Table 4.

Comparison of the most recent and pivotal clinical studies (ordered by study type/phase of trial and chronological date order)

References	Cell donor	Sample size/control	OA severity	BM-MSC dosage	Follow-up	Significant findings
Case reports
Centeno et al. [35]	Autologous	n = 1 Male 56 years No control	OA causing significant on-going pain and disability (ungraded)	1 Injection: 22.4 × 106 cells suspended in PBS Dexamethasone injection administered following BM-MSC injection	Baseline, 1 and 3 months	No adverse events reported MRI: ↑ cartilage and meniscus growth
Emadedin et al. [38]	Autologous	n = 6 Female Mean age: 54 years No control	Grade 4	1 Injection: 20–24 × 106 cells suspended at a density of 5 × 106	Baseline, 6 and 12 months	No local or systemic adverse events WOMAC: ↓ at 6 and 12 months Mean walking distance: ↑ at 6 months No local or systemic adverse events Mean walking distance: ↑ at 6 and 30 months after treatment WOMAC: ↓ at 6, 12 and 30 months compared with baseline. ↓ in WOMAC physical function sub scores at 6 and 12 months
Emadedin et al. [39] Long term-follow up	Autologous	n = 6 Female Mean age: 54 years No control	Grade 4	1 Injection: 20–24 × 106 cells suspended at a density of 5 × 106	Baseline, 6, 12 and 30 months
Mehrabani et al. [67]	Autologous	n = 1 Female 47 years No control	Grade 4 Unresponsive to NSAIDs	1 Injection: 36 × 106 cells provided and transferred in 2 ml of media	3, 6 and 12 months	No local or systemic adverse events MRI: ↑ thickness of cartilage on distal condyle of femur and proximal tibia at 6 and 12 months
Preliminary reports
Davatchi et al. [36]	Autologous	n = 4 2 Females: 57 and 54 years 2 Males: 55 and 65 years	Grade 2–3	1 Injection: 8–9 × 106 in a mean volume of 5.5 mL	Baseline, 1 week, then every month up to 1 year	No adverse events reported
Davatchi et al. [37] 5-year follow-up	Autologous	n = 3 No control	Grade 2–3	1 Injection: 8–9 × 106 in a mean volume of 5.5 mL	5 years	No adverse events reported
Pilot studies
Orozco et al. [40]	Autologous (according to Good Manufacturing Processes: GMP)	n = 12 6 Females 6 Males Mean age: 49 ± 5 No control	Grade 2 to 4 Unresponsive to conservative treatment for 6 months	1 Injection: 40 × 106 cells suspended in ringer-lactate at 5 × 106	Baseline, 3, 6 and 12 months	Mild adverse events: the first 1–6 days and occurred frequently (50% of patients) = controlled by ibuprofen VAS: Pain ↓ at 3 months with progressive improvement during the subsequent 9 months (statistically significant at all time points compared with basal pain level) WOMAC: All subscales ↓ at 12 months compared to baseline Lequesne algofunctional index: Correlation between improvement and the initial score (p < 0.01) MRI: Mean PCI ↓ from 19.5 to 15.4 during first 6 months and further ↓ to 14.3 at 12 months (11 out of 12 patients) Correlation between VAS and PCI
Orozco et al. [40] Two-year follow up	Autologous (GMP)	n = 12 6 Females 6 Males Mean age: 49 ± 5 No control	Grade 2 to 4 Unresponsive to conservative treatment for 6 months	1 Injection: 40 × 106 cells suspended in ringer-lactate at 5 × 106	2 years	Results of the follow-up reaffirm the conclusions for the first-year results regarding feasibility and safety
Phase I/II studies
Rich et al. [41] Clinical trial-Phase I/II	Autologous (ex-vivo)	n = 50 20 Females 30 Males Mean age: 57.8 ± 14.1 No control	Grade 2–4	1 Injection: 40 × 106 suspended in ringer-lactate at 5 × 106	Day 8, 3, 6 and 12 months	No local or systemic adverse events VAS: ↓ at 6 and 12 months The pattern of 1-year improvement was parallel for VAS, WOMAC and Lequesne algofunctional index MRI: Mean PCI ↓ from 25 to 5 at 12 months post-injection
