Intradiscal mesenchymal stem cell therapy for degenerative disc disease: a systematic review and meta-analysis of randomized trials
Article information
Abstract
Study Design
Systematic review and meta-analysis.
Purpose
To investigate the efficacy and safety of intradiscal mesenchymal stem cell (MSC) injection for patients with degenerative disc disease (DDD).
Overview of Literature
DDD is a common cause of chronic low-back pain (CLBP). Studies have shown that MSC therapy may have the potential to reverse the degenerative process. However, the evidence remains inconclusive.
Methods
A comprehensive search of Europe PMC, Cochrane Library, Scopus, and Medline was conducted up to April 30, 2025, using relevant keywords. Randomized controlled trials (RCTs) comparing intradiscal MSC injections with sham/placebo in patients with DDD were included. Random-effects models were used to calculate the odds ratios and mean differences (MDs).
Results
Seven RCTs were included. MSC injection was associated with significantly greater reductions in Visual Analog Scale pain scores (MD, 6.67; 95% confidence interval [CI], −9.31 to −4.02; p<0.00001, I2=29%) and Oswestry Disability Index (ODI) (MD, −4.05%; 95% CI, −5.24 to −2.87; p<0.00001, I2=32%) compared with sham/placebo. The magnitude of improvement increased with longer follow-up. No significant differences were observed between groups in treatment-emergent adverse events (AEs), serious AEs, treatment discontinuation due to AEs, or mortality.
Conclusions
Intradiscal MSC injection appears to be a safe and effective therapy, offering modest improvements in pain and function for patients with CLBP due to DDD who do not respond to conservative treatments.
Introduction
Low-back pain (LBP) is a highly prevalent condition worldwide [1]. In 2020, an estimated 619 million individuals were affected, with projections rising to 843 million by 2050 [1]. Beyond pain, LBP is associated with depression, anxiety, reduced daily activity, and occupational limitations, all of which impair quality of life [2].
LBP is commonly categorized by symptom duration: acute (<6 weeks), subacute (6–12 weeks), or chronic (>12 weeks) [3]. A major cause of chronic LBP (CLBP) is degenerative disc disease (DDD), characterized by deterioration of intervertebral discs due to aging or injury [3]. DDD occurs more frequently in the lumbar than the cervical spine, and its prevalence increases with age [3]. A cohort study reported DDD in 71% of men and 77% of women under 50 years, increasing to >90% in both sexes over 50 years [4].
The first line of management of DDD is conservative, including lifestyle modification (e.g., weight reduction), physiotherapy, analgesics, including acetaminophen and nonsteroidal anti-inflammatory drugs, as well as epidural injections [5,6]. If these measures fail, surgical intervention is considered [5,6]. However, surgery is often met with resistance due to concerns about complications, prolonged recovery, and substantial costs [7]. Consequently, stem cell-based therapy was formulated as a potential alternative for patients with DDD who do not respond to conservative treatment and wish to avoid surgery.
Mesenchymal stem cells (MSCs) are skeletal progenitor cells with the capability to differentiate into multiple cell lineages, including osteocytes, chondrocytes, and adipocytes [8,9]. MSCs can be isolated from diverse tissues, most notably bone marrow, adipose tissue, and umbilical cord blood [8,9]. In the context of disc degeneration, MSCs can exert paracrine factors that reduce inflammation, promote collagen and proteoglycan synthesis, inhibit aberrant nerve ingrowth, and facilitate analgesia, thereby offering potential clinical benefit in patients with DDD [8,9]. However, evidence regarding the efficacy of intradiscal MSC injection remains ambiguous. This study aims to systematically evaluate the efficacy and safety of MSC therapy in patients with DDD.
