Clinical application of concentrated bone marrow aspirate in orthopaedics: A systematic review

Table 1 Studies evaluating concentrated bone marrow aspirate in the treatment of full thickness chondral lesions

Ref.	Tissue	BMAC preparation	Concen-tration	Study design/methods/follow up	Outcomes measured	Results	LOE
Enea et al[4]	Knee	60 mL BMA from iliac crest processed with MarrowStim Concentration Kit (Biomet) resulting in 3-4 mL of BMAC. Chondral lesion debrided and microfracture performed. Biocollagen MeRE collagen membrane (Bioteck) cut to match shape and immersed in BMAC until implantation. 10:1 mixture of 1-2 mL fibrin glue and BMAC laid on lesion. Membrane inserted and placed. 2-3 mL of fibrin glue-BMAC injected over and left to solidify	NS	n = 9. Arthroscopic microfracture covered with collagen membrane immersed in autologous BMAC from iliac crest. Follow up: 29 mo	Biopsy cartilage evaluated by surgeon using criteria of international cartilage repair society. The following items were utilized: Cartilage repair assessment, MRI, IKDC, Lysholm, VAS (pre and post op), Tegner (pre and post op). Four patients had second look arthroscopy and biopsy	Significant clinical improvement (P < 0.005). Cartilage macroscopic assessment at 12 mo revealed all repairs appeared almost normal. Histo-analysis showed hyaline-like cartilage repair in 1 lesion, fibrocartilaginous repair in 2 lesions and a mixture of both in 1 lesion. Post op MRIs (6-9 mo out) all showed reconstitution of original cartilage. Bone marrow edema and/or subchondral irregularities observed in all cases. Non-homogeneous cartilage signal and fissuring observed in 2 of 3 cases	IV
Enea et al[5]	Knee	60 mL of BMA from the iliac crest was obtained and processed with MarrowStim Concentration Kit (Biomet) to obtain 3-4 mL of BMAC. Cartilage was treated with arthroscopic microfracture and the defect was covered with PGA-HA scaffold matrix (Chondrotissue) seeded with autologous BMAC. 10:1 mixture of 1-2 mL of fibrin glue and BMAC was then applied to lesion bed. PGA-HA soaked in BMAC was then applied with 2-3 mL additional fibrin glue-BMAC mixture dispersed over the matrix until solidification at 2-3 min	NS	n = 9 (Outerbridge type III/IV) Consecutively treated with arthroscopic Polyglycolic acid/hyaluronan - covered microfracture and BMAC. Follow up: 22 mo	Clinical scoring, IKDC, Lyshold, VAS, Tegner, cartilage microscopic examination at 12 mo, MRI at 8-12 mo post op. 5 patients underwent second look and 2 had biopsy	All patients but one showed improvement in clinical scoring from pre-op sto last follow-up (22 mo). All other variables increased from baseline to latest follow-up. Nineteen cartilage exams appeared normal, three almost normal, and one abnormal at 12 mo. Histo showed hyaline-like cartilage repair tissue formation in one case. MRI showed complete defect filling	IV
Gigante et al[6]	Knee	NA	NA	n = 5. MACI augmented with BMAC	Second look arthroscopy biopsy, CRA, ICRS II Visual Histological Assessment Scale	Normal ICRS/CRA at arthroscopic evaluation and had mean overall histological ICRS II of 59.8 ± 14.5. Hyaline-like matrix only found in one case. Mixture of hyaline/fibrocartilage was found in one case and fibrocartilage was found three cases	IV