Soler et al. [42] Prospective, open-label, single-dose, single-arm clinical trial-Phase I/II final results	Autologous (ex-vivo)-Infusion of XCEL-M-Alpha	n = 15 9 Females 6 Males Mean age: 52 No control	Grade 2 (n = 9) Grade 3 (n = 6)	1 Injection: 40 × 106 ± 10 × 106 XCEL-M-ALPHA was infused within 6 h from delivery	Day 8, 3, 6- and 12-months	Mild adverse events (local discomfort and back pain from bone marrow extraction) VAS: Pain ↓ and daily activity ↑ at 8 days until the end of the study period HAQ (questionnaire): Over time ↓ from baseline across whole population at 12 months WOMAC: ↓ at 12 months Lequesne algofunctional index: ↓ at 6 and 12 months
Al Najar et al. [34] Prospective open-label study-Phase I/II	Autologous	n = 13 7 Females 6 Males Mean age: 50 years No control	Grade 2–4	2 Injections (1 month apart): 30.5 × 106 cells suspended in 09% normal saline	Adverse events: day 1, 7, 14, 28, 60 and then every 6 months until month 24 Normalised KOOS: baseline, 1, 2, 4, 6, 12 and 24 months after first injection MRI: baseline, 6 and 12 months	2 local adverse events within 2 h of injection, 1 6 h after injection (all resolved with ice/mild analgesia in 48 h) Normalised KOOS: Symptoms and pain ↓, daily life activity, sport and QoL ↑ at 6, 12 and 24 months MRI: ↑ Mean tibial and femoral plate thickness (mm) at 12 months (1 female deteriorated by MRI despite of KOOS improvement)
Randomised clinical trails (RCT)
Vangsness et al. [43] Double-blinded, randomised, controlled clinical study	Allogenic (GMP) Obtained from donors (screened and tested according to the US FDA)	n = 55 63% were male Group A = 18 Group B = 18 Control = 19 Mean age: 46 years	All underwent subtotal meniscectomies-any previous knee ligament reconstruction needed to have had a stable result	Group A: 50 × 106 cells suspended in 2 mL of HA, human serum albumin and PlasmaLyte A to a volume of 5 ml Group B: 150 × 106 suspended in 2 mL of HA, human serum albumin and PlasmaLyte A to a volume of 5 ml Control: A vehicle control compromised the same HA solution without BM-MSCs	Baseline, 6 weeks, 6 months, 1 year and 2 years post-operatively	427 adverse events among 55 patients. 272 were mild, 28 sever and 1 life-threatening 1-year post VAS: Knee pain ↓ for all patient’s compared with baselines in all groups. Significant differences were observed at 2 years for group A, and at 1 and 2 years for B Lysholm knee scale: Total scores relative to baseline ↓ at all follow-ups Meniscus volume > 15%: At 12 months, both the control compared with group A and overall comparison were significant (> 15%) and at 2 years the overall group comparison was significant
Vega et al. [44] Randomised controlled, comparator multi-centre-Phase I/II study	Allogenic (GMP)	n = 30 17 Females 13 Males Mean age: 57 ± 9 Experimental group = 15 Control = 15	Grade 2–4 Unresponsive to conventional treatments for at least 6 months prior to recruitment	1 Injection: Experimental group: 40 × 106 cells from a 5 × 106 cell/mL suspension Control: 60 mg of HA in 3 mL	Baseline, day 8, 3, 6 and 12 months	Minor adverse events during first 7 days in both groups = 53–60% of patients AS: ↓ in experimental group at 6 and 12 months. Control group ↓ at 12 months WOMAC: Pain and general WOMAC ↓ at 6 and 12 months for experimental group Lequesne algofunctional index: ↓ at 6 and 12 months in experimental group MRI: Poor Cartilage Index (PCI) ↓ at 12 months in experimental group
