Multiple intra-articular injections of umbilical cord mesenchymal stem cells in treating severe knee osteoarthritis: a triple-blind randomized controlled trial

Abstract

Background

Umbilical cord mesenchymal stem cells (UC-MSCs), with multidirectional differentiation potential, immunomodulatory properties, and paracrine repair functions, have emerged as a promising therapeutic approach for knee osteoarthritis (KOA). This study aimed to compare the 1-year clinical efficacy, structural repair, and safety of multiple intra-articular injections of UC-MSCs versus sodium hyaluronate (SH) in patients with advanced KOA (Kellgren-Lawrence grade III–IV).

Methods

In this single-center, prospective, triple-blind randomized controlled trial, 50 eligible patients were randomly allocated to the UC-MSCs group (n = 25) or the SH group (n = 25). The UC-MSCs group received 4 intra-articular injections at 2-week intervals to sustain continuous paracrine effects and tissue repair, while the SH group received 5 weekly injections following its standard clinical dosing regimen. Evaluations were performed at baseline and 1, 3, 6, and 12 months post-treatment, including the visual analog scale (VAS), Western Ontario and McMaster University Osteoarthritis Index (WOMAC), Lequesne score, subjective satisfaction score, and magnetic resonance imaging (MRI)-based radiological assessments (MOCART cartilage repair score, bone marrow edema remission rate, and meniscal injury repair rate).

Results

The UC-MSCs group showed a significantly greater and sustained reduction in VAS scores without rebound: the mean VAS score decreased from 7.2 ± 1.1 to 3.36 ± 0.8 at 12 months (a reduction of 3.84 ± 1.02 points), compared with a decrease from 7.0 ± 1.2 to 4.6 ± 0.9 in the SH group (a reduction of 2.42 ± 0.95 points) (P < 0.001). The 12-month WOMAC total score decreased by 23.2 ± 6.5 points in the UC-MSCs group (from 62.4 ± 8.7 to 39.2 ± 7.4), which was significantly larger than the 15.6 ± 5.8-point reduction in the SH group (from 60.8 ± 9.2 to 45.2 ± 8.1) (P < 0.05). For radiological outcomes, the mean MOCART score improved by 13.6 ± 4.2 points in the UC-MSCs group (from 18.2 ± 5.3 to 31.8 ± 6.1), versus 7.8 ± 3.5 points in the SH group (from 17.6 ± 5.1 to 25.4 ± 5.7) (P < 0.05). The remission rate of femoral bone marrow edema was 80.0% in the UC-MSCs group, significantly higher than 44.0% in the SH group (P = 0.007). No serious adverse events — defined per CTCAE v5.0 as joint infection, severe allergic reaction, tumorigenesis, or events requiring hospitalization — were observed in either group.

Conclusion

Compared with SH, multiple intra-articular injections of UC-MSCs provide more durable pain relief and greater functional improvement in severe KOA, coupled with significant advantages in cartilage repair, bone marrow edema resolution, and meniscal healing. This therapy represents an innovative and effective alternative to traditional conservative management for advanced knee osteoarthritis.

Background

Knee osteoarthritis (KOA) affects over 300 million adults worldwide [1, 2] and is pathologically characterized by irreversible articular cartilage degeneration, synovial inflammation, and subchondral bone remodeling [3]. These pathological changes commonly lead to persistent knee pain, joint stiffness, and restricted physical mobility [4]. For patients with mild-to-moderate KOA, first-line conservative management includes physical therapy, oral nonsteroidal anti-inflammatory drugs (NSAIDs), and intra-articular injections of corticosteroids, hyaluronic acid (HA), or platelet-rich plasma (PRP) [5]. Sodium hyaluronate (SH), as a natural component of synovial fluid, can lubricate the joint cavity and improve the intra-articular microenvironment, thereby temporarily relieving clinical symptoms [6]. Nevertheless, similar to most conventional conservative regimens, SH monotherapy fails to halt or reverse the progressive process of cartilage degeneration [7]. For patients with end-stage severe KOA graded Kellgren-Lawrence (K-L) III–IV, partial or total knee arthroplasty (TKA) is an effective and definitive treatment to restore joint function [8]; however, the procedure is inherently invasive and carries measurable perioperative risks (e.g., periprosthetic infection, thromboembolism), particularly in elderly patients or those with multiple comorbidities [9].

In recent years, emerging technologies and research have opened up many new directions for OA treatments. As a chronic low-grade inflammatory condition [10], reducing inflammation and oxidative stress, improving immune cell function, and modulating chondrocyte metabolism have become promising research directions for KOA treatment [11]. Mesenchymal stem cells (MSCs) are adult stem cells with multipotent differentiation potential and paracrine function, which can be isolated from various tissues such as bone marrow, umbilical cord, adipose tissue, and synovium [12]. With combined properties of multidirectional differentiation, immunomodulation, and paracrine repair, MSCs have emerged as a novel strategy in regenerative medicine, demonstrating promising clinical value for KOA treatment [13].

Human umbilical cord mesenchymal stem cells (UC-MSCs), as a relatively new source of therapeutic MSCs, have multiple advantages. They can be widely obtained from discarded umbilical cords, without additional trauma and ethical controversy [14]. Meanwhile, the strong proliferation ability makes UC-MSCs capable of rapid proliferation while maintaining stem cell characteristics [15]. Low expression of major histocompatibility complex class II molecules (MHC-II) and low immunogenicity can reduce the risk of transplant rejection [16]. Compared with bone marrow mesenchymal stem cells (BMSCs), UC-MSCs can secrete higher concentrations of anti-inflammatory factors while having better cartilage differentiation efficiency [17].

At present, high-quality clinical evidence regarding multiple intra-articular UC-MSCs injection for end-stage severe KOA remains insufficient. There is still a lack of prospective controlled trials head-to-head comparing UC-MSCs with routinely used conservative treatments. Therefore, this single-center prospective triple-blind randomized controlled trial was designed to systematically compare the 1-year clinical efficacy and imaging outcomes between intra-articular UC-MSCs and sodium hyaluronate injections in patients with K-L III–IV severe KOA. The findings aim to provide comprehensive evidence-based references for the standardized clinical application of UC-MSCs in the treatment of advanced knee osteoarthritis (Fig. 2).

Fig. 2.