Gobbi et al[7]	Patello-femoral	60 mL of BMA from ipsilateral iliac crest concentrated by BMAC Harvest Smart PreP2 system to obtain concentration of BMC 4-6 times baseline value	4-6 × baseline	(1) MACI n = 19; (2) BMAC n = 18. Both with HYAFF1 scaffold. Follow up: 3 yr	XR, MRI, IKDC score, KIOOS score, VAS, Tegner	Both groups showed significant improvements in all scores from preop to final follow up (P = 0.002). There was no difference between the two groups except in the IKDC subjective scores which favored BMAC group (P = 0.015). MRI showed complete filling of defect in 76% of MACI and in 81% of BMAC	III
Gobbi et al[8]	Knee	60 mL of BMA from ipsilateral iliac crest concentrated by BMAC Harvest Smart PreP2 system to obtain concentration of BMC 4-6 times baseline value. Activated using batroxobin enzyme to form sticky clot. Implanted and covered with collagen-based membrane scaffold (ChondroGide) and sealed with fibrin glue (Tissucol)	4-6 × baseline	n = 25. Cartilage transplantation with mulipotent stem cells and collagen type I/III matrix	XR, MRI, VAS, IKDC, KOOS, Lysholm, Marx, Tegner	Significant improvement at follow up across all measures. < 45-year-old and smaller lesions = better results. MRI = good stability of implant, hyaline-like cartilage found is histo analysis of biopsied tissue	IV
Gobbi et al[9]	Knee	60mL BMA from ipsilateral iliac crest (PreP2) and concentrated to 4-6 times baseline value, after activation with batroxobin enzyme (Plateltex Act) and pasted into lesion Covered with collagen type I/III matrix (Chondro-Gide) and sealed with fibrin glue (Tissucol)	4-6 × baseline	n = 15. One step surgery with BMAC and Collagen I/III matrix (chondro-gide)	XR, MRI at 1 and 2 yr. VAS, IKDC, KOOS, Lysholm, Marx, SF-36, Tegner at 6, 12, 24 mo. 3 had second look biopsy	Significant improvement at follow up across all measures (P < 0.0005). Single lesion and smaller lesions had better improvement. MRI showed greater hyaline-like tissue in all patients. Hyaline-like cartilage on histology in 3 biopsies	IV
Krych et al[10]	Distal femur	NS	NS	(1) n = 11 control scaffold; (2) n = 23 scaffold + PRP; (3) n = 12 scaffold + BMAC. Follow up: 12 mo	MRI, T2 mapping	BMAC and PRP groups had superior cartilage infill (P = 0.002, P = 0.03). BMAC demonstrated mean T2 value closer to that of superficial hyaline cartilage (P = 0.01)	III

BMA: Bone marrow aspirate; NS: Not significant; CRA: Cartilage repair assessment; MRI: Magnetic resonance imaging; MACI: Matrix-induced autologous chondrocyte implantation; PRP: Platelet-rich plasma.

Table 2 Studies evaluating concentrated bone marrow aspirate in the treatment of osteochondral defects

Ref.	Tissue	BMAC preparation	Concen-tration	Study design/methods/follow up	Outcomes measured	Results	LOE
Buda et al[11]	OCL of talus	Scaffold was a hyaluronic acid membrane loaded with previously cultured chondrocytes (ACI) or with BMAC. Platelet rich fibrin gel was produced the day before surgery using Vivostat System 1 (vivolution A/S). Harvested and processed 120 mL of the patient’s venous blood to obtain 6 mL of platelet rich fibrin gel. 60 mL BMA was harvested from posterior iliac crest using Smart PRepI to obtain 6 mL of BMAC. 1 g powder mixed with 2 mL BMAC and 1 mL platelet rich fibrin gel. The hyaluronic acid membrane was cut and loaded with 2 mL BMAC and 1 mL platelet rich fibrin gel. A layer of platelet rich fibrin gel was placed over the implant once in place to provide additional stability	NS	n (total) = 80: (1) n = 40 - autologous chondrocytes implantation; (2) n = 40 with BMAC. Follow up: 48 mo	Clinical scores, XR, MRI Mocart score, T2 mapping	Groups had similar results at 48 mo. No statistically significant difference in clinical outcomes. Return to sport was slightly better with BMAC. MRI MOCART score was similar in both groups. T2 mapping highlighted a higher presence of hyaline like values and lower incidence of fibrocartilage in BMAC group	IV