Materials and Methods
Eligibility criteria
This review was conducted in accordance with the PRISMA (Preferred Reporting Items for Systematic reviews and Meta-Analyses) statement guidelines [10]. The protocol was registered in PROSPERO (CRD420251067756). Studies were included if they met the following PICOS-based inclusion criteria: (1) Population: patients with CLBP primarily attributed to DDD and unresponsive to conservative (non-surgical) treatments; (2) Intervention: intradiscal injection of MSCs; (3) Control: any therapeutic intervention other than MSCs, including a placebo; (4) Outcomes: efficacy parameters: Visual Analog Scale (VAS) pain score, Oswestry Disability Index (ODI), proportion of patients achieving ≥30% and ≥50% reduction in VAS scores, and proportion achieving ≥10- and ≥15-point reduction in ODI; safety parameters: treatment-emergent adverse events (TEAEs), serious adverse events (SAEs), treatment discontinuation due to adverse events (AEs), and mortality; and (5) Study design: randomized controlled trials (RCTs).
Studies were excluded if they met any of the following criteria: (1) conducted in pediatric populations; (2) included patients with a history of spinal fusion, disc replacement, or other surgical procedures; (3) investigated CLBP not attributable to DDD; (4) used cell-based therapies other than stem cells or progenitor cells; (5) reported insufficient data on the pre-specified outcomes; (6) lacked a comparison group; (7) published only as pre-prints; and (8) non-primary studies (e.g., reviews, editorials, conference abstracts).
Search strategy and study selection
A comprehensive literature search was conducted without language restrictions, covering studies published up to April 30, 2025. Four major databases were searched: Europe PMC, Medline, Scopus, and the Cochrane Library. The search strategy used the following terms (Supplement 1): “(mesenchymal stem cells OR mesenchymal stromal cells OR mesenchymal precursor cells OR mesenchymal progenitor cells) AND (degenerative disc disease OR disc degeneration OR chronic back pain) AND (trial OR clinical trial OR randomized trial OR randomized controlled trial OR RCT).” Two authors independently conducted the initial search and verified references in the reference manager to ensure consistency and completeness. Citation tracking was also performed by reviewing bibliographies, following citation links, and exploring pertinent articles. Gray literature sources were additionally screened. Title and abstract screening were independently performed by two authors, with irrelevant studies excluded. Full-text articles were then assessed for eligibility by the same two reviewers. Any discrepancies were resolved through discussion with a third author.
Data extraction
Two authors independently extracted key information from the eligible studies, including participant characteristics (age, sex distribution, body mass index [BMI]) and study characteristics (first author, year of publication, country, intervention details, type of control, follow-up duration, and sample size).
For the purpose of this review, MSCs were defined as adult multipotent stem cells found in various connective and other tissues, with the capacity to differentiate into cell lineages such as osteoblasts, chondrocytes, and adipocytes. Disc progenitor cells were defined as stem cells found within the intervertebral disc that can differentiate into various disc cell types and contribute to disc regeneration. These cells express markers characteristic of MSCs, suggesting overlapping biological properties.
The efficacy outcomes in this review included the VAS score, ODI, and the number of patients demonstrating improvement on both measures. VAS is a tool for quantifying subjective pain intensity on a linear scale ranging from 0 to 100, with higher scores indicating greater pain severity. ODI is a self-administered questionnaire that assesses functional disability associated with LBP. It yields a percentage score ranging from 0% to 100%, with higher values reflecting greater disability related to LBP.
The safety outcomes in this review included TEAEs, SAEs, treatment discontinuation due to AEs, and mortality. TEAEs were defined as any AEs that newly appeared or worsened following initiation of treatment. SAEs were defined as any AEs that resulted in death, were life-threatening, required hospitalization or prolongation of existing hospitalization, or caused persistent or substantial disability or incapacity, regardless of their relationship to the study treatment. Treatment discontinuation due to AEs was defined as the cessation of therapy or intervention as a direct result of an AE. Mortality was defined as the number of deaths that occurred during the follow-up period.
Risk of bias assessment
Two reviewers independently assessed the risk of bias for each included study. For RCTs, the Cochrane Risk of Bias version 2 (RoB 2) tool (Cochrane, London, UK) was applied [11]. This instrument evaluates five domains of potential bias and categorizes each as “low risk,” “some concerns,” or “high risk” of bias [11].
