Evidence designation: Condition-specific. No class-wide designation is scientifically appropriate. Most streams are investigational—human clinical evidence; direct ligament regeneration remains preclinical or very early, and CARTISTEM has a product-specific South Korean authorization for a defined cartilage indication. [3–24]
Definitions and terminology
Musculoskeletal medicine includes disorders of joints, cartilage, tendons, ligaments, menisci, bone, and the intervertebral disc; these tissues differ sufficiently that an intervention studied in one cannot be presumed effective in another. [1–21]
The scientifically preferred term mesenchymal stromal cells (MSCs) describes culture-adherent, multipotent stromal populations meeting defined minimal laboratory criteria, whereas “mesenchymal stem cells” remains common but may imply stemness that has not been demonstrated for every administered cell. [22]
Autologous products originate from the treated patient, whereas allogeneic products originate from a donor; stromal vascular fraction (SVF), bone-marrow aspirate concentrate (BMAC), culture-expanded MSCs, and proprietary mesenchymal precursor cells are compositionally and operationally different interventions. [3–8,11–21]
For this topic, investigational means that human clinical evidence exists but remains insufficient for routine standard-care designation; it does not mean that the intervention has been proved effective, ineffective, safe, unsafe, legitimate, or illegitimate. [23,24]
Biological rationale
MSCs may influence injured musculoskeletal tissues through paracrine signaling, immunomodulation, altered inflammatory-cell activity, angiogenic signaling, extracellular-matrix regulation, and support of resident repair cells. [22,25]
Direct, durable replacement of articular chondrocytes, tenocytes, ligament fibroblasts, osteoblasts, or nucleus-pulposus cells after clinical administration is less securely demonstrated than short-lived trophic or signaling effects. [22,25]
The rationale therefore varies by indication: symptom modulation may be relevant in osteoarthritis, scaffold-assisted tissue formation in focal cartilage or bone defects, tendon-to-bone healing after rotator-cuff repair, and inflammatory or matrix effects in disc degeneration. [4–21] In this context, the extracellular matrix includes the structural environment of musculoskeletal tissues; the nucleus pulposus is the central disc tissue discussed in disc applications, and synovial inflammation is one proposed osteoarthritis treatment target.
Biological plausibility supports clinical testing, but it cannot establish the magnitude, durability, or patient importance of a treatment effect. [4–21]
Product identity and characterization
A clinically interpretable MSC report should identify tissue source, donor relationship, isolation and culture method, passage or population doublings, fresh or cryopreserved status, viability, identity, purity, potency testing, dose, carrier, route, and concomitant procedure. [22,25]
These details are often incomplete, and products described by the same broad name may differ in cellular composition, secretory activity, viability after thawing, senescence, immunogenicity, and functional potency. [22,25]
SVF contains a heterogeneous mixture of stromal, vascular, immune, and other cells and should not be treated as equivalent to a defined culture-expanded MSC product. [7,12]
Likewise, BMAC contains relatively few MSCs among many hematopoietic and marrow-derived components, while proprietary mesenchymal precursor-cell products use manufacturer-specific selection and manufacturing processes. [13,16–18]
Mechanism of action
In osteoarthritis, proposed actions include reduction of synovial inflammatory signaling, modulation of pain-associated mediators, and support of cartilage homeostasis, but clinical imaging changes have not established generalized regeneration of durable hyaline cartilage. [3–8]
In focal cartilage repair, cells are commonly delivered with microfracture, fibrin, or a matrix, so marrow recruitment, scaffold mechanics, rehabilitation, and the implanted cells may all contribute to the result. [9,10]
In tendon and ligament applications, proposed mechanisms include altered inflammation, collagen organization, vascular responses, and tendon-to-bone integration; the human evidence does not establish routine regeneration of a torn ligament or tendon by injection alone. [11,12,14]
In bone defects and nonunion, implanted cells may provide osteogenic activity and trophic support, but fixation, debridement, infection control, defect geometry, graft, scaffold, and mechanical stability remain major determinants of union. [19–21]
In disc disease, proposed benefits include modulation of inflammation and matrix metabolism, yet survival in the hypoxic, acidic, nutrient-poor disc environment and the relationship between imaging and pain remain uncertain. [15–18]
Why Products Called “MSCs” Differ
Products described as “stem cells” or MSCs can begin with different tissues and follow very different manufacturing paths. SVF and BMAC are heterogeneous mixtures, whereas culture-expanded products contain a more selected stromal population; donor-derived and scaffold-based products add further variables. Identity, viability, sterility, endotoxin, and potency testing may also differ. These distinctions matter because clinical evidence belongs to the product that was actually studied. A result obtained with one source, processing method, dose, carrier, or route should not be transferred automatically to another.