Buda et al[12]	OCL of knee	Combined with either MAST or HA matrix	NS	n = 30. One step arthroscopic BMAC transplant with scaffold. Follow up: 29 mo	Clinical inspection, MRI, IKDC, KOOS	Good clinical outcome and osteochondral regeneration on MRI and biopsies in both groups	IV
Buda et al[13]	OCL of talus	Scaffolds either: (1) porcine collagen powder SpongostanI Powder (J and J) mixed with autologous cell concentrate and platelet gel; or (2) hyaluronic acid membrane (fidia advanced biopolymers) with addition of platelet gel. Platelet rich fibrin gel was produced the day before surgery using Vivostat System 1 (vivolution A/S). Harvested and processed 120 mL of the patient’s venous blood to obtain 6 mL of platelet rich fibrin gel. 60 mL BMA was harvested from posterior iliac crest using Smart PRepI to obtain 6mL of BMAC. 1 g powder mixed with 2 mL BMAC and 1ml platelet rich fibrin gel. The hyaluronic acid membrane was cut and loaded with 2 mL BMAC and 1 mL platelet rich fibrin gel. A layer of platelet rich fibrin gel was placed over implant once in place to provide additional stability	NS	n = 64. One step arthroscopic BMAC transplant with scaffold (collagen powder of hyaluronic acid membrane) and platelet gel. Follow up: 53 mo	AOFAS scale score, radiographic, scaffold type, lesion area, previous surgery, lesion depth	Mean preop AOFAS was 65.2. Regardless of scaffolding type all patients showed similar pattern of clinical improvement at each follow-up. No correlation between area of lesion and pre-op AOFAS score but did observe relationship between area and AOFAS at each follow up post-operatively. No relationship between AOFAS score and depth of lesion	IV
Buda et al[14]	OCL of knee	Scaffold either MAST or HA matrix + PRF	NS	n = 20. Follow up: 24 mo	Clinical, MRI	Significant improvement at 12 and 24 mo, satisfactory MRI	IV
Giannini et al[15]	OCL of talus	Porcine collagen powder (J and J) or hyaluronic membrane scaffold. 60 mL of bone marrow harvested from posterior iliac crest and concentrated by SmartPrep to 6 mL of BMC. One step delivery system	NS	n = 49 received either BMA with collagen scaffold or BMA with HA membrane scaffold. Follow up: 48 ± 6 mo	AOFAS, radiograph, MRI	AOFAS improved P < 0.0005. T2 mapping analysis showed regenerated tissue with T2 values similar to hyaline cartilage in a mean of 78% of the repaired lesion area	IV
Giannini et al[16]	OCL of talus	One step arthroscopic transplantation. Platelet gel using Vivostat system. 60 mL BMA harvested from posterior iliac crest. Concentrated using SmartPReP in order to obtain 6 mL of concentrate. Scaffold: Either collagen powder (Spongostan1 Powder) or hyalyronic acid membrane. Scaffold was loaded with 2 mL BMAC and 1 mL PRF	NS	n = 25 in BMAC group. Study also compared to ACI	AOFAS, histology	Statistically significant improvement in mean AOFAS scores post-operatively (P < 0.0005). Only 1 superficial infection noted. Nearly homogeneous regenerated tissue on MOCART MRI in 82% of cases. Hypertrophy found in 2 cases on histology	IV