Statistical analysis
Dichotomous outcomes were analyzed using odds ratios (ORs) with 95% confidence intervals (95% CI), using the Mantel-Haenszel method. Continuous outcomes were analyzed as mean differences (MDs) with 95% CIs. A random-effects model was applied owing to the anticipated heterogeneity related to differences in MSC dosage, preparation, and patient demographics. Heterogeneity was assessed using the I2 statistic [12]. Values of ≤25%, 26%–50%, and >50% were interpreted as low, moderate, and high heterogeneity, respectively [12]. For studies reporting medians with interquartile ranges or (ranges), data were converted to mean±standard deviation using the method described by Wan et al. [13] For efficacy outcomes (VAS score and ODI), subgroup analyses were conducted according to MSC dosage (low ≤6×106 cells/mL vs. high >6×106 cells/mL), MSC origin (autologous vs. allogenic), and the type of preparation (crystalloid solution vs. hyaluronic acid [HA]). For safety outcomes with significant heterogeneity, sensitivity analyses were conducted by excluding studies with differing follow-up durations. Publication bias was assessed when ≥10 studies were available for a given outcome. All statistical analyses were performed using Review Manager (RevMan) ver. 5.4 (Cochrane).
Ethics approval
This is a systematic review and meta-analysis study. The Faculty of Medicine, Universitas Indonesia Research Ethics Committee has confirmed that no ethical approval is required.
Results
Study selection and characteristics
Database searches identified 44 records: Europe PMC (n=192), Scopus (n=147), Medline (n=94), and Cochrane Library (n=9). After removing duplicates and screening titles/abstracts, 384 records were excluded. Of the 58 full-text articles examined, 51 were excluded for the following reasons: review articles (n=22), no control group (n=14), non-human studies (n=7), surgical patient population (n=3), not using MSCs (n=2), protocol-only publications (n=2), and non-randomized design (n=1). Ultimately, seven RCTs enrolling 594 patients with CLBP attributable to DDD were included in the final analysis [14–20] (Fig. 1). Two RCTs employed a double-blind design, while the remaining five did not specify blinding and were therefore categorized solely as prospective RCTs. Four studies were conducted in the United States, two in Spain, and two were multinational (France, Italy, Germany, and Spain). Sample sizes in the MSCs group ranged from 12 to 143 participants, while control groups ranged from 10 to 132. Follow-up durations varied from 3 to 42 months. Substantial variation was observed in MSC dosing. For example, Amirdelfan et al. [14] reported outcomes separately for patients receiving 6 million versus 18 million MSCs, while Gornet et al. [16] divided patients into low-dose (3,000,000 cells/mL allogeneic injectable disc progenitor cell therapy [IDCT]) and high-dose (9,000,000 cells/mL allogeneic IDCT) groups. Five RCTs used placebo (normal saline) as control, while two used sham injections (mepivacaine 1%) as the comparator. Table 1 provides an overview of the baseline characteristics for each study included in this analysis. Supplement 2 provides further details regarding the eligibility criteria of the included RCTs.
PRISMA (Preferred Reporting Items for Systematic reviews and Meta-Analyses) diagram of the detailed process of selection of studies for inclusion in the systematic review and meta-analysis.
Baseline characteristics of the included studies
Quality assessment of included studies
Using the RoB 2 tool (Cochrane), four out of the seven RCTs were judged to have a “low” risk of bias across all five domains (Fig. 2). The remaining three RCTs were rated as having “some concerns” in the “deviations of intended interventions” domain due to the lack of information regarding blinding of participants and intervention providers, coupled with the lack of data on protocol deviations. These RCTs also displayed “some concern” in the “measurement of the outcome” domain due to a lack of information regarding the blinding of outcome assessors, which could have potentially influenced outcome evaluation (Fig. 2).
Risk of bias assessment using RoB ver. 2 tool from Cochrane Collaborations.
Efficacy parameters
Visual Analog Scale
Meta-analysis of seven RCTs showed that patients with DDD receiving intradiscal MSC injections had significantly lower VAS scores compared with those receiving sham/placebo (MD, −6.67; 95% CI, −9.31 to −4.02; p<0.00001, I2=29%, random-effect model) (Fig. 3, Table 2).
