This research aimed to study the safety and efficacy of adipose-derived mesenchymal stem cells (ADMSCs) for knee osteoarthritis (OA).

We used six databases to search for records and then screened them for eligibility. In both randomized and non-randomized studies, the risk of bias was assessed. The data were then retrieved and used in single-arm and double-arm analyses using Comprehensive Meta-Analysis (CMA) Version 3.0 and RevMan Version 5.3, respectively.

Based on the study's inclusion criteria, we included 15 studies with a total of 463 patients. According to our single-arm analyses, there was a significant improvement in quality of life (QOL) among the three dose subgroups (high, medium, and low doses), as measured by SF-36 scores after a year of follow-up [low dose: mean (M)=−23.99; 95% confidence interval (CI) [−31.49 to −16.49]; P<0.001; medium dose: M=−15.96; 95% CI [−23.5 to−8.42]; P<0.001; high dose: M=−19.31; 95% CI [−28.02 to −10.59], P<0.001] and the knee injury and osteoarthritis outcome score (KOOS) QOL sub-score after six months following ADMSCs injection in the low-dose group (M=24.9; 95% CI [4.3 to 45.6]; P<0.05). Moreover, after three months of follow-up, we detected significant pain reduction as measured by the numeric pain rating scale (NPRS), with no significant difference between the low and medium doses (low dose: M=−3.12; 95% CI [−5.09 to −1.14]; P<0.01; medium dose; M=−2.17; 95% CI [−3.13 to −1.21]; P<0.001). However, after a year, the results were no longer significant. Despite finding no significant difference between them after 6 and 12 months post-treatment in the Visual Analogue scale (VAS) scale and the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) total score, double-arm analyses revealed significant pain reduction in the ADMSCs group over the control after 12 months as estimated by the WOMAC pain sub-score (mean difference (MD)=−1.85, 95% CI [−3.55, −0.15], P<0.05). After six months, the low dosage group's knee functions and activity levels improved significantly, as determined by the WOMAC physical function and stiffness subscales (M=−23.79; 95% CI [−38.43 to −9.16]; P=0.001; M=−10.25; 95% CI [−17.31 to −2.59]; P<0.01, respectively), as well as the KOOS scores after a year (P<0.01 for all KOOS subscales). In the ADMSCs injections group, there were no serious adverse effects [event rate (ER)=0.11; 95% CI [0.03–0.3]; P=0.001].

In the present single-arm meta-analysis, ADMSCs were associated with significant reduction in pain and improvement in QOL and knee functions in patients with knee OA. However, double arm analyses did not confirm these positive findings, which may be returned to the small sample size of included patients. Therefore, to introduce ADMSCs into clinical practice and establish guidelines for their use, more randomized controlled clinical trials with large sample sizes and long-term follow-ups are needed.