Giannini et al[17]	OCL of talus	Porcine collagen powder (J and J) or hyaluronic membrane scaffold. 60 mL of bone marrow harvested from posterior iliac crest and concentrated by SmartPrep to 6 mL of BMC. One step delivery system	NS	(1) n = 23 - Collagen scaffold + BMA; (2) n = 25 HA membrane scaffold + BMA. Follow up: 29 mo (24-35)	AOFAS, histology	AOFAS improved, Histology showed regenerated tissue in various degrees of remodeling	IV
Gobbi et al[18]	OCL of knee	Hyaluronic acid-based scaffold was used with BMAC	6 × baseline	n = 25 HA-BMAC, n = 25 microfracture. Observed prospectively for 5 yr	Patient-reported scoring tools: IKDC Subjective Knee Evaluation, KOOS, Lysholm Knee Questionnaire, and Tegner activity scale	Microfracture - 64% normal/nearly normal according to IKDC objective score at 2 yr and declined to 28% at 5 yr	II
		60 cc of BMA from Iliac Crest spun to 6 × normal concentration. Batroxobin enzyme used to activate BMAC				HA-BMAC - 100% normal/nearly normal objective IKDC at 2 yr, 100% at 5 yr for ALL outcomes measured
Hannon et al[19]	OCL of talus	60 mL of BMA from ipsilateral iliac crest, concentrated by Arteriocyte Magellan Autologous Platelet Separator System to obtain 3 mL of BMAC	NS	(1) n = 12 BMS; (2) n = 22 BMAC+BMS. Follow up: 48.3 mo for BMAC + BMS, 78.3 mo for BMS	AOFAS, FAOS, SF-12, MOCART	Mean FAOS and SF-12 PCS scores improved pre to post operatively (P < 0.01) for both groups. MOCART score significantly higher in cBMA + BMS (P = 0.023). T2 relaxation values in cBMA + BMS group significantly higher with measurements of adjacent cartilage	III
Kennedy et al[2]	OCL of talus	60 mL of BMA from ipsilateral iliac crest, concentrated by commercially available BMAC centrifuge system to obtain 4 mL of pluripotent cells	NS	n = 72. AOT with BMAC. Follow up: 28 mo	FAOS, SF-12	FAOS, SF-12 significantly improved from pre to post-op	III

KOOS: Knee injury and Osteoarthritis Outcome Score; NS: Not significant; OCL: Osteochondral lesions; BMA: Bone marrow aspirate; MRI: Magnetic resonance imaging.

Table 3 Studies evaluating concentrated bone marrow aspirate in the treatment of osteoarthritis

Ref.	Tissue	BMAC preparation	Concen-tration	Study design/methods/follow up	Outcomes measured	Results	LOE
Centeno et al[20]	Knee	60 mL of BMA from iliac crest was obtained toproduce 1-3 mL of BMAC. 60 cc of heparinized IV venous blood drawn to be used for isolating PRP and platelet lysate. Lipoaspirate - miniliposuction of the posterior superior buttocks or lateral thigh was performed under ultrasound and minimally processed (centrifuged) adipose tissue was injected into the articular space. Preparations were injected into the articular space of the knee together (5-10 cc) between the meniscus on the most painful side and over lying collateral ligament	NS	Data from registry. (1) n = 616 - BMAC+ PRP vs (2) BMAC + PRP + adipose graft. Outcomes and complication questionnaires at 1, 3, 6, 12 mo completed. 2 groups (A-BMAC and PRP protocol, B BMAC and PRP plus adipose fate graft (lipoaspirate)	LEFS, NPS, subjective percentage improvement rating, frequency and type of adverse events	Mean LEFS score increased in both groups and mean NPS decreased in both groups. AE rates were 6% without graft and 8.9% with graft. No difference between groups. Addition of adipose graft did not provide a detectible benefit over BMAC alone	IV