Schematic diagram of the process of UC-MSCs therapy and the mechanism of UC-MSCs in treating KOA

Methods

Study design and ethics statement

This study protocol was registered on ClinicalTrials.gov (Registration Number: NCT05160831) and approved by the Ethics Committee of the Army Medical University (Approval Number: KY2021087, August 7, 2021). The study was strictly conducted in compliance with the Good Clinical Practice (GCP) guidelines and the Helsinki Declaration. This study adhered to relevant regulatory requirements. In China, cell-based therapies are regulated by a dual framework: the National Medical Products Administration (NMPA) governs product registration, GMP manufacturing and marketing trials, while the National Health Commission (NHC) oversees investigator-initiated research. As an ethically approved investigator-initiated trial, our quality control system was benchmarked against international standards, including the stringent quality, safety and GMP rules enforced by the European Medicines Agency (EMA) Committee for Advanced Therapies (CAT) for advanced therapy products. All participants provided written informed consent after thoroughly reviewing the study protocol.

Participants

Eligibility criteria

Inclusion criteria.

1. Aged 18 to 70 years old.

2. Diagnosed with primary knee osteoarthritis classified as Kellgren-Lawrence grade III or IV.

3. Able to perform daily living activities and walk independently.

4. No history of mental illness, with full ability to understand and cooperate with trial procedures and follow-up assessments.

5. Absence of local knee joint infection and systemic active infectious diseases.

6. No contraindications to knee joint puncture and intra-articular injection.

7. Persistent knee pain with a baseline VAS score ≥ 5 points despite standard conservative treatment (NSAIDs, physical therapy) for at least 3 months.

8. Willing and able to complete scheduled 1-, 3-, 6-, and 12-month follow-up evaluations.

Exclusion criteria.

1. Body mass index (BMI) > 30 kg/m².

2. Severe knee osteoarthritis with genu varum > 10° or genu valgum > 8° requiring surgical correction.

3. Pregnant or lactating females, or females with a positive pregnancy test within 7 days before intervention.

4. Severe dysfunction of the heart, cerebrovascular system, liver or kidney.

5. Confirmed immunodeficiency or autoimmune diseases.

6. Receiving long-term immunosuppressive agents or systemic corticosteroid therapy currently.

7. Participation in any other clinical trial within the past 6 months or during the study period.

8. History of previous knee surgery including arthroscopy, meniscectomy, osteotomy or joint arthroplasty.

9. Received intra-articular injection of glucocorticoid, PRP, stem cells or other biological preparations within the preceding 6 months.

10. Coagulopathy or bleeding tendency that cannot tolerate joint puncture.

Randomization and blinding

According to established criteria, a total of 50 patients with primary knee osteoarthritis were enrolled between January 2021 and June 2024 and randomly assigned to two groups: the control group (sodium hyaluronate group, n = 25) and the experimental group (UC-MSCs group, n = 25).

This study employed a computer-generated block randomization sequence (block size = 4) prepared by an independent statistician, to ensure 1:1 group allocation and balanced sample size throughout enrollment (Figs. 1).

Fig. 1.

Consort diagram. Subjects were randomized to a 1:1 ratio. No patients withdrew from the study after randomization till the end of follow-up duration

The blinding was maintained at three levels:

1. All patients were not informed of their assigned group or injection type;

2. the physician administering the injection was unaware of the treatment regimen;

3. and the outcome investigators were not aware of the group assignment.

Interventions

Preparation and characterization of UC-MSCs

The human umbilical cord-derived mesenchymal stem cells (UC-MSCs) employed in this study were isolated from fresh umbilical cord tissues procured from one single healthy parturient undergoing cesarean delivery to mitigate potential inter-donor variability. The donor provided written informed consent in advance and received comprehensive serological and pathogen screening to exclude hepatitis B, hepatitis C, human immunodeficiency virus, syphilis, and other infectious diseases (Fig. 2). No economic compensation was provided to the donors; instead, only the routine infectious disease screening costs were covered by the research expenses.

Fresh umbilical cord tissues were fully rinsed with sterile normal saline to remove residual blood contamination and cut into 1–2 cm segments. After stripping blood vessels, Wharton’s jelly was isolated, minced mechanically, and transferred into cell culture flasks containing dedicated stem cell growth medium. Primary UC-MSCs migrated from Wharton’s jelly after 7 days of primary culture and were harvested for subculture. Cells were cultured in DMEM/F12 medium supplemented with 10% fetal bovine serum (FBS, DOWOBIO, Shanghai, China) and incubated at 37 °C with 5% CO₂. The culture medium was refreshed every 3 days routinely.

Notably, while fetal bovine serum (FBS) has historically served as the standard supplement for ex vivo expansion of mesenchymal stem cells (MSCs), clinical-grade cell manufacturing for knee osteoarthritis (OA) and other indications is increasingly shifting toward xeno-free culture systems to align with Good Manufacturing Practice (GMP) standards and clinical translation requirements.

This paradigm shift is driven by well-recognized risks associated with xenogeneic components: potential immunogenic reactions in recipients, transmission of animal-derived pathogens, and inherent batch-to-batch variability, all of which can compromise product safety, consistency, and reproducibility. Furthermore, major global regulatory agencies increasingly recommend or mandate chemically defined, human-sourced media for advanced therapy medicinal products. The adoption of xeno-free conditions not only improves the safety profile of MSC-based therapies but also facilitates regulatory approval, process standardization, and scalable production.

In the present study, despite the use of FBS-containing medium, we implemented stringent preclinical quality control at passage 3—including comprehensive pathogen screening and strict endotoxin limitation—to minimize the potential hazards introduced by xenogeneic serum.

All UC-MSCs used in this trial were prepared from a single qualified donor as one uniform P3 batch to minimize donor-related variability. Only cells at the third passage (P3) were adopted for subsequent quality inspection and clinical injection.

Prior to clinical application, strict preclinical quality control was performed for P3 UC-MSCs, including sterility test for culturable bacteria and mycoplasma, detection of Treponema pallidum, hepatitis B virus, hepatitis C virus, human immunodeficiency virus, cytomegalovirus and Epstein-Barr virus. The endotoxin level was strictly controlled at ≤ 0.5 EU/mL to meet clinical injection standards.

Flow cytometry was applied to identify the phenotypic characteristics and purity of P3 UC-MSCs. After centrifugation and resuspension in staining buffer, cells were incubated with specific surface marker antibodies for flow cytometric detection. Phenotypic identification results showed that the positive expression rates of MSC-specific surface markers CD73, CD90, CD44 and CD105 were all > 95%, while hematopoietic lineage markers CD45, CD34 and CD14 were negatively expressed, consistent with the phenotypic criteria of human mesenchymal stem cells.

After all quality indicators were confirmed qualified, UC-MSCs were resuspended in sterile normal saline at a standardized concentration of 1.5 × 10⁷ cells/mL with a total volume of 3 mL. The prepared cell suspension was aseptically loaded into a 5 mL sterile syringe and kept at room temperature for immediate intra-articular injection.