Preclinical evidence
Laboratory and animal models generally support chondrogenic, tenogenic, ligament-healing, osteogenic, and disc-matrix hypotheses, but model injury, cell dose, species, and outcome timing frequently differ from human disease. [4,14,19,22,25]
A rabbit anterior-cruciate-ligament study, for example, reported enhanced macroscopic and histologic repair when autologous MSCs were seeded on a collagen scaffold, but it cannot establish human ligament regeneration, clinical benefit, or safety. [26]
Preclinical evidence is therefore best used to justify product-specific clinical trials and mechanistic measurements rather than to predict clinical success. [22,25,26]
Human clinical evidence
Knee osteoarthritis has the largest musculoskeletal evidence base, but results are inconsistent across products and designs. [3–8]
Figure 1 placeholder — Knee osteoarthritis
Small uncontrolled dose-escalation studies of autologous adipose-derived expanded cells reported symptom improvement and, in some cohorts, structural signals, while a 60-participant randomized dose-ranging study of pooled allogeneic bone-marrow MSCs described an efficacy trend primarily in its 25-million-cell arm rather than a consistent dose response. [3–5]
A 39-participant placebo-controlled SVF trial reported greater WOMAC improvement than placebo, but SVF is heterogeneous and the sample was small. [7]
The 135-participant ADIPOA2 trial found strict OARSI/OMERACT responder rates of 47.3% with autologous expanded adipose MSCs and 54.8% with placebo at six months, yielding a relative risk of 0.86 (95% CI 0.58–1.28; P=.46), with no significant overall secondary between-group differences. [8]
Hip osteoarthritis currently has a much thinner and more heterogeneous human evidence base, so knee findings should not be transferred to the hip. [27]
Figure 2 placeholder — Hip osteoarthritis
For focal knee cartilage defects, a 14-participant randomized pilot comparing matrix-induced autologous MSC implantation with autologous chondrocyte implantation and an 80-participant trial adding adipose-derived cells and fibrin to microfracture reported favorable signals, but both evaluated multicomponent surgical strategies. [9,10]
Figure 3 placeholder — Cartilage defects
Rotator-cuff studies include a 24-participant randomized injection trial and a 20-participant dose-escalation study of autologous adipose-derived MSCs; their small samples and differing products, controls, tear types, and delivery strategies preclude a routine-care conclusion. [11,12]
Figure 6 placeholder — Rotator-cuff disease
The principal randomized meniscal study enrolled 55 patients after partial medial meniscectomy and reported limited MRI meniscal-volume and subgroup pain signals after allogeneic bone-marrow MSC injection; it does not establish benefit for untreated traumatic or degenerative meniscal tears. [13]
Figure 7 placeholder — Meniscal injury
A 17-participant post-ACL-reconstruction pilot found no cell-related serious adverse events and exploratory symptom and structural signals after 75 million allogeneic mesenchymal precursor cells in hyaluronan, but it studied post-traumatic osteoarthritis modulation rather than regeneration of a torn ACL. [14]
In disc disease, small randomized studies reported responder signals after intradiscal allogeneic bone-marrow MSC or mesenchymal precursor-cell products, whereas the larger RESPINE trial did not conclusively demonstrate efficacy and a 404-participant program did not meet its primary and secondary efficacy endpoints across all participants. [15–18]
Figure 10 placeholder — Intervertebral-disc disease
In delayed union, nonunion, and pseudoarthrosis, small studies report feasibility and union, but most combine cells with fixation, surgery, PRP, demineralized bone matrix, bioceramics, or allogeneic bone scaffold. [19–21]
Figure 9 placeholder — Nonunion
A 20-participant phase IIa pseudoarthrosis trial found no significant CT consolidation difference between an autologous expanded MSC tissue-engineered product and iliac-crest autograft, supporting feasibility more clearly than superiority. [21]
Trial design and limitations
The recurrent limitations are small samples, open-label designs, baseline-to-follow-up comparisons, numerous outcomes, incomplete product reporting, short safety surveillance, and selective emphasis on favorable doses, time points, or subgroups. [3–21]
An improvement from baseline is not proof of efficacy because symptoms fluctuate and may improve through regression to the mean, concurrent rehabilitation, procedural context, or placebo effects. [3,4,6,12,15,19,20]