Centeno et al[21]	Knee	10-15 cc whole bone marrow aspirate harvested from 6-8 sites on posterior iliac crest (3-4 each side). Centrifuged and cells isolated. Patient heparinized blood for PRP and PL. Aspirates mixed together and injected into joint. Cell counts were counted four times and average was taken under microscope for total nucleated cell count	Lower and higher cell count groups defined using threshold of 4 × 104 cells	Data from registry. n = 373 patients that received BMAC combined with PRP and PL injections for 424 OA knees	Clinical scales assessed at baseline, 1, 3, 6, 12 and annually thereafter. NPS, LEFS, pain and functional outcome measures	Significant positive results with treatment for all pain and functional metrics. Higher cell group reported lower post treatment numeric pain scale values (P < 0.001). No significant difference detected for other metrics	IV
Haleem et al[22]	Femoral condyle	20 mL BMA from iliac crest isolated with density gradient (Ficoll-Paque), supplemented with 10% fetal bovine serum and penicillin streptomycin. Microfracture performed and sclerotic bone curetted. Autologous periosteal flap harvested from anteromedial ispilateral proximal tibia to fit defect size and stuffed into place. 1 mL platelet concentrate and 1 mL fibrinogen and 1 mL thrombin placed with BMAC PR fibrin glue	NS	n = 5, treated with BMAC + PRF	At 6 and 12 mo: Lysholm and Revised HHS Knee Score, XR and MRI. 2 patients had follow up arthroscopy at 12 mo rated by ICRS	All patients had statistically significant improvement at 6 and 12 mo (P < 0.005). No statistically significant difference between 6 and 12 mo post op in clinical scores. ICRS were near normal for 2 patients who consented to arthroscopy. MRI of 3 patients at 12 mo showed complete defect filling and complete surface congruity with native cartilage. Two patients showed incomplete congruity. BMAC on platelet rich fibrin gel as a scaffold may be effective to promote repair of articular cartilage defects	IV
Koh et al[23]	Knee	60 mL BMA from Iliac crest processed with MarrowStim Concentration Kit (Biomet) to obtain 3-4 mL of BMAC. Adipose tissue harvested from buttocks through liposuction. All fluid removed from knee arthroscopically. Lesion filled with cell suspension and held stationary for 10 minutes with defect facing upwards. Adherence of MSC confirmed. No marrow stimulation procedures were performed	Average of 3.8 × 106 (2.5-6.1 × 106)	n = 37 knees using second-look arthroscopy after mesenchymal stem cell implantation for cartilage lesions done 12 mo post op	IKDC, Tegner, cartilage repair assessed using ICRS grading	IKDC and Tegner sores significantly improved (P < 0.001). ICRS overall repair grades 2/37 were normal, 7/37 were near normal, 20/37 abnormal, 8/37 severely abnormal.). Patient satisfaction: 33/34 reported good to excellent satisfaction. High BMI (> 27.5) and large lesion (> 5.4 cm2) had significant prediction of poor clinical and arthroscopic outcomes (P < 0.05)	IV
Shapiro et al[24]	Knee	52 mL BMA from iliac crest concentrated in Arteriocyte Magellan Autologous Platelet Separator System centrifuge to yield 6 mL of cellular product	NS	n = 25 BMAC, n = 25 saline (patients had bilateral knee pain)	OARSI measure, VAS score, safety outcomes, pain relief, function	OARSI and VAS decreased significantly from baseline at 1wk, 3 mo, 6 mo (P < 0.019), no difference in pain relief	II

BMA: Bone marrow aspirate; MRI: Magnetic resonance imaging; NS: Not significant; OA: Osteoarthritis; BMI: Body mass index; VAS: Visual analogue scale; OARSI: Osteoarthritis Research Society International.