Injection procedure

All injections were performed by the same orthopedic surgeon (with > 5 years of KOA injection experience) using a standardized lateral suprapatellar approach. To preserve blinding integrity despite different injection volumes, all study medications were prepared and dispensed into identical 5 mL unlabeled syringes by an independent unblinded research nurse. The administering physician and all outcome assessors had no access to preparation records, ensuring the robustness of the triple-blind design. All patients were prohibited from receiving other treatments within 12 months. Full weight-bearing was allowed immediately after injection, while high-impact vigorous exercise (e.g., running, deep squatting, heavy weight-bearing climbing) was restricted for 2 weeks after each injection.

UC-MSCs injection (Group A)

All patients received a total of four intra-articular injections once every 2 weeks. The injections were carried out by the same experienced orthopedic surgeon. Patients were placed in a supine position. After routine disinfection and draping, sterile syringe was inserted through the lateral suprapatellar approach. After draining the joint effusion, 3 mL of UC-MSCs suspension (1.5 × 10⁷ cells/mL, total 4.5 × 10⁷ cells per injection) was slowly injected. Ten times of knee flexion and extension movements were performed to ensure the even distribution. Oral administration of celecoxib (200 mg/d) was applied for the first three days after injection.

Sodium hyaluronate injection (Group B)

All patients received the intra-articular injection once a week, for a total of 5 injections. The injection process was the same as Group A. Medical grade sodium hyaluronate injections (2.5 mL; 25 mg, Shandong Bosilen Furuida Pharmaceutical Co., Ltd.) were used in all patients.

Outcome measures

All patients were evaluated their symptom and function recovery before injection and 1, 3, 6, and 12 months post-operatively.

All outcome measures were selected in accordance with the Outcome Measures in Rheumatology (OMERACT) consensus recommendations for KOA clinical trials, covering the core domains of pain, physical function, quality of life, and structural repair.

Clinical outcomes

Primary outcomes:

1. Visual Analog Scale (VAS) was used to quantify the pain degree on a scale of 0–10 (0 for painless and 10 for most unbearable pain). Pain scores were recorded at each time point to analyze the trend and duration of pain relief ; Simultaneously calculate the proportion of patients with VAS scores ≤ 3 points (indicating significant or complete pain relief) at 12 months postoperatively.

2. 2) The Western Ontario and McMaster University Osteoarthritis Index (WOMAC) was applied to evaluate the joint function improvement by scoring from three dimensions: pain (0–20 points), stiffness (0–8 points), and daily activity function (0–68 points), calculating the scores for each dimension and the total score (0–96 points).

Secondary outcomes:

1. The Lequesne score (0–24 points) was used to assess the impact of osteoarthritis related symptoms on patients’ quality of life, which could be comprehensively reflected by evaluating indicators such as joint pain, swelling, limited mobility, and walking ability. The higher the score, the more severe the condition.

2. 2) Subjective satisfaction score (0–10 points) was used to evaluate the patient’s subjective satisfaction for the treatment, with 0 points representing “completely dissatisfied” and 10 points representing “completely satisfied”.

Radiological outcomes

Baseline MRI was acquired within 7 days before the first injection, and follow-up MRI was performed at 6 and 12 months post-treatment (± 7-day window) with identical scanning parameters. All images were independently scored by two blinded board-certified musculoskeletal radiologists; discrepancies were adjudicated by a third senior radiologist.

Magnetic resonance imaging (MRI) with multiple sequences was used to perform multidimensional evaluation of intra-articular anatomical structures, including cartilage, bone marrow, meniscus, and synovium.

The Magnetic Resonance Observation of Cartilage Repair Tissue Score (MOCART) was used to assess the quality of cartilage repair from 8 dimensions (total score of 100 points), including cartilage defect filling rate, surface smoothness, integration with surrounding cartilage, and cartilage thickness. The condition of bone marrow edema was assessed based on the T2WI lipid suppression sequence. Femur, tibia and patella were assessed separately, analyzing the differences in the improvement of bone marrow edema in different regions by UC-MSCs. According to Stoller’s grading system (Grade I: punctate high signal intensity within the meniscus; Level II: Linear high signal not reaching the joint surface; Grade III: high signal reaching the joint surface, indicating tearing), the repair in meniscus injury was recorded grading before and after treatment.

Safety assessment

The adverse events (AE) were divided into perioperative AE and postoperative AE. Perioperative AE referred to adverse events that occurred from the first injection to the end of the fourth injection. Postoperative AE referred to all adverse events that occurred after the last injection to 12-month follow-up. Record the type of adverse event (such as pain, swelling, fever, infection, allergic reactions, etc.), occurrence time, duration, and treatment methods. AEs were classified per CTCAE v5.0 (Common Terminology Criteria for Adverse Events). Mild: Symptoms not interfering with daily activities; Moderate: Symptoms interfering with activities; Severe: Symptoms preventing activities or requiring hospitalization.

Statistical analysis

The sample size was calculated based on the primary endpoint of 12-month changes in VAS pain score. According to preliminary pilot data, the mean reduction of VAS score was assumed to be 2.5 points in the sodium hyaluronate group and 4.5 points in the UC-MSCs group. With a two-sided significance level of α = 0.05, statistical power of 80% (β = 0.2), and an estimated 10% loss to follow-up, a total of 25 patients per group were required.

All statistical analyses were performed using IBM SPSS Statistics 23.0. Continuous variables conforming to normal distribution were presented as mean ± standard deviation (SD); non-normally distributed continuous variables were expressed as median with interquartile range (IQR). Baseline demographic and clinical characteristics were compared between the two groups to assess baseline balance.

Independent-samples t-test was used for inter-group comparison of normally distributed continuous variables, while the Mann-Whitney U test was applied for non-normal data. Categorical variables and enumeration data (such as adverse event incidence, remission rate and meniscus repair rate) were compared using the Chi-square (χ²) test, and Fisher’s exact test was adopted when the theoretical frequency was insufficient.

Repeated-measures analysis of variance (RM-ANOVA) was used to compare dynamic changes of VAS, WOMAC, Lequesne, satisfaction and MOCART scores across multiple time points (baseline, 1, 3, 6, 12 months). Mauchly’s sphericity test was performed beforehand; the Greenhouse-Geisser correction was applied if the sphericity assumption was violated. Bonferroni post-hoc multiple comparison correction was used for pairwise comparisons between time points and between groups.

