A minimally invasive procedure that temporarily blocks small blood vessels around the knee cut osteoarthritis pain roughly in half for most patients and kept it down for at least a year, according to a German study published in Radiology, the journal of the Radiological Society of North America (RSNA). The study, which tested a newer type of dissolving embolic particle, reported no moderate or severe complications across 239 procedures.
The procedure, called genicular artery embolization (GAE), is performed by interventional radiologists. Working through a small catheter inserted at the groin, the physician finds the abnormal, overgrown blood vessels that develop in the lining of an arthritic knee and injects tiny particles to shut them down. Researchers believe those new vessels travel alongside pain-sensing nerves and fuel low-grade inflammation, so cutting off the blood supply may calm both. GAE does not repair cartilage or replace the joint; it targets pain.
What the study tested is the embolic agent. The most commonly used material has been imipenem-cilastatin, an antibiotic that forms crystals when mixed with contrast dye. The authors note it is used off-label, is not always available, raises antibiotic-resistance concerns, and comes in uneven particle sizes. Permanent plastic microspheres are the main alternative, but because they never dissolve, they may raise the risk of particles lodging where they shouldn’t and of prolonged inflammation. The CharitÉ team instead used size-calibrated gelatin microspheres (Nexsphere-F, made by NextBiomedical) designed to dissolve within hours.
The researchers enrolled 194 patients treated between July and November 2024. Their median age was 69, and 114 were women. All had knee pain that had persisted at least three months despite conservative care such as anti-inflammatory drugs, injections, and physical therapy, and each case was reviewed by a board of orthopedic surgeons and interventional radiologists. Most treated knees had moderate to severe arthritis on X-ray, and 80% of patients had had symptoms for more than two years. Forty-five patients had both knees treated.
Every procedure was technically successful. Mild, self-limiting side effects occurred after 16 of the 239 procedures (6.7%). Fifteen involved temporary skin discoloration over the treated area that cleared within 24 hours, and one patient developed a superficial bruise-like collection of blood at the groin. The authors note that skin discoloration has been reported in about 16% of cases in a prior meta-analysis of GAE and in up to 65% of cases with permanent particles.
On a 0-to-10 pain scale, patients’ median score fell from 7 before treatment to 4 at six weeks and 3 at six and 12 months, a median drop of 4 points. At 12 months, 80% of patients who completed follow-up had improved by at least 2 points, the threshold the study used for a clinically meaningful change. Scores on the Knee Injury and Osteoarthritis Outcome Score (KOOS), a 0-to-100 measure, also rose across every domain. Median daily-activity scores climbed from 53 to 71.5, pain scores from 44 to 65, symptom scores from 51 to 68, sports and recreation from 15 to 36, and quality of life from 19 to 40. Depending on the domain, 55% to 69% of patients reached the meaningful-improvement threshold on KOOS at one year. (The paper’s abstract states the range as 55% to 80%, apparently folding in the pain-scale result; the full text reports 55% to 69% for KOOS alone.)
In the RSNA release, Fleckenstein said many knee-arthritis patients face a real treatment gap once injections stop working and surgery is not an option, and he described GAE as a new option between injections and joint replacement. Those are the lead author’s characterizations; the study itself did not compare GAE with injections, surgery, or any other treatment.
The authors flag several limitations. The study had no control or sham group, so it cannot separate the procedure’s effect from placebo response, regression to the mean, or natural ups and downs in symptoms. It was conducted at a single center, relied on patient-reported outcomes rather than imaging or objective function tests, and followed patients for only one year. Follow-up fell to 79% at 12 months, and patients who missed visits were excluded from that time point’s analysis rather than counted, which the authors say could overstate the benefit. They also caution that the meaningful-improvement thresholds have not been validated specifically for GAE and should be read as descriptive.
That caution matters because the broader evidence is mixed. According to the ScienceDaily article, a sham-controlled trial published in European Radiology this year found no significant difference between GAE and a sham procedure, a European Journal of Radiology review pooling three sham-controlled trials found no significant advantage, and a June 2026 analysis of 45 studies rated the overall evidence as low or very low certainty. On the other side, the Society of Interventional Radiology issued a position statement in May 2026 supporting GAE for adults with symptomatic knee osteoarthritis who have not responded to conservative care and are not candidates for knee replacement or want to delay it, while calling for larger randomized trials.