Forest plot that demonstrates the comparison between intradiscal injection of mesenchymal stem cells (MSCs) and sham/placebo for patients with degenerative disc disease (DDD) in terms of reduction of Visual Analog Scale (VAS) from baseline. SD, standard deviation; IV, inverse variance; CI, confidence interval; df, degree of freedom.
Summary of meta-analysis results for intradiscal injection of MSCs compared to sham/placebo for patients with CLBP due to DDD
When stratified by follow-up time, the benefit of MSCs became evident only in the later follow-up period, especially at 12 months (MD, −9.82; 95% CI, −15.67 to −3.97; p=0.001, I2=18%, random-effect model), 24 months (MD, −11.85; 95% CI, −19.42 to −4.27; p=0.002, I2=47%, random-effect model), and 36 months (MD, −17.04; 95% CI, −30.35 to −3.72; p=0.01, I2=70%, random-effect model) (Fig. 3, Table 2). The magnitude of pain reduction appeared to increase with longer follow-up. In contrast, no significant differences were observed during the early period of follow-up: 1 month (MD, −0.74; 95% CI, −6.06 to 4.57; p=0.78, I2=0%, random-effect model), 3 months (MD, −3.74; −10.08 to 2.60; p=0.25, I2=24%, random-effect model), and 6 months (MD, −3.49; 95% CI, −8.02 to 1.04; p=0.13, I2=0%, random-effect model) (Fig. 3, Table 2).
A subgroup analysis based on MSC dosage showed a greater reduction in VAS scores among studies using higher doses of MSCs (MD, −8.99; 95% CI, −12.77 to −5.22; p<0.00001, I2=16%) compared to those using lower doses (MD, −4.61; 95% CI, −8.09 to −1.12; p=0.01, I2=34%). When examined by follow-up interval, low-dose MSC administration did not yield statistically significant reductions in VAS scores at 1 month (p=0.77), 3 months (p=0.74), 6 months (p=0.20), 12 months (p=0.15), 24 months (p=0.10), or 36 months (p=0.16). In contrast, high-dose MSCs demonstrated significant reductions beginning at 12 months (MD, −13.56; p=0.001), with effects sustained at 24 months (MD, −13.02; p=0.02) and 36 months (MD, −27.72; p<0.0001) (Supplement 3).
Subgroup analysis by MSC source showed that only allogenic MSCs were associated with significant reductions in VAS scores (MD, −7.38; 95% CI, −9.85 to −4.91; p<0.00001, I2=15%), whereas autologous MSCs were not (MD, 5.21; 95% CI, −1.56 to 11.98; p=0.13, I2=0%) (Supplement 3). Specifically, autologous MSCs failed to show meaningful effects at either 1 month (p=0.52) or 3 months (p=0.14), while allogenic MSCs showed significant improvement by 3 months (MD, −7.12; 95% CI, −13.47 to −0.77; p=0.03, I2=0%) (Supplement 3). Notably, the evidence for autologous MSCs was limited to a single, small-scale RCT with only 3 months of follow-up, highlighting the need for larger trials with extended follow-up durations.
Subgroup analysis by MSC preparation showed significant effects for both MSCs combined with crystalloid solution (MD, −3.37; 95% CI, −5.73 to −1.01; p=0.005, I2=0%) and those combined with HA (MD, −13.39; 95% CI, −17.80 to −8.99; p<0.00001, I2=6%), with the effect size being more pronounced in the HA group (Supplement 3). When stratified by follow-up duration, studies using crystalloid mixtures demonstrated significant improvement only at 12 months (MD, −5.95; 95% CI, −11.31 to −0.59; p=0.03, I2=0%), with no meaningful differences observed at 1 month (p=0.70), 3 months (p=0.58), 6 months (p=0.31), 24 months (p=0.06), or 36 months (p=0.26) (Supplement 3). In contrast, MSCs administered with HA showed significant and sustained reductions in VAS scores beginning at 12 months (MD, −20.11; 95% CI, −30.87 to −9.36; p=0.0002, I2=0%) and persisting through 24 months (MD, −21.73; 95% CI, −32.65 to −10.81; p<0.0001, I2=0%) and 36 months (MD, −20.87; 95% CI, −31.95 to −9.80; p=0.0002, I2=0%) (Supplement 3).