Intention-to-treat (ITT) analysis was conducted for all enrolled participants. Per-protocol (PP) analysis was additionally performed for patients who completed the whole intervention and follow-up without major protocol violation. Missing follow-up data were processed by multiple imputation with 5 imputation cycles to minimize statistical bias.

A two-tailed P-value < 0.05 was considered statistically significant.

Results

Baseline characteristics of participants

A total of 70 patients were screened, with 50 patients enrolled (25 Group A, 25 Group B). No patients were lost to follow-up (100% retention rate). Baseline characteristics were balanced between two groups, including age, gender, BMI, disease duration, and baseline clinical/radiological scores (all P > 0.05; Table 1).

Table 1.

Comparison of baseline general characteristics between two groups of patients (x ± s)

Characteristic	UC-MSCs Group (n = 25)	Sodium Hyaluronate Group (n = 25)	t / χ2	P-value
Demographic Features
Age, years	58.6 ± 8.2 (38–69)	56.9 ± 7.5 (36–68)	t = 0.736	0.465
Sex, n (%)			χ²=0.089	0.766
- Male	11 (44.0)	12 (48.0)
- Female	14 (56.0)	13 (52.0)
Body Mass Index (BMI), kg/m²	26.3 ± 2.1 (22.1–29.8)	25.8 ± 2.4 (21.5–29.6)	t = 0.821	0.416
Disease Duration, years	4.5 (2.3–6.8) [Median (IQR)]	4.2 (2.1–6.5) [Median (IQR)]	Z = 0.364	0.716
Kellgren-Lawrence (KL) Grade, n (%)			χ²=0.168	0.682
- Grade III	16 (64.0)	15 (60.0)
- Grade IV	9 (36.0)	10 (40.0)
Affected Knee, n (%)			χ²=0.254	0.614
- Left	13 (52.0)	11 (44.0)
- Right	12 (48.0)	14 (56.0)
Baseline Scores
VAS Score (0–10)	7.2 ± 1.1 (5–9)	7.0 ± 1.2 (5–9)	t = 0.628	0.533
WOMAC Total Score (0–96)	62.4 ± 8.7 (45–78)	60.8 ± 9.2 (43–76)	t = 0.657	0.514
- WOMAC Pain Subscale (0–20)	14.3 ± 2.5 (10–19)	13.8 ± 2.7 (9–18)	t = 0.683	0.498
- WOMAC Stiffness Subscale (0–8)	5.2 ± 1.3 (3–8)	5.0 ± 1.2 (3–7)	t = 0.552	0.583
- WOMAC Function Subscale (0–68)	42.9 ± 6.8 (32–58)	42.0 ± 7.1 (30–57)	t = 0.476	0.636
Lequesne score (0–24)	12.6 ± 2.4 (8–17)	12.2 ± 2.5 (8–16)	t = 0.569	0.571
Baseline Radiological Features
MOCART Score (0–100)	18.2 ± 5.3 (10–28)	17.6 ± 5.1 (9–27)	t = 0.445	0.658
Meniscal Tear (Stoller Grade ≥ II), n (%)	19 (76.0)	18 (72.0)	χ²=0.093	0.760

Clinical outcomes

Primary outcomes

Both groups showed significant improvements in VAS and WOMAC scores over time (Fig. 3).

Fig. 3.

Clinical outcome scores over 12 months in knee OA patients treated with UC-MSCs vs. sodium hyaluronate (n = 25 per group). A VAS pain score; B Lequesne index; C WOMAC total score; D WOMAC pain subscale; E WOMAC stiffness subscale; F WOMAC function subscale. Data are presented as mean ± standard error. (* Indicates that the difference between the two groups is statistically significant)

VAS score

VAS scores decreased significantly over time in both groups (P < 0.05), with a greater magnitude of reduction observed in the UC-MSCs group. Meanwhile, scores in the SH group exhibited a rebound increase after 6 months (from 3.8 ± 0.8 to 4.6 ± 0.9). Intra-group differences were statistically significant from the 1-month time point onward (P < 0.05), as shown in Table 2; Fig. 3A. At 12 months, 92.0% (23/25) of patients in the UC-MSCs group had a VAS score ≤ 3 points, compared with 64.0% (16/25) in the SH group (χ²=5.76, P < 0.05). The mean VAS score of the UC-MSCs group was significantly lower than that of the SH group at 12 months (P < 0.001), and the significant analgesic response rate (92.0% vs. 64.0%) was also higher (P < 0.05), confirming the superior long-term analgesic benefit of UC-MSCs.

Table 2.

Clinical outcomes before and after injections (x ± s)

Outcome Measure	Group	Baseline	1 Month	3 Months	6 Months	12 Months	RM-ANOVA (Group×Time Interaction)	P-value (Group×Time)
VAS Pain Score (0–10)	UC-MSCs (n = 25)	7.2 ± 1.1	4.8 ± 1.0	3.1 ± 0.8	2.1 ± 0.7	1.9 ± 0.6	F = 42.86	< 0.001*
	Sodium Hyaluronate (n = 25)	7.0 ± 1.2	5.5 ± 1.1	4.6 ± 0.9	3.8 ± 0.8	4.6 ± 0.9
WOMAC Total Score (0–96)	UC-MSCs (n = 25)	62.4 ± 8.7	45.2 ± 7.3	32.6 ± 6.1	24.8 ± 5.2	18.0 ± 4.5	F = 38.52	< 0.001*
	Sodium Hyaluronate (n = 25)	60.8 ± 9.2	51.3 ± 7.8	44.7 ± 6.9	36.5 ± 6.3	48.6 ± 7.1
WOMAC Pain Subscale (0–20)	UC-MSCs (n = 25)	14.3 ± 2.5	10.2 ± 2.1	6.8 ± 1.7	4.5 ± 1.3	3.1 ± 1.0	F = 35.19	< 0.001*
	Sodium Hyaluronate (n = 25)	13.8 ± 2.7	11.6 ± 2.3	9.5 ± 1.9	7.2 ± 1.5	9.2 ± 1.8
WOMAC Stiffness Subscale (0–8)	UC-MSCs (n = 25)	5.2 ± 1.3	3.8 ± 1.0	2.9 ± 0.8	2.2 ± 0.7	1.8 ± 0.6	F = 2.84	0.096*
	Sodium Hyaluronate (n = 25)	5.0 ± 1.2	3.9 ± 1.1	3.1 ± 0.9	2.5 ± 0.8	2.9 ± 0.9
WOMAC Function Subscale (0–68)	UC-MSCs (n = 25)	42.9 ± 6.8	31.2 ± 5.7	22.9 ± 4.8	15.8 ± 3.9	13.1 ± 3.2	F = 40.23	< 0.001*
	Sodium Hyaluronate (n = 25)	42.0 ± 7.1	35.8 ± 6.2	32.1 ± 5.5	27.0 ± 4.9	36.5 ± 5.8
Lequesne Score (0–24)	UC-MSCs (n = 25)	12.6 ± 2.4	9.8 ± 2.0	7.2 ± 1.7	4.5 ± 1.3	4.2 ± 1.2	F = 28.67	< 0.001*
	Sodium Hyaluronate (n = 25)	12.2 ± 2.5	10.5 ± 2.1	9.1 ± 1.9	7.6 ± 1.5	8.0 ± 1.6