Table 4 Studies evaluating concentrated bone marrow aspirate in bone healing

Ref.	Tissue	BMAC preparation	Concen-tration	Study design/methods/follow up	Outcomes measured	Results	LOE
Bastos Filho et al[25]	Tibia/femur nonunion	11G × 10 cm bone marrow aspiration needle into posterior iliac crest to obtain a total of 100 to 110 mL for each patient - concentrated to 20 mL with Sepax system	NS	n = 6 patients with nonunion of tibia or femur. Four received percutaneous infusion of autologous bone marrow aspirated without Sepax processing. Two received with processing. Follow up to 6 mo	Clinical examination and radiographic evaluation at 2, 4, 6 mo. Clinical criteria included full weight bearing tolerance and absence of pain upon palpation at the fracture site. Radiographic healing checked with AP, lateral and oblique films to look for bone callus. Patient satisfaction questionnaire scale from 0-10	Bone consolidation obtained in all the patients. Bone callus observed in the radiographic between 3 and 24 wk, average 13.8 wk in group without processing. Mean satisfaction increased in all patients	II
Desai et al[26]	Nonunion/delayed union of tibia	Total of 60 cc bone marrow aspirated from iliac crest with 16 gauge Jamshidi needle (Harvest system). Concentrated to 10 cc for injection	101.48 ± 64.13/cc	n = 49 patients with tibial nonunion had BMAC injection with DBM and/or rhBMP-2. Follow up until radiographic union or another procedure was performed	Radiographic healing (bridging of 3 out of 4 cortices on AP and lateral films)	No difference in healing rate between patients with fracture gaps less than and greater than 5 mm	III
Garnavos et al[27]	Humeral shaft delayed union	With the use of a 10 cm long and 3 mm wide biopsy needle, 60 mL of bone marrow was aspirated from each patient’s iliac wing and was centrifuged to provide 10 mL of concentrated mesenchymal stem cells. The concentrated bone marrow mixed with 10 cc of DBM putty	NS	n = 5. Intramedullary nailing with antegrade/unreamed technique was performed for 4 patients. One patient was treated previously with retrograde/unreamed nailing left in situ. The concentrated mixture was infused percutaneously in the area of nonunion with a biopsy needle under fluoroscopy. Patients were followed up every 4-6 wk for 12 mo	Patients were assessed for union process, discomfort, level of activities and functional improvement	There were no peri-or postoperative complications. Sound union was obtained in all cases from 12 to 20 wk after the operation. At final followup, all patients had regained a satisfactory range of shoulder and elbow motion. They had also returned to pre-injury level of activities and were happy with their treatment and outcome	IV
Guimaraes et al[28]	Femoral shaft nonunion	11G × 10 cm needle used for aspiration from iliac crest. The marrow samples were harvested in small amount (2 mL) and the contents of each syringe were pooled in the container of the bone-marrow-collection kit containing anticoagulant solution. The final volume of bone marrow aspirate (200 mL) was then filtered through a sequence of successively smaller-diameter mesh filters. The cells were finally collected in a blood transfer pack unit. The aspirated material was reduced to a final volume of 40 mL by removing most of the RBC the plasma by centrifugation	9.8 ± 4.3 × 106 vs 20.2 ± 8.6 × 106	n = 16 patients with aseptic nonunion of femur were treated with injection of BM-MSCs who had locked IMN. Follow up: 3-8 mo	Radiographic RUST scores	Bone union occurred in 8 of 16 patients according to RUST. The grafts used in patients whom treatment failed contained significantly lower number of total nucleated cells (9.8 ± 4.3 × 106 vs 20.2 ± 8.6 × 106)	IV
Hernigou et al[29]	Ankle nonunion	150 mL of bone marrow aspirate obtained from anterior portion of the ipsilateral iliac crest then treated with a cell separator	27.3 ± 14.6 × 106	n = 86 ankle nonunion in diabetic patients treated with BM-MSCs vs n = 86 diabetic matched nonunion treated with a standard bone iliac crest autograft	Time of union, callus volume, complication, morbidity of graft harvesting vs bone marrow aspiration in diabetic patients	70 out of 86 patients (82.1%) treated with BMC achieved healing with a low number of complications; 53 (62.3%) of patients treated with iliac bone graft had healing and major complications were observed: Amputations, osteonecrosis of fracture wound edge, infections	III