Proportion of patients achieving ≥30% pain reduction
Meta-analysis of 3 RCTs demonstrated that intradiscal MSC injection was associated with a significantly higher proportion of patients achieving ≥30% pain reduction (VAS) compared with sham/placebo (OR, 1.82; 95% CI, 1.43–2.33; p<0.00001, I2=21%, random-effect model) (Table 2).
When stratified by follow-up time, significant differences were observed starting from 6 months (OR, 1.67; 95% CI, 1.08–2.57; p=0.02, I2=0%, random-effect model), with benefits persisting at 12 months (OR, 2.17; 95% CI, 1.40–3.34; p=0.0005, I2=0%, random-effect model), 24 months (OR, 2.34; 95% CI, 1.01–5.39; p=0.05, I2=65%, random-effect model), and 36 months (OR, 2.80; 95% CI, 1.12–6.99; p=0.03, I2=54%, random-effect model) (Table 2). In contrast, no significant differences were detected during early follow-up at 1 month (OR, 1.32; 95% CI, 0.60–2.91; p=0.50, random-effect model) or 3 months (OR, 0.95; 95% CI, 0.44–2.03; p=0.89, random-effect model) (Table 2).
Proportion of patients achieving ≥50% pain reduction
Meta-analysis of two RCTs showed that intradiscal MSC injection was associated with a significantly higher proportion of patients achieving ≥50% pain reduction (VAS) compared with sham/placebo (OR, 1.96; 95% CI, 1.38–2.79; p=0.0002, I2=33%, random-effect model) (Table 2).
When analyzed by follow-up interval, significant differences were observed at 6 months (OR, 1.73; 95% CI, 1.02–2.92; p=0.04, I2=0%, random-effect model) and 36 months (OR, 2.08; 95% CI, 1.04–4.15; p=0.04, I2=25%, random-effect model) (Table 2). No significant differences were found at 12 months (OR, 2.80; 95% CI, 0.87–8.98; p=0.08, I2=71%, random-effect model) or 24 months (OR, 2.26; 95% CI, 0.69–7.45; p=0.18, I2=68%, random-effect model) (Table 2). Data for 1- and 3-month follow-up were unavailable in the included studies.
Oswestry Disability Index
Meta-analysis of seven RCTs demonstrated a significantly greater reduction in ODI scores from baseline among DDD patients receiving intradiscal MSC injections compared with sham/placebo (MD, −4.05%; 95% CI, −5.24 to −2.87; p<0.00001, I2=32%, random-effect model) (Fig. 4, Table 2).
Forest plot that demonstrates the comparison between intradiscal injection of mesenchymal stem cells (MSCs) and sham/placebo for patients with degenerative disc disease (DDD) in terms of reduction of Oswestry Disability Index (ODI) from baseline. SD, standard deviation; IV, inverse variance; CI, confidence interval; df, degree of freedom.
When stratified by follow-up time, significant improvements were evident as early as 3 months (MD, −3.16%; 95% CI, −5.41 to −0.91; p=0.006, I2=0%, random-effect model), and persisted at 6 months (MD, −4.68%; 95% CI, −7.88 to −1.48; p=0.004, I2=0%, random-effect model), 12 months (MD, −5.05%; 95% CI, −8.39 to −1.71; p=0.003, I2=10%, random-effect model), 24 months (MD, −6.73%; 95% CI, −10.29 to −3.18; p=0.0002, I2=38%, random-effect model), and 36 months (MD, −11.58%; 95% CI, −16.08 to −7.08; p<0.00001, I2=68%, random-effect model) (Fig. 4, Table 2). The magnitude of ODI reduction increased with longer follow-up durations. No significant difference between groups was observed at the 1-month follow-up (MD, −1.33%; 95% CI, −3.54 to 0.89; p=0.24, I2=0%, random-effect model) (Fig. 4, Table 2).