* indicating that the difference is statistically significant (P < 0.05)

WOMAC score

Both groups of patients showed a significant decrease in WOMAC scores over time (P < 0.05). The UC-MSCs group had significantly lower pain (A), function (C), and total scores than the SH group at 3, 6, and 12 months (P < 0.05). There was no significant difference in stiffness (B) scores between the groups (P > 0.05), as shown in Table 2; Fig. 3C and F. In addition, unlike the good therapeutic effect of continuous improvement in WOMAC scores in the UC-MSCs group, the WOMAC scores in the SH group showed a trend of first decreasing and then increasing, indicating that their improvement in pain function gradually lost over time.

Secondary outcomes

Lequesne score

Both groups of patients showed a significant decrease in Lequesne scores over time (P < 0.05). The UC-MSCs group had significantly lower scores than the SH group at 6 and 12 months (P < 0.05), while there was no significant difference between the groups at 1 and 3 months (P > 0.05), as shown in Table 2; Fig. 3B.

Subjective satisfaction score

The satisfaction scores of UC-MSCs group at all time points were significantly higher than those of SH group (P < 0.05) (Table 3), while UC-MSCs group’s scores reached their peak at 3–6 months after treatment (8.3 ± 0.9 points), and remained at a high level at 12 months (8.1 ± 1.1 points). On the contrary, The scores in SH group reached peak at 6 months after treatment (5.5 ± 1.2 points) and slightly decreased at 12 months (5.3 ± 1.0 points). In subgroup analysis, the overall satisfaction score of K-L III patients was higher than that of K-L IV patients (compared within the same group), while UC-MSCs group was significantly better than SH group in both severity levels (P < 0.05), indicating that UC-MSCs have better subjective satisfaction with KOA of different severity levels.

Table 3.

Comparison of subjective satisfaction scores between two groups (x ± s)

Time Point	UC-MSCs Group (n = 25)	Sodium Hyaluronate Group (n = 25)	Independent Samples t-test		RM-ANOVA (Group×Time Interaction)	P-value (Group×Time)
			t Value	P-value
Baseline	2.1 ± 0.8	2.2 ± 0.7	0.452	0.653	F = 56.34	< 0.001*
1 Month Post-op	5.8 ± 1.2	4.2 ± 1.1	4.517	< 0.001*
3 Months Post-op	7.6 ± 1.0	5.1 ± 1.3	7.329	< 0.001*
6 Months Post-op	8.3 ± 0.9	5.5 ± 1.2	9.156	< 0.001*
12 Months Post-op	8.1 ± 1.1	5.3 ± 1.0	9.874	< 0.001*
Subgroup Analysis (12 Months Post-op)
KL Grade III	8.5 ± 0.8	5.6 ± 0.9	10.213	< 0.001*
KL Grade IV	7.5 ± 1.0	4.8 ± 1.1	7.952	< 0.001*

* indicating that the difference is statistically significant (P < 0.05)

Radiological outcomes

MOCART score

Both groups of patients showed a significant increase in MOCART scores over time (P < 0.05). The UC-MSCs group had significantly higher scores than the SH group at 6 and 12 months (P < 0.05), while there was no significant difference between the groups at 1 and 3 months (P > 0.05). At 12 months, there was no progression of cartilage defects in all patients in the UC-MSCs group, while 3 cases (12.0%) in the SH group showed mild enlargement of cartilage defects, as shown in Table 4; Fig. 4A. At 12 months, the UC-MSCs group demonstrated a mean improvement of 13.6 points from baseline, while the SH group showed an improvement of 7.8 points, confirming superior structural cartilage repair with UC-MSCs (Fig. 4B) (Table 5).

Table 4.

MOCART scores before and after injections (x ± s)

Time Point	UC-MSCs Group (n = 25)	Sodium Hyaluronate Group (n = 25)	Between-Group Comparison (P-value)	RM-ANOVA (Group×Time Interaction)	P-value (Group×Time)
Baseline	18.2 ± 5.3	17.6 ± 5.1	0.658	F = 28.45	< 0.001*
1 Month	22.5 ± 4.8	19.8 ± 4.5	0.032*
3 Months	26.3 ± 4.2	21.2 ± 4.0	0.001*
6 Months	30.5 ± 3.9	23.7 ± 3.8	< 0.001*
12 Months	31.8 ± 6.4	25.4 ± 5.7	0.002*

* indicating that the difference is statistically significant (P < 0.05)

Abbreviations: MOCART = Magnetic Resonance Observation of Cartilage Repair Tissue; RM-ANOVA = Repeated Measures Analysis of Variance; SD = Standard Deviation

Fig. 4.

A MOCART scores at different follow-up points after treatment; B The left knee MRI images of a 60 year old female (KL IV) before and after treatment: compared with before treatment (pre), the femoral bone marrow edema (red arrow) gradually decreased. The cartilage injury (yellow arrow) did not worsen during the 12-month follow-up period, and there was no significant change in knee joint effusion (green arrow). (1/T1 sagittal plane; 2/T2 sagittal plane; 3/T2 coronal position; 4/T2 cross-section.) (*: There was a significant difference compared to the previous time point (P < 0.05))

Table 5.