Imaging endpoints require particular restraint: cartilage thickness, MOCART scores, meniscal volume, disc grade, and CT density are useful measurements but do not automatically demonstrate normal tissue composition, mechanical function, durability, or patient benefit. [3,7,9,10,13,15,21]
Negative primary outcomes must remain primary in the synthesis; exploratory subgroups, post hoc thresholds, or secondary outcomes may generate hypotheses but should not be promoted as if they reversed the main result. [8,16,18,21]
Musculoskeletal Evidence by Condition
The evidence must be interpreted by condition rather than by the broad label “MSC therapy.” Knee osteoarthritis and intradiscal treatment have randomized human trials, but their results are mixed. Focal cartilage repair is usually combined with surgery, while tendon, rotator-cuff, meniscal, ligament, bone-defect, and nonunion studies remain limited or procedurally confounded. The colors organize the illustration only; they do not rank treatments or indicate recommendations. Each evidence statement applies to its named condition, tested product, study design, comparator, outcomes, and follow-up.
Manufacturing variables
Manufacturing can change the administered biological product through donor selection, tissue harvest, enzymatic digestion, density separation, plastic adherence, media supplements, oxygen tension, culture duration, passage, washing, formulation, cryopreservation, thawing, and transport. [22,25]
Release testing should address identity, purity, viability, sterility, endotoxin, mycoplasma, adventitious agents, and a potency assay plausibly connected to the intended mechanism. [22,25]
A cell count without viable-cell recovery, assay method, and product composition is not a complete dose description. [22,25]
Poor product comparability is a central reason that a positive or negative trial should not be generalized to every preparation marketed as MSCs. [3–21]
Dose and administration
Reported knee-OA doses range from approximately 2 million to 150 million cells per intra-articular administration, with no reproducible class-wide dose-response relationship. [3–8]
Other applications use intratendinous, intraosseous, intradiscal, scaffold-based, or surgically implanted delivery, and these routes create different retention, mechanical, procedural, and safety considerations. [9–21]
Carriers such as hyaluronan, fibrin, bioceramics, collagen, or bone scaffold may influence both biological behavior and outcomes. [9,10,14,17–21,26]
The current evidence does not establish one optimal source, dose, interval, route, number of treatments, rehabilitation program, or retreatment schedule across musculoskeletal indications. [3–21]
Safety and contraindications
Across the small trials reviewed, transient local pain, swelling, or arthralgia was commoner than serious treatment-attributed toxicity, but the combined sample sizes and follow-up remain inadequate to exclude uncommon infection, immune, thrombotic, ectopic-tissue, neoplastic, or procedure-related events. [3–21]
The post-ACL study observed temporary increases in anti-HLA reactivity in some allogeneic-cell recipients, illustrating that absence of immediate clinical rejection does not make allogeneic products immunologically inert. [14]
Intradiscal injection adds risks related to needle access and disc infection; surgical scaffold implantation adds anesthesia, wound, fixation, and graft-related risks that should not be attributed solely to the cells. [15–21]
Potential contraindications and precautions depend on the exact product and procedure and may include active infection, uncontrolled malignancy, significant immunologic risk, coagulopathy or anticoagulation incompatible with the procedure, pregnancy exclusions, severe comorbidity, and inability to complete rehabilitation; trial criteria must not be converted automatically into universal clinical rules. [3–21,23]
Regulatory status
In the United States, FDA states that regenerative-medicine therapies have not been approved to treat orthopedic conditions including osteoarthritis, tendonitis, disc disease, and back, hip, knee, neck, or shoulder pain. [23]
That statement concerns US marketing authorization and does not itself prove that every studied product is ineffective; conversely, an investigational trial or facility registration is not authorization for routine commercial marketing. [23]
South Korea’s MFDS identifies CARTISTEM as an allogeneic umbilical-cord-blood-derived MSC product, 7.5 million cells per vial, approved in January 2012 for specified ICRS grade IV knee articular-cartilage defects associated with osteoarthritis or repeated trauma and administered surgically or arthroscopically. [24]