Subgroup analysis by MSC dosage showed greater ODI improvement in studies using high-dose MSCs (MD, −7.56%; 95% CI, −10.76 to −4.35; p<0.00001, I2=48%) compared with low-dose MSCs (MD, −3.20%; 95% CI, −4.61 to −1.79; p<0.00001, I2=0%). Dose-related subgroup findings by follow-up periods are presented in Supplement 3.
When stratified by MSC source, significant effects were observed only in studies using allogenic MSCs (MD, −5.32%; 95% CI, −7.04 to −3.59; p<0.00001, I2=28%), while autologous MSCs showed no significant benefit (MD, −2.06%; 95% CI, −4.86 to 0.73; p=0.15, I2=35%). Notably, the autologous MSC data were derived from a single small RCT with only 3 months of follow-up, underscoring the need for larger studies with longer observation periods. Detailed results by follow-up period are presented in Supplement 3.
Subgroup analysis by MSC preparation type showed that MSCs delivered with HA produced greater ODI improvement (MD, −7.60%; 95% CI, −10.38 to −4.81; p<0.00001, I2=16%) than MSCs delivered with crystalloid solution (MD, −3.54%; 95% CI, −5.24 to −1.85; p<0.0001, I2=23%). Further subgroup findings by follow-up duration are also provided in Supplement 3.
Proportion of patients achieving ≥10-point ODI reduction
Meta-analysis of two RCTs demonstrated that intradiscal MSC injection was associated with a significantly higher proportion of patients achieving ≥10-point ODI reduction compared with sham/placebo (OR, 2.13; 95% CI, 1.61–2.82; p<0.00001, I2=6%, random-effect model) (Table 2).
When stratified by follow-up time, significant differences were observed beginning at 6 months (OR, 2.07; 95% CI, 1.18–3.64; p=0.01, I2=2%, random-effect model), 12 months (OR, 2.73; 95% CI, 1.36–5.47; p=0.005, I2=26%, random-effect model), 24 months (OR, 2.12; 95% CI, 1.07–4.32; p=0.03, I2=26%, random-effect model), and 36 months (OR, 4.57; 95% CI, 1.81–11.54; p=0.001, I2=0%, random-effect model) (Table 2). No significant differences were found during early follow-up at 1 month (OR, 1.33; 95% CI, 0.53–3.34; p=0.55, random-effect model) or 3 months (OR, 1.39; 95% CI, 0.64–2.99; p=0.40, random-effect model) (Table 2).
Proportion of patients achieving ≥15 point ODI reduction
Data for this outcome were available from a single RCT, which reported results separately for patients receiving 6 million MSCs and 18 million MSCs. Pooled analysis of these two groups showed that intradiscal MSC injection was associated with a significantly higher proportion of patients achieving ≥15-point ODI reduction compared with sham/placebo (OR, 4.45; 95% CI, 2.76–7.15; p<0.00001, I2=0%, random-effect model) (Table 2).
Significant differences favoring MSCs were observed at 6 months (OR, 3.93; 95% CI, 1.63–9.49; p=0.002, I2=0%, random-effect model), 12 months (OR, 4.00; 95% CI, 1.59–10.09; p=0.003, I2=0%, random-effect model), 24 months (OR, 4.96; 95% CI, 1.81–13.55; p=0.002, I2=0%, random-effect model), and 36 months (OR, 5.30; 95% CI, 1.94–14.48; p=0.001, I2=0%, random-effect model) (Table 2). No data were available for the 1-month and 3-month follow-up.
Safety parameters
Treatment-emergent adverse events
Meta-analysis of two RCTs showed no significant difference in the incidence of TEAEs between DDD patients receiving intradiscal MSC injections and those receiving sham/placebo (OR, 1.07; 95% CI, 0.64–1.77; p=0.80, I2=0%, random-effect model) (Table 2).
Serious adverse events
Meta-analysis of three RCTs showed no significant difference in the incidence of SAEs between DDD patients receiving intradiscal MSC injections and those receiving sham/placebo (OR, 0.58; 95% CI, 0.18–1.91; p=0.37, I2=56%, random-effect model) (Table 2).