Remission rate of bone marrow edema and recovery of meniscus repair at 12 months after treatment

Outcome Indicator	Subcategory	UC-MSCs Group (n = 25)	Sodium Hyaluronate Group (n = 25)	Statistical Value	P-value
Bone Marrow Edema Remission¹	Femoral Bone Marrow Edema	20/25 (80.0%)	11/25 (44.0%)	χ²=7.28	0.007
	Tibial Bone Marrow Edema	15/25 (60.0%)	9/25 (36.0%)	χ²=3.24	0.070
	Patellar Bone Marrow Edema	9/25 (36.0%)	4/25 (16.0%)	χ²=2.56	0.110
Meniscus Repair²	Total Meniscal Tear (Stoller Grade ≥ II)	14/19 (76.0%)	6/18 (36.0%)	χ²=8.24	0.004
	Meniscal Tear (Stoller Grade III)	6/10 (60.0%)	1/7 (14.3%)	χ²=4.82	0.028

* indicating that the difference is statistically significant (P < 0.05)

Abbreviations: MRI = Magnetic Resonance Imaging; T2WI = T2-weighted imaging

Remission rate of bone marrow edema

At 12 months, the remission rate of femoral bone marrow edema in the UC-MSCs group was 80.0% (20/25), significantly higher than the 44.0% (11/25) in the SH group (χ ²=7.28, P = 0.007); The remission rates of patellar bone marrow edema in the two groups were 36.0% (9/25) and 16.0% (4/25), respectively, with no significant difference between the groups (χ ²=2.56, P = 0.11). Inter group comparison showed that the overall remission rate of tibial bone marrow edema in the UC-MSCs group (60.0%) was higher than that in the SH group (36.0%), but the difference was not statistically significant (χ ²=3.24, P = 0.07) (Table 5).

Effect of meniscus repair

The repair rate of meniscus injury (Stoller grade ≥ II) in both groups of patients increased over time, while the repair effect of the UC-MSCs group was significantly better than those of the SH group (Table 6).

Table 6.

Safety and adverse events

Adverse Event Type	UC-MSCs Group (n = 25)	Sodium Hyaluronate Group (n = 25)	χ² Value	P-value
Perioperative Adverse Events*
Knee Joint Pain	25/25 (100.0%)	17/25 (68.0%)	8.96	0.003
Knee Joint Swelling
- After 1st injection	21/25 (84.0%)	10/25 (40.0%)	11.52	0.001
- After last injection (4th/5th)	11/25 (44.0%)	3/25 (12.0%)	7.59	0.006
Fever (37.5–38.0 °C)	4/25 (16.0%)	0/25 (0.0%)	4.34	0.037
Postoperative Adverse Events†
Serious Adverse Events‡	0/25 (0.0%)	0/25 (0.0%)	-	-
Total Incidence of Adverse Events§	25/25 (100.0%)	19/25 (76.0%)	4.77	0.029

There was no significant difference in the distribution of meniscus injuries between the two groups at baseline (19 cases in the UC MSCs group, 18 cases in the sodium hyaluronate group). At 3 months after treatment, the overall repair rate between two groups was not statistically significant (χ ²=2.31, P = 0.129); At 6 months and 12 months, the repair rates of the UC-MSCs group were significantly higher than the SH group.

For the Stoller III grade meniscus tear, subgroup analysis showed that the repair rate of the UC-MSCs group was higher than that of the SH group at all time points, while the difference reached statistical significance at 12 months (P < 0.05), indicating that UC-MSCs have better repair ability for full-thickness meniscus tears than sodium hyaluronate.

Safety and adverse events

Both groups did not experience serious adverse events such as infection, severe allergic reactions, or tumor occurrence. The perioperative adverse events in the UC-MSCs group were mainly transient knee joint pain (100%), swelling (84.0% for the first injection and 44.0% for the fourth injection), and fever (16.0%). The SH group mainly experienced knee joint pain (68.0%) and swelling (40.0% for the first injection and 12.0% for the fifth injection), with no cases of fever. There was no significant difference in the incidence of adverse events between the two groups (P > 0.05), as shown in Table 6.

Discussion

This randomized controlled trial compared the clinical efficacy and radiological outcomes of multiple intra-articular injections of UC-MSCs versus sodium hyaluronate (SH) for severe knee osteoarthritis (K-L grade III–IV), and demonstrated that UC-MSCs achieve superior pain relief, functional improvement, and structural protection with a favorable safety profile over 12 months of follow-up. The results indicate that the therapeutic effects of UC-MSCs are more durable and stable compared with SH. Furthermore, radiological evaluation confirms that UC-MSCs confer significant advantages in structural repair and delaying degeneration. We conclude that for patients with K-L grade III–IV severe knee osteoarthritis, multiple intra-articular injections of UC-MSCs outperform traditional sodium hyaluronate injections, improving outcomes from both symptomatic relief and structural repair perspectives. This regimen is particularly suitable for patients with severe KOA who are unwilling to undergo total knee arthroplasty or who seek non-surgical alternatives to delay surgical intervention.

As a chronic low-grade inflammation, knee OA was characterized by cartilage degeneration and bone hyperplasia, which can lead to persistent pain and joint dysfunction [1, 18]. Sodium hyaluronate, as a commonly used conservative treatment drug in clinical practice, could temporarily alleviate symptoms by supplementing joint fluid viscoelasticity and reducing joint friction, thus may temporarily improve the joint function [19]. However, like most conservative treatments currently available, SH did not have the ability to delay degeneration, let alone promote cartilage repair [20]. Our results confirmed the advantages of multiple intra-articular injections of UC-MSCs in the treatment of severe knee osteoarthritis by comparing with sodium hyaluronate, suggesting that UC-MSCs provide a new direction for KOA treatment from the perspectives of tissue repair and anti-inflammatory effects.

Clinical efficacy and safety

Our study suggested that both UC-MSCs and SH can improve pain symptoms and joint function in KOA patients, which was consistent with the results of most studies using mesenchymal stem cells (MSCs) to treat knee osteoarthritis [21–23]. The randomized controlled study conducted by ZoghAli et al. confirmed the efficacy of UC-MSCs compared to corticosteroid for knee osteoarthritis, including significant improvement in VAS scores and knee mobility [24]. However, their relatively short follow-up period (24 weeks) resulted in insufficient evaluation of long-term efficacy and imaging outcomes. Zhao et al.‘s retrospective study applied microfracture combined with UC-MSCs injection or UC-MSCs injection alone to treat tiny cartilage defects in the knee joint, achieving satisfactory therapeutic effects in the short term [25]. The single-arm clinical trial conducted by Dilogo et al. also confirmed that UC-MSCs had satisfactory clinical efficacy for KOA, while this efficacy was most evident at 6 months after injection [26]. In our study, the UC-MSCs group demonstrated better and more lasting pain relief and functional improvement compared to the SH group. In terms of pain control, the UC-MSCs group indicated a more significant decrease without rebound, while the VAS score of SH group increased again after 6 months. At 12 months, 92.0% of patients in the UC-MSCs group achieved VAS score ≤ 3, which was much higher than the 64.0% in the SH group. In terms of function improvement, the WOMAC scores and Lequesne score of the UC-MSCs group were significantly lower than those of the SH group at 3, 6, and 12 months (P < 0.05), while the therapeutic effect could last until 12 months after treatment. Meanwhile, the differences in therapeutic efficacy were also reflected in subjective satisfaction. The satisfaction scores of the UC-MSCs group at all time points (peak 8.3 points) were significantly higher than those of the SH group (peak 5.5 points).