Hernigou et al[30]	Tibial shaft nonunion	Bone marrow aspirated from anterior iliac crest total of 300 mL then concentrated to 50 mL	18 ± 7 million	BMAC injected into 60 noninfected atrophic nonunion of tibia. Follow up until union	Radiographic union; healing time; volume of callus	Patients who did not achieve union had significantly lower number of progenitor cells comparing to the 53 patients who achieved union. There was positive correlation between the volume of mineralized callus at 4 mo and the number and concentration of fibroblast colony-forming units in the graft; there was a negative correlation between the time needed to obtain union and the concentration of CFU in the graft	IV
Ismail et al[31]	Long bone nonunion	40 mL of bone marrow was aspirated from posterior iliac crest and transferred into a container prefilled with 5000 U/mL of heparin. Aspirate was diluted with phosphate-buffered saline at a ratio of 1:1 and centrifuged at room temperature at 3000 rpm for 30 min. The collected buffy coat was washed and transferred into a culture flask containing Dulbecco's Modified Eagle Medium supplemented with 10% fetal bovine serum. Cells were incubated at 37 °C at 5% CO2 with a routine culture medium change every two to three days. Subculture was performed between days 7 and 10. Mixed with 5 g/cm3 defect of HA granules	14-18 million BMSCs	n (total) = 10. n = 5, treated with combination of 15 million BM-MSCs, 5 g/cm3 (HA) granules and internal fixation. n = 5, control subjects were treated with iliac crest autograft, 5 g/cm3 HA granules with internal fixation. Follow up = 12 mo	VAS, LEFS, DASH score. Radiological assessments for union were conducted by a blinded radiologist using two radiological scoring systems: The Lane-Sandhu and Tiedeman radiological scores	No significant differences in post-op pain between the two groups. The treatment group demonstrated initial radiographic and functional improvements. Statistically significant differences in functional scores were present during the first (P = 0.002), second (P = 0.005) and third (P = 0.01) month. Both groups achieved similar outcomes by the end of one year follow up	III
Le Nail et al[32]	Open tibia fracture	Hernigou’s technique. Bone marrow from posterior iliac crest by needle aspiration. Around 500 mL concentrated by centrifugation to obtain 50 mL	171 ± 107 × 106 vs 118 ± 28 × 106	n = 43 cases of open tibial fractures with initial surgical treatment that developed nonunion or delayed union, subsequently treated with injection of BMAC	Clinical success (consolidation without any subsequent procedure): Non painful callus palpation and a full weight bearing without any contention system. Radiographic bone healing 3 out of 4 cortices	23 successes (53.5%) within 17 wk after BMAC	IV
Thua et al[33]	Long bone nonunion	BMA (300-350 mL) were obtained by Jamshidi vacuum. Both posterior iliac crests of patients were harvested under loco-regional anaesthesia. BMAC was produced via density gradient centrifugation using the Sorvall centrifuge at 3670 rpm for 7 min. Afterwards, a total volume of 8 mL BMAC was mixed with freeze-dried allograft cancellous bone chips. BMAC was incubated for 15 min with bone chips as a composite of BMAC-ACB prior to transplantation	2.43 ± 1.03 (× 106) CD34 cells/mL (staining)	n (total) = 27. n = 9 control treated with autologous cancellous bone graft from iliac crest. n = 18 clinical trial group treated with BMSCs and allograft cancellous bone chips. Correction and optimization of fixation device were done for previously failed procedures. Patients were followed up in outpatient clinic for 1, 3, 6, 9 12, 18, 24 mo	Functional outcomes, radiographic outcomes based on modified Lane and Sandhu radiological scoring system	Bone consolidation was obtained in 88.9% and mean interval between cell transplantation and union was 4.6 ± 1.5 months in autograft group. Bone union rate was 94.4% in group of composite BMAC-ACB implantation. The time to union in BMAC-ACB grafting group was 3.3 ± 0.9 mo, and led to faster healing when compared to the autograft	III

NS: Not significant; BM-MSC: Bone marrow-derived mesenchymal stem cell; BMA: Bone marrow aspirate; RBC: Red blood cell; CFU: Colony-forming units; BMSC: Bone marrow derived stroma cell.