Given the notable heterogeneity detected for this outcome, a sensitivity analysis was performed by excluding the RCT by Gornet et al. [16], which had a shorter follow-up duration (24 months vs. 36 months in the other two studies). This reduced heterogeneity from 56% to 0%, but the results remained nonsignificant (OR, 1.33; 95% CI, 0.68–2.60; p=0.40, I2=0%, random-effect model).
Treatment discontinuation due to AEs
Meta-analysis of two RCTs revealed that treatment discontinuation due to AEs did not differ significantly between DDD patients receiving intradiscal MSC injection and those receiving sham/placebo (OR, 1.42; 95% CI, 0.28–7.19; p=0.67, I2=0%, random-effect model) (Table 2).
Mortality
Three RCTs reported data on mortality. No deaths were recorded in either the MSC or sham/placebo groups; therefore, a meta-analysis could not be performed.
Discussion
This systematic review and meta-analysis demonstrated that intradiscal MSC injection is associated with significant reductions in pain and improvements in function among patients with CLBP due to DDD compared with sham/placebo. Pain relief was reflected in lower VAS scores and a higher proportion of patients achieving ≥30% and ≥50% pain reduction, with benefits becoming more apparent at longer follow-up intervals (12–36 months). Functional improvement, measured by ODI, was evident earlier (beginning at 3 months) and persisted through 36 months. Subgroup analyses further suggested that clinical benefit was greater with high-dose MSCs (>6×106 cells/mL) compared with low-dose, allogeneic sources compared with autologous, and HA suspensions compared with crystalloid solutions. These findings point toward potential dose–response and preparation-related effects. Despite statistical significance, the magnitude of change may not reach thresholds considered clinically meaningful. The mean reduction in ODI (−4.05 points) and VAS (−6.67 points) falls short of established minimal clinically important differences, estimated at approximately 10–15 points for ODI and 15–20 points on a 100-mm VAS. This suggests that while the data may show measurable changes, the patients may not perceive the improvements as substantial in daily life. Therefore, the findings, although statistically robust, may fall short of translating into clinically relevant benefits, underscoring the need for cautious interpretation in real-world application. In terms of safety, MSC injections appeared well-tolerated. No significant differences were observed in TEAEs, SAEs, treatment discontinuation due to AEs, or mortality between MSC and sham/placebo groups, suggesting a favorable safety profile in the short- to mid-term.
The presence of “some concerns” in three of the seven RCTs, primarily due to insufficient reporting of blinding and deviations from the intended protocol, raises potential threats to the validity of the findings. Given that the primary outcomes (VAS and ODI) are subjective, they are highly susceptible to bias arising from inadequate blinding. Participants’ and assessors’ awareness of the intervention may influence perceptions of pain and disability, leading to inflated treatment effects. In addition, protocol deviations, such as inconsistent adherence or unbalanced co-interventions, could confound outcome assessments and compromise internal validity. Collectively, these limitations may have contributed to an overestimation of efficacy in the meta-analysis, reducing confidence in the reliability of the synthesized evidence.
Disc degeneration begins with a decline in the number of large notochordal cells within the nucleus pulposus, which alters intervertebral disc cellular function [8,23]. This triggers a degenerative cascade characterized by matrix loss, upregulation of matrix metalloproteinases, a transition from type II to type I collagen, and reduced proteoglycan content [8,23]. These morphological changes impair water-binding capacity, leading to disc dehydration and subsidence [8,23]. The resulting decline in intradiscal pressure compromises the mechanical integrity of the spine, contributing to local instability and microtrauma [8,23]. Degeneration is also accompanied by increased expression of pro-inflammatory and catabolic cytokines, along with various inflammatory mediators (e.g., interleukin-1 [IL-1], IL-6, IL-12, IL-17, tumor necrosis factor-alpha [TNF-α], interferon-beta [IFN-β], and IFN-γ) [23,24]. These pro-inflammatory cytokines drive cellular remodeling, fibrosis, and the ingrowth of vascularized granulation tissue. Importantly, they also activate nociceptive nerve endings within the disc, promoting aberrant innervation and the development of discogenic pain [23,24].