As for the comparison of radiological evaluations, UC-MSCs have significant advantages in structural repair and delayed degeneration. The case report by Song et al. shows that using UC-MSCs intra-articular injection to treat patellar cartilage defects can not only improve symptoms and function for a long time, but also observe significant cartilage regeneration images 18 months after surgery [27]. Hernigou et al.‘s RCT not only proposed that MSCs injection can delay or avoid TKA of the contralateral joint in patients with bilateral osteoarthritis, but also demonstrated through imaging evaluation that bone marrow edema significantly subsided after subchondral bone injection of MSCs [28]. Our research further confirmed that UC-MSCs could achieve satisfactory relief of bone marrow edema. The research results of Fiorin et al. supported that intra-articular injection of UC-MSCs can effectively improve knee joint function and cartilage contour in KOA at 6 and 12 months, achieving satisfactory imaging results [29]. Our findings were consistent with most current research. In addition, both groups demonstrated improvement in MOCART scores, while the UC-MSCs group showed a significantly greater increase than SH group. All patients in the UC-MSCs group had no progression of cartilage defects, while 12.0% of patients in the SH group had enlarged defects, confirming that UC-MSCs can effectively delay cartilage degeneration compared to SH. The meniscus repair rate in the UC-MSCs group was 76.0% (grade III tear: 60.0%), significantly higher than that in the SH group (overall 36.0%, grade III tear 14.3%), suggesting a promotion of the recovery of damaged meniscus structure. The remission rate of femoral bone marrow edema was significantly higher in the UC-MSCs group, and a similar trend was observed for tibial bone marrow edema (44.0% for femur and 36.0% for tibia), although the difference did not reach statistical significance, somehow verifying the potential advantages of UC-MSCs in promoting the resolution of bone marrow edema.

In terms of clinical safety, UC-MSCs were equally safe as sodium hyaluronate and posed no serious risks. In our study, neither group of patients experienced serious adverse events such as infection, allergies, or tumors. The main adverse reactions in the UC-MSCs group were transient knee pain (100%), swelling (84.0% after the first injection → 44.0% after the fourth injection), and low-grade fever (16.0%), which were relieved by NSAIDs or spontaneously. During the follow-up process, no serious adverse reactions were found in all patients, confirming the clinical safety of UC-MSCs. Although clinical studies on MSCs from other sources have also confirmed their safety [30], the complications and adverse reactions caused by MSCs acquisition should not be neglected. Sajadi et al.‘s double-blind clinical study [23] confirmed the clinical efficacy of adipose derived mesenchymal stem cells (AD-MSCs) in KOA, while also reported adverse reactions after acquisition, including transient inflammation and rash. Kim et al. reported on similar complications related to acquisition in a phase III randomized double-blind controlled trial [31]. As the collection of AD-MSCs is mainly achieved through liposuction or fat extraction, in addition to complications such as local pain, bleeding, scar formation, serious or lethal complications like infection, fat embolism may also be caused by improper intraoperative procedures [32]. The research of Toan et al. indicated that after the extraction of autologous bone marrow stem cells (BMSCs) [33], all participating patients had no acute or chronic complications. However, whether it was extracting iliac bone or extracting bone marrow from the iliac spine, pain and bleeding in the acquisition area should be inevitable, which can affect the early functional exercise and activity. We suggested that compared to MSCs from other sources, UC-MSCs can achieve satisfactory clinical efficacy while minimizing acquisition-related complications.

Potential mechanisms underlying the therapeutic effects of UC-MSCs on KOA

Of note, all mechanistic interpretations in this section are inferred from published preclinical studies and existing literature, and are not experimentally validated in the present clinical trial. The following discussion is provided to offer a plausible biological context for the clinical and radiological findings observed in our cohort.

For a long time, osteoarthritis was regarded as a simple “wear and tear” disease [34, 35], and MSCs were traditionally thought to exert therapeutic effects by differentiating into chondrocytes to directly replace damaged tissue [36]. However, accumulating preclinical and translational studies have proposed that immunomodulatory and anti-inflammatory effects mediated primarily by paracrine actions may represent a core mechanism underlying the clinical benefits of MSCs in KOA [37]. Pathologically, KOA is a degenerative disease characterized by articular cartilage degradation, chronic synovial inflammation, and subchondral bone remodeling [38], driven by the synergistic effects of mechanical stress, inflammatory imbalance, and oxidative damage [39].

As a chronic low-grade inflammatory condition, KOA progression is mainly driven by pro-inflammatory factors such as IL-1β, TNF-α, and IL-6 secreted by M1 macrophages and synovial fibroblasts [40]. These factors can activate the NF-κB pathway and upregulate the expression of matrix metalloproteinases (MMP-13, ADAMTS-5), further accelerating cartilage matrix degradation [41]. UC-MSCs are hypothesized to modulate local immune homeostasis via two complementary pathways, as evidenced by in vitro and in vivo experimental studies: on the one hand, CD200 and PD-L1 expressed on UC-MSCs have been shown to bind CD200R and PD-1 on the surface of M1 macrophages, inhibiting their pro-inflammatory activation [42]; on the other hand, IL-1RA secreted by UC-MSCs may competitively bind to IL-1 receptors, blocking the IL-1β-mediated inflammatory cascade, and molecules such as TSG-6 and PGE2 may promote the polarization of M1 macrophages toward the M2b phenotype [43]. While M2b macrophages may secrete small amounts of IL-6, their high expression of IL-10 could significantly inhibit synovitis and cartilage-degrading enzyme activity [44].

Preclinical studies have provided supporting evidence for these anti-inflammatory effects [45]. Xiao et al. demonstrated in a mouse KOA model that hUC-MSCs significantly reduced serum levels of IL-1 and IL-1β, as well as cartilage levels of MMP-1 and MMP-13 [46]. Deng et al. reported that both single and multiple intra-articular injections of UC-MSCs reduced periarticular osteophyte formation and articular cartilage degeneration in a rabbit KOA model, while increasing chondrogenic marker expression and reducing synovial TNF-α and IL-6 levels [47]. Klyucherev et al. showed that both UC-MSC-derived exosomes and matrix-bound nanovesicles downregulated inflammatory markers in macrophages and synovial fibroblasts [48]. Tong et al. also found that repeated UC-MSC administration attenuated KOA progression in rats by reducing inflammatory cell infiltration and synovial hyperplasia [49].