Vangsness et al. [43] Double-blinded, randomised, controlled clinical study	Allogenic (GMP) (screened and tested according to the US FDA)	n = 55 63% were male Group A = 18 Group B = 18 Control = 19 Mean age: 46 years	All underwent subtotal meniscectomies-any previous knee ligament reconstruction needed to have had a stable result	Group A: 50 × 106 cells suspended in 2 mL of HA, human serum albumin and PlasmaLyte A to a volume of 5 ml Group B: 150 × 106 suspended in 2 mL of HA, human serum albumin and PlasmaLyte A to a volume of 5 ml Control: A vehicle control compromised the same HA solution without BM-MSCs	Baseline, 6 weeks, 6 months, 1 year and 2 years postoperatively	427 adverse events among 55 patients. 272 were mild, 28 sever and 1 life-threatening 1-year post VAS: Knee pain ↓ for all patient’s compared with baselines in all groups. Significant differences were observed at 2 years for group A, and at 1 and 2 years for B Lysholm knee scale: Total scores relative to baseline ↓ at all follow-ups Meniscus volume > 15%: At 12 months, both the control compared with group A and overall comparison were significant (> 15%) and at 2 years the overall group comparison was significant
Vega et al. [44] Randomised controlled, comparator multi-centre- Phase I/II study	Allogenic (GMP) Obtained from three healthy donors	n = 30 17 Females 13 Males Mean age: 57 ± 9 Experimental group = 15 Control = 15	Grade 2–4 Unresponsive to conventional treatments for at least 6 months prior to recruitment	1 Injection: Experimental group: 40 × 106 cells from a 5 × 106 cell/mL suspension Control: 60 mg of HA in 3 mL	Baseline, day 8, 3, 6 and 12 months	Minor adverse events during first 7 days in both groups = 53–60% of patients VAS: ↓ in experimental group at 6 and 12 months. Control group ↓ at 12 months WOMAC: Pain and general WOMAC ↓ at 6 and 12 months for experimental group Lequesne algofunctional index: ↓ at 6 and 12 months in experimental group MRI: Poor Cartilage Index (PCI) ↓ at 12 months in experimental group
Gupta et al. [30] Randomised, double-blind multicentre placebo-controlled phase II study	Allogenic (GMP ex-vivo) Stempeucel®	n = 60 15 in each dose group (4 groups) Randomised into two groups within the 4 main groups (2:1) for Stempeucel® and placebo (control)	Grade 2–3	1 Injection: Group 1: 25 × 106 of Stempeucel® Group 2: 50 × 106 of Stempeucel® Group 3: 75 × 106 of Stempeucel® Group 4: 150 × 106 of Stempeucel® Each group control: PlasmaLyte placebo	Baseline, 1 week, 1, 3 and 6 months Clinical data unblinded after 6 months but followed-up until 12 months	Adverse events: 97 mild to moderate adverse events were reported in 40 subjects
Espinosa et al. [49] Randomised clinical trial Phase I/II	Autologous (GMP) co-administered with HA	n = 30 Control group:10 Low dose:10 High dose:10 Active control	Inclusion of ≥ 2 Range: 2–4	1 Injection: Control: 60 mg HA (in a volume of 4 ml) Low BM-MSC dose: 10 × 106 cells in 1.5 ml ringer’s lactate solution + 4 ml of HA injection High BM-MSCs dose: 100 × 106 cells in 3 ml ringer’s lactate solution + 4 ml of HA injection	Baseline, 3, 6 and 12 months	Articular pain requiring anti-inflammatory treatment at 24 h post-injection in 1, 3 and 6 patients in the control, low-dose and high-dose groups respectively Range of motion: ↑ in BM-MSCs treated groups (effect seen earlier in higher dose) VAS score ↓ in low and high dosage groups at all follow-up times WOMAC: Control: Pain ↓ at 3 and 6 months, and function ↑ at 3 and 6 months Low dose: Stiffness ↓ at 6 and 12 months High dose: All WOMAC sub-scores improved significantly at 12 months X-ray: Knee joint space ↓ in control group at 12 months