Consistent with the aforementioned pathophysiology, pain relief and functional enhancement from MSC therapy in DDD may occur through several mechanisms. First, MSCs can differentiate into nucleus pulposus cells (NPCs) to replace degenerated cells and support surviving NPCs by producing cytokines such as transforming growth factor-β1 [25,26]. Second, they can promote synthesis of proteoglycans and type II collagen, both of which are reduced in DDD, thereby helping to restore extracellular matrix (ECM) integrity and slow disease progression [27]. Third, MSCs exhibit strong immunomodulatory effects, secreting anabolic growth factors that stimulate adjacent intervertebral disc cells, enhance ECM production, and suppress inflammatory and catabolic pathways in the disc microenvironment [27,28]. It should be noted that these mechanisms remain largely hypothetical and have not yet been definitively confirmed in clinical studies.
The findings of our systematic review and meta-analysis are consistent with a prior meta-analysis by Xie et al. [29], who also reported greater reductions in both VAS and ODI scores among patients with DDD treated with intradiscal MSC injections compared with controls. However, several important distinctions exist between the two analyses. The meta-analysis by Xie et al. [29] included only three RCTs with a total of 128 participants, providing relatively weak evidence. In contrast, our meta-analysis incorporated seven RCTs with 594 participants, thereby offering a substantially larger evidence base. In addition, Xie et al. [29] presented outcomes only at 3, 6, 12, and ≥24 months of follow-up. Our analysis expands on this by including data at 1 month and 36 months, showing that the efficacy of MSC therapy increases with longer follow-up duration. This broader evidence base provides a more comprehensive comparison between MSCs and sham/placebo.
This study has several limitations. Although seven RCTs were included, most had modest sample sizes, with only one RCT enrolling more than 100 participants per group. Furthermore, the majority of RCTs were conducted in the USA or Spain, with limited representation from Asian populations, which restricts the generalizability of the findings. Publication bias could not be formally assessed, as the number of eligible studies was below the recommended threshold of 10 for reliable evaluation using funnel plots or statistical tests. Therefore, the possibility of unpublished or selectively reported studies influencing the results cannot be entirely excluded. Finally, several RCTs were rated as having “some concerns” in risk of bias, particularly in the domains of deviations from intended interventions and outcome measurement, which warrants cautious interpretation of the pooled findings. Future large-scale, multicenter RCTs, especially from Asian countries, are needed to validate and extend the evidence presented in this review.
Conclusions
This systematic review and meta-analysis demonstrated that intradiscal MSC injections reduce pain and improve function in patients with CLBP attributable to DDD. The therapy appears safe, with no significant differences in AEs compared with sham/placebo. These findings suggest that MSC therapy may represent a potential treatment option for patients with CLBP unresponsive to conservative management.
Key Points
This meta-analysis included seven randomized controlled trials evaluating intradiscal mesenchymal stem cell (MSC) injections for chronic low-back pain due to degenerative disc disease.
MSC therapy significantly improved pain (Visual Analog Scale) and disability (Oswestry Disability Index) compared with sham/placebo.
No significant differences were observed in adverse events or mortality.
Intradiscal MSC injection appears to be a safe and effective option for patients unresponsive to conservative treatments.
Notes
Conflict of Interest
No potential conflict of interest relevant to this article was reported.
Author Contributions
Conceptualization: AJR, SWM, RT, FAKC. Methodology: RT, FAKC. Formal analysis and investigation: AJR, SWM, RT, FAKC. Writing–original draft preparation: RT, FAKC. Writing–review and editing: AJR, SWM. Resources: AJR, SWM, RT, FAKC. Supervision: AJR, SWM. Final approval of the manuscript: all authors
Supplementary Materials
Supplementary materials can be available from https://doi.org/10.31616/asj.2025.0354.
Supplement 1. Literature search strategy.
asj-2025-0354-Supplement-1.pdfSupplement 2. Eligibility criteria for participants among the included studies.
asj-2025-0354-Supplement-2.pdfSupplement 3. Summary of subgroup analysis results for intradiscal injection of MSCs compared to sham/placebo for patients with CLBP due to DDD.
asj-2025-0354-Supplement-3.pdf