Regarding cartilage protection and regeneration, previous studies suggest that UC-MSCs may exert bidirectional effects by inhibiting cartilage degradation and promoting matrix synthesis [50]. Growth factors such as IGF-1 and TGF-β secreted by UC-MSCs may activate the Smad2/3 pathway in chondrocytes [51], promoting the synthesis of type II collagen and proteoglycans, while inhibiting chondrocyte hypertrophy and cartilage calcification. In vivo work by Zhou et al. indicated that UC-MSC-derived extracellular vesicles delayed OA progression, reduced osteophyte formation, and increased cartilage matrix protein expression, while enhancing chondrocyte proliferation and migration and inhibiting apoptosis in vitro [52]. Oxidative stress-induced chondrocyte apoptosis is an important cause of cartilage loss [42], and UC-MSCs have been shown to delay KOA progression by inhibiting chondrocyte apoptosis in multiple animal models [43].

For subchondral microenvironment remodeling, subchondral bone sclerosis and osteophyte formation can exacerbate uneven mechanical stress in the joint. UC-MSCs may break the vicious cycle of “injury-degeneration” in KOA by regulating subchondral bone metabolism and improving tissue permeability [53, 54]. Furthermore, BMP-2 and OPG secreted by UC-MSCs may promote osteoblast activation and inhibit osteoclast activity, increasing subchondral bone volume fraction and trabecular thickness to further reduce cartilage loading pressure [55].

It is also worth noting that the therapeutic effect of UC-MSCs appears to be dose-dependent, with individual variability [56]. A phase I/II randomized controlled trial by Matas et al. showed that repeated MSC dosing achieved more satisfactory clinical efficacy than a single dose [57], and preclinical work by Deng et al. also demonstrated that multiple injections yielded better subchondral bone repair than single administration [47], which aligns with the design rationale of our multiple-injection regimen.

Overall, these proposed mechanistic pathways are consistent with the sustained symptomatic relief and structural improvement observed in our clinical cohort. However, the specific molecular mechanisms underlying the efficacy of intra-articular UC-MSCs for severe KOA remain to be further validated by dedicated basic research.

This study has several inherent limitations that should be acknowledged. First, it was conducted at a single center with a relatively small sample size of 50 patients, which may limit the generalizability of the findings and reduce the statistical power for subgroup analyses. Second, the follow-up duration was restricted to only 12 months; longer-term observation over 2–5 years is still required to further verify the sustained efficacy of UC-MSCs, the long-term progression of cartilage degeneration, and the potential oncogenic risk of stem cell application. Third, only a fixed cell concentration and single injection protocol of UC-MSCs were adopted in the present study. We did not set gradient cell doses or different injection intervals; thus, the optimal therapeutic dose and administration strategy could not be determined. Fourth, the UC-MSCs in this trial were expanded in FBS-supplemented medium rather than a fully xeno-free, chemically defined culture system. Although strict quality testing was performed to exclude microbial contamination and control endotoxin levels, the use of xenogeneic serum still introduces potential risks of batch-to-batch variation and residual immunogenicity, which may hinder standardized GMP-compliant production and further clinical translation of the cell product. Follow-up studies will adopt GMP-grade xeno-free culture systems to optimize the manufacturing protocol and enhance the translational potential of UC-MSCs for KOA treatment.

From the perspective of clinical translation and marketing authorization, further optimization of the manufacturing and quality control system is required to align with regional regulatory requirements. In Europe, centralized evaluation by EMA-CAT mandates full GMP compliance, consistent batch-to-batch quality, and complete traceability for all advanced cell therapies. In China, formal marketing approval of cell therapy products falls under the jurisdiction of the NMPA, with parallel clinical application specifications issued by the NHC.

Building on the findings of this study, our follow-up work will transition to xeno-free, chemically defined GMP-grade culture systems to eliminate risks associated with xenogeneic serum and improve product reproducibility. Meanwhile, we will further standardize the full workflow of cell isolation, culture, quality inspection and clinical administration to fully meet the regulatory standards of both domestic and international agencies, so as to accelerate the translational application of UC-MSCs in the management of severe knee osteoarthritis.

Conclusions

Compared with intra-articular sodium hyaluronate injection, multiple intra-articular injections of UC-MSCs yielded superior and long-lasting improvements in pain relief and joint function among patients with Kellgren-Lawrence grade III–IV severe knee osteoarthritis. In terms of radiological structural repair, UC-MSCs demonstrated prominent advantages in preventing cartilage defect progression, alleviating femoral bone marrow edema, and promoting the healing of meniscus injuries, especially full-thickness Stoller grade III meniscal tears. Meanwhile, the overall safety profile of UC-MSCs is comparable to sodium hyaluronate, with no severe adverse events observed during the 12-month follow-up. This regimen provides an innovative and effective alternative to conventional conservative treatment for advanced knee osteoarthritis. It is particularly suitable for patients with severe KOA who are unwilling to undergo total knee arthroplasty or seek non-surgical alternatives to delay surgical intervention. Nevertheless, this study was limited by its single-center design, relatively small sample size, and only one-year follow-up duration. Further multicenter, large-sample randomized controlled trials with prolonged follow-up are warranted to validate the long-term efficacy, optimal injection frequency and cell dosage of UC-MSCs. In-depth mechanistic investigations will also help standardize the clinical application protocol of UC-MSCs and facilitate its widespread translation in the management of severe knee osteoarthritis.

Author contributions

C.C. conceived and designed the trial, performed clinical outcome assessments, analyzed data, and drafted the main manuscript text. S.W. and X.L. recruited participants, collected clinical data, and conducted patient follow-up. X.M. and G.C. performed the isolation, culture, and quality control of UC-MSCs. J.D. conducted statistical analyses and prepared all figures and tables. F.W. supervised the entire study, critically revised the manuscript for intellectual content, and approved the final version for submission. All authors reviewed and approved the final manuscript.

Funding

The author(s) declare that financial support was received for the research and/or publication of this article. The authors acknowledge the funding provided by The National Natural Science Foundation of China (52335006) and Chongqing Organ Intelligent Biological Manufacturing Engineering Research Center Open Project (CQERC-OIBM20241018).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

This study protocol was registered on ClinicalTrials.gov (Registration Number: NCT05160831) and approved by the Ethics Committee of the Army Medical University (Approval Number: KY2021087, August 7, 2021). The study was strictly conducted in compliance with the Good Clinical Practice (GCP) guidelines and the Helsinki Declaration. All participants provided written informed consent after thoroughly reviewing the study protocol.

Consent for publication

All authors provided informed consent for the publication of the article.

Competing interests

The authors declare no competing interests.