CARTISTEM’s status cannot be generalized to other cord-tissue products, Wharton’s-jelly preparations, injections, doses, indications, or jurisdictions. [24]
European classification of substantially manipulated or nonhomologously used cell products as advanced-therapy medicinal products is product- and use-specific and is not equivalent to a marketing authorization. [28]
Before publication, every additional jurisdictional statement must be verified against the relevant regulator’s primary product database and dated; an unresolved search must be labeled unresolved rather than converted into a claim of approval or nonapproval. [23,24,28]
Clinical interpretation
The most defensible current designation for knee osteoarthritis, focal cartilage repair outside a specifically authorized product, tendon and rotator-cuff disease, meniscal use, bone defects, nonunion, and disc disease is investigational—human clinical evidence. [3–21]
Direct human ligament-regeneration evidence remains preclinical or very early, while hip-OA evidence remains especially sparse. [14,26,27]
The evidence supports continued carefully controlled clinical development, but it does not support the statement that MSCs as a class regenerate musculoskeletal tissues or provide a proven alternative to established care. [3–21]
Standard-care alternatives remain condition-specific and may include education, activity modification, weight management, physical therapy, analgesic or anti-inflammatory strategies, bracing, selected injections, repair or reconstruction, osteotomy, fixation or bone grafting, and arthroplasty; comparison must be made against the appropriate alternative for the individual condition and stage. [29–32]
A treatment decision should therefore begin with diagnosis, severity, prognosis, standard alternatives, patient goals, and the exact product—not with the generic label “stem cells.” [3–32]
Knowledge gaps
Priority gaps include adequately powered multisite randomized trials; credible sham or active comparators; prespecified clinically meaningful responder definitions; complete adverse-event denominators; and follow-up long enough to assess durability and delayed harm. [3–21]
Trials also need harmonized reporting of donor and manufacturing variables, viable administered dose, potency, concomitant procedures, rehabilitation, rescue treatment, and failures. [22,25]
Mechanistic studies should connect validated product attributes with clinical outcomes rather than assuming that surface markers or in-vitro differentiation predict patient benefit. [22,25]
Finally, replication must occur with materially comparable products; otherwise, both favorable and unfavorable results remain confined to the tested intervention. [3–25]
Bottom-line evidence statement
Musculoskeletal MSC research contains credible human signals, important null results, and substantial uncertainty. [3–21]
Knee osteoarthritis and intradiscal therapy have progressed furthest into randomized development, yet neither supports a broad class-wide efficacy claim, and focal cartilage, tendon, meniscal, ligament, bone, and nonunion applications remain constrained by small studies or multicomponent procedures. [3–21]
Regulatory authorization must be stated by exact product, indication, route, jurisdiction, and verification date, while scientific conclusions must continue to rest on study design, effect size, reproducibility, safety, and product comparability. [23,24,28]
Compact evidence summary
This table is a navigation aid, not a substitute for the study-level evidence appendix or the cited primary reports.
Complete supporting evidence: The public Topic 8.2 evidence appendix will be linked here after assignment of its permanent publication URL.
Supporting evidence file: 0802_Public_Evidence_Appendix.xlsx — permanent website URL to be assigned during publication.
From Biological Rationale to Clinical Proof
A plausible mechanism is the beginning of clinical research, not its conclusion. Cell signaling and laboratory findings may justify animal studies and early human trials, but reliable clinical interpretation requires adequately powered controlled studies with prespecified outcomes. Product changes, different doses or routes, and different patient populations can break the chain of inference. Imaging may provide useful structural information, yet it is not automatically a direct measure of pain, function, durability, healing, or safety. Patient-important outcomes and adverse-event surveillance remain essential.
Figure 4 placeholder — Tendon injury
Figure 5 placeholder — Ligament injury
Figure 8 placeholder — Bone defects