Intra-Articular Injection of Umbilical Cord-Derived Mesenchymal Stem Cells Is Safe and Effective for Moderate to Severe Knee Osteoarthritis With Synovitis: A Double-Blinded and Randomized Controlled Trial

Major Points and Findings:

This small pilot trial from Wuhan, China compared umbilical cord-derived mesenchymal stem cells (UC-MSCs) with both a true placebo and hyaluronic acid under double-blind conditions. In total 55 people were randomized to one injection and followed for six months.

The data are more modest than the title suggests. In the full study population, UC-MSCs did not improve pain, function or quality of life more than placebo or hyaluronic acid. The positive findings come from two subgroups identified after baseline MRI. The first was people whose knee also showed synovitis (inflammation of the joint lining), and the second was those within that group with a pain score of 4 or more out of 10. In those subgroups the UC-MSC patients improved and their synovitis eased on MRI. The finding needs confirmation, and the authors ask for a phase III trial.

Aim:

To evaluate the safety (primary endpoint) and efficacy (secondary endpoint) of a single intra-articular injection of allogeneic UC-MSCs for knee osteoarthritis, against placebo and hyaluronic acid (HA).

Methods:

This was a randomized, double-blind, three-arm trial at Wuhan Union Hospital, conducted from January 2021 to June 2022 (ChiCTR2000039017).

Participants were adults aged 40 to 75 with osteoarthritis in one knee, Kellgren-Lawrence (KL) grade 1 to 3, symptoms for more than 6 months and no relief from basic physical treatment. Exclusions included:

• Varus or valgus deformity of 10 degrees
• Grade III meniscus injury, or cruciate or collateral ligament injury
• Arthroscopy, or steroid or HA injection into the joint, in the past year
• Pain in the hip or ankle on the same side, infection, previous malignancy
• Long-term immunosuppressive or anticoagulant drugs
• BMI of 30 or above

Sixty-nine people were screened and 55 randomized. One withdrew before treatment, leaving 18 per group. Mean age was about 55, mean BMI about 24, and most were women. The disease was mild. Almost half had KL grade 1, only 5 of 54 had grade 3, and mean baseline pain was 3.5 to 4.0 out of 10.

UC-MSCs were grown from donated cord tissue in serum-free medium under GMP conditions and used at passage 5. They were tested for mycoplasma, endotoxin and microbial contamination before release.

Each patient received one injection under ultrasound guidance:

• UC-MSCs: 20 million cells in 2.5 mL of saline with 1% human serum albumin
• HA: 2.5 mL sodium hyaluronate (25 mg)
• Placebo: 2.5 mL of the same saline and albumin solution, without cells

The injecting surgeon took no part in assessment. A blinded physician saw patients at 2, 6, 12 and 24 weeks.

Outcomes: Safety was assessed by adverse events graded by CTCAE, laboratory tests and ECG. Efficacy measures were pain VAS (0 to 10), WOMAC, the Knee Society score (KSS) and EQ-5D-5L. MRI was scored with MOAKS for cartilage, synovitis and bone marrow lesions.

Results:

Safety: Fifteen placebo, 16 HA and 17 UC-MSC patients completed follow-up. There were no serious adverse events, joint infections, deaths or tumours, and nobody stopped because of side effects.

• Any adverse event: 0 with placebo, 3 (18.8%) with HA, 5 (29.4%) with UC-MSCs (P = 0.091). Injection-related: 0, 2 and 5 (P = 0.055).
• These were knee pain in 4 UC-MSC and 2 HA patients, settling within 2 days with rest or paracetamol, and one fever of 37.8°C about 6 hours after a UC-MSC injection, gone by the next morning. All were grade 1 except one grade 2 in the UC-MSC group.

The differences were not statistically significant, but no events at all occurred with placebo.

In the full study population there were no significant differences among the three groups at 6 months for VAS, WOMAC, KSS or EQ-5D. The UC-MSC group improved more on VAS and EQ-5D, but not significantly. HA did not help significantly either.

Synovitis subgroup: On baseline MRI, 11 placebo, 10 HA and 13 UC-MSC patients had synovitis. At 24 weeks:

• Synovitis had cleared in 8 of 13 (61.5%) UC-MSC patients, 3 of 10 (30%) HA patients and 2 of 11 (18.2%) placebo patients.
• The change from baseline was significant only in the UC-MSC group (P = 0.003, against P = 0.309 for HA and P = 0.325 for placebo). The direct three-way comparison at 24 weeks was not significant (P = 0.140).
• The UC-MSC group showed a significant WOMAC improvement from baseline at final follow-up (P = 0.036) and an EQ-5D improvement. Its larger fall in VAS was not significantly different from the other groups.

Among patients with synovitis plus pain of 4 or more, a still smaller subgroup, placebo and HA patients showed no significant change on any scale. UC-MSC patients had significantly lower pain at 6 and 24 weeks and better EQ-5D at 24 weeks.

Joint structure: No new cysts, tumours or abnormal bone formation were seen. Cartilage on MRI improved in 2 patients in every group and was unchanged in 67% to 77%, and bone marrow lesions behaved similarly across groups. There was no sign of cartilage regrowth at 6 months, which the authors say matches a meta-analysis of randomized trials that found insufficient short-term MRI evidence of cartilage regeneration after MSC injection.

The abstract reports “significantly better improvements” in VAS, WOMAC and EQ-5D in the synovitis subgroup. The full text presents these as changes from baseline within the UC-MSC group and states that the VAS difference between groups was not significant. The title says “moderate to severe knee osteoarthritis”, yet grade 4 knees were not eligible and most participants had grade 1 or 2 disease. “Moderate to severe” refers to pain. This summary follows the full text.

Conclusions:

The authors conclude that a single 20 million cell UC-MSC injection was safe and well tolerated over 6 months, and that it gave pain relief and functional improvement in knee osteoarthritis combined with synovitis, especially with moderate to severe pain. They state several limitations. This was a pilot trial with 55 participants and a real risk of missing true differences. Follow-up of 6 months is too short, and they recommend at least 12. There was no corticosteroid or physiotherapy comparator.

One further limitation concerns the subgroups. The synovitis and pain subgroups are not among the planned analyses in the methods section and contain 13 people per arm or fewer. Such findings need to be repeated before they can be relied on.

Background Information:

Synovitis is an independent risk factor for osteoarthritis progression and is closely tied to pain. The same group’s earlier animal work found that UC-MSCs slowed osteoarthritis by suppressing synovitis, and this trial’s MRI finding is consistent with that.

On dose and side effects, the authors compare their 29.4% rate of post-injection pain with 53% to 60% after 40 million bone marrow MSCs in one study, and with more pain and swelling at higher doses of adipose MSCs in another. They suggest larger cell numbers bring more cell death inside the joint, which can itself trigger inflammation.