Table 5 Studies evaluating concentrated bone marrow aspirate in tendon repair

Ref.	Tissue	BMAC preparation	Concen-tration	Study design/methods/follow up	Outcomes measured	Results	Level of evidence
Hernigou et al[34]	Rotator cuff	150 mL BMA from iliac crest mixed with an anticoagulant solution (citric acid, sodium citrate, dextrose). MSCs were injected in the tendon at the junction between the bone and tendon (4 mL), and in the bone at the site of the footprint (8 mL). Each patient in the MSC-treated group received a total of 12 mL of bone marrow concentrate	51000 ± 25000 cells in 12 mL of injected BMC	n = 45 received MSCs during repair. n = 45 matched control group of 45 patients who did not receive MSCs. Follow up: 3, 6, 12, 24 mo and 10 yr	RTC healing and re-tear rate confirmed by ultrasound and MRI	45/45 repairs with MSC augmentation had healed by six months vs 30/45 repairs without MSC treatment by 6 mo. Intact rotator cuffs were found in 39/45 patients in the MSC-treated group, but just 20/45 patients in the control group. Patients with a loss of tendon integrity at any time up to the ten-year follow-up milestone received fewer MSCs as compared with those who had maintained a successful repair during the same interval	III
Hernigou et al[35]	Tendon-bone interface rotator cuff	NS	NS	n = 125 symptomatic patients. n = 75 control patients. Assessed the level of MSCs in the tuberosity of the shoulder of patients undergoing a rotator cuff repair	Mesenchymal stem cell content at the tendon–bone interface tuberosity was evaluated by bone marrow aspiration collected in the humeral tuberosities of patients at the beginning of surgery	A significant reduction in MSC content (from moderate, 30%-50%, to severe > 70%) at the tendon–bone interface tuberosity relative to the MSC content of the control was seen in all rotator cuff repair study patients. Severity of the decrease was statistically correlated to the delay between onset of symptoms and surgery, number of involved tendons, fatty infiltration stage and increasing patient age	III
Mazzocca et al[36]	Rotator cuff	MSCs were exposed to either insulin or tendon-inducing growth factors or were left untreated to serve as a control. The BMA was overlaid onto a 17.5% sucrose gradient and centrifuged for 5 min at 1500 rpm (205 g), and the resulting pink middle layer was obtained. After the isolation of bone marrow, MSCs were exposed to a 1-time dose of 10-9-mol/L, 10-10-mol/L, 10-12-mol/L, or 10-13-mol/L insulin from bovine pancreas or were left untreated to serve as a control	NS	n = 11 patients undergoing arthroscopic RCR. After the determination of the optimal dose of insulin, MSCs were (1) exposed to the hormone insulin; (2) exposed to the growth factors IGF-1, bFGF, and GDF-5, which served as a positive control for MSCs’ differentiation into a tendon; or (3) left untreated to serve as a negative control. In the growth factor group, MSCs were treated with a 1-time dose, 10 ng/L, of IGF-1, bFGF, and GDF-5 or 10–10-mol/L insulin	Cell count, gene expression, protein analysis, and immunocytochemical analysis. Confirmation of protein levels was verified on immunocytochemistry analysis by 4 independent evaluators blinded to group assignment	MSCs treated with insulin showed increased gene expression of tendon-specific markers (P > 0.05), increased content of tendon-specific proteins (P > 0.05), and increased receptors on the cell surface (P > 0.05) compared with control cells. Histologic analysis showed a tendon-like appearance compared with the control cells	III
Mazzocca et al[37]	Rotator cuff	Isolation 1: one 5 min centrifugation at 1500 rpm in which BMA was overlaid onto a 17.5% sucrose gradient in a 50-mL conical tube followed by extraction of CTPs in the fractional layer. Isolation 2:30 min centrifugation at 1500 rpm followed by fractionated layer extraction of CTPs using a Histopaque gradient	Nucleated cells harvested from fractionated layer were counted and plated on 100 mