December 2022. Participants had Kellgren–Lawrence grade 1–3 osteoarthritis and VAS pain of 6 or higher for over six months despite conservative care. The mean age was about 64, roughly 92% were women, and mean BMI was about 32. Patients received I/C (n=30) or 100–300 µm Embosphere microspheres (n=30) under local anesthesia, with patients blinded. Operators targeted MRI-guided synovial blush with small 0.3 mL aliquots. Outcomes were tracked at 30, 90 and 365 days.
Technical success was 100%. Procedures averaged about 22–25 minutes with roughly 10.5 minutes of fluoroscopy, with no group differences. Microspheres required less embolic volume (1.5 vs 2.0 mL, p=0.003). In the whole cohort, VAS fell from 8.13 to 2.74 at 30 days and 3.92 at one year. WOMAC pain dropped from 13.93 to 6.91, and KOOS pain rose from 30.64 to 59.56 (all p<0.001). Benefit was largest at 30 days and partly faded afterward.
Between-group differences in the main pain measures were not significant at any time point. At 365 days, VAS was 3.95 (I/C) vs 3.88 (microspheres; p=0.928), WOMAC pain 7.50 vs 6.27 (p=0.323), and KOOS pain 53.76 vs 65.81 (p=0.081), a numerical but non-significant edge for microspheres. Some secondary KOOS subscales favored microspheres early (symptoms p=0.009 and sports p=0.025 at 30 days; quality of life p=0.048 at 30 days), but these disappeared by 12 months. The authors note the microsphere group had more bilateral treatment (56.6% vs 30.0%, p=0.037).
No major adverse events occurred, and MRI showed no osteonecrosis, ischemia or non-target embolization. Temporary skin discoloration occurred in 39 of 82 knees (47.6%), at similar rates in both groups (45.2% vs 50%). One case was painful and resolved within 24 hours, and one prolonged case resolved after 180 days. Access hematomas occurred in 2 knees per group. Medication use fell significantly, with opioid users dropping from 11 to 0 (I/C) and 6 to 1 (microspheres), and NSAID users from 20 to 4 and 27 to 7. About 78% of evaluated patients met the KOOS success criterion at 12 months.
The results suggest the choice of embolic agent matters little for one-year pain relief, which is useful where I/C is unavailable. Caveats are important. There was no sham control, outcomes were subjective, the trial was small and single-center, six patients were lost to follow-up, and questionnaire completion was incomplete. Pain also drifted upward from the 30-day low, so longer follow-up is needed.
Key Findings
- Technical success was 100% across 82 treated knees, with an average of 2.74 genicular arteries embolized per knee.
- Overall VAS pain fell from 8.13 at baseline to 2.74 at 30 days and 3.92 at 365 days (all p<0.001).
- No significant I/C vs microsphere difference at 365 days: VAS 3.95 vs 3.88 (p=0.928); WOMAC pain 7.50 vs 6.27 (p=0.323); KOOS pain 53.76 vs 65.81 (p=0.081).
- Microspheres needed less embolic volume than I/C (1.5 vs 2.0 mL, p=0.003); procedure time was similar (22.2 vs 24.5 min, p=0.191).
- Temporary skin discoloration occurred in 47.6% of knees (45.2% I/C vs 50% microspheres, no significant difference); no osteonecrosis, ischemia or non-target embolization on MRI.
- Opioid users fell from 11 to 0 in the I/C group and from 6 to 1 in the microsphere group; NSAID users fell from 20 to 4 and 27 to 7 (p<0.001 within groups).
- Early KOOS subscale advantages for microspheres (symptoms p=0.009, sports p=0.025 at 30 days) were not sustained at 12 months.
Methodology
This was a prospective, randomized, single-center, single-blind trial in Brazil. It enrolled 60 patients (82 knees) with KL grade 1–3 knee osteoarthritis and VAS pain of 6 or higher for more than 6 months, randomized in blocks of six via sealed envelopes to I/C or 100–300 µm Embosphere microspheres. Outcomes (VAS, WOMAC, KOOS, medication use, MRI) were assessed at baseline, 30, 90 and 365 days. The trial was designed as non-inferiority with a 20-percentage-point margin and 26 patients needed per arm. Analysis used Generalized Estimating Equation models with Bonferroni correction, accounting for bilateral treatment, in SPSS 20.
Study Limitations
The trial was single-center and small, with no sham control and a single-blind design, so placebo effects on subjective pain scores cannot be excluded. Groups differed in bilateral treatment (56.6% vs 30.0%), six patients were lost to follow-up, and questionnaire data were incomplete at later time points. Follow-up was limited to 12 months, and skin discoloration was not defined by a standardized scale. No conflict-of-interest disclosures were included in the supplied text.
Enjoyed this summary?
Get the latest longevity research delivered to your inbox every week.
Enter your email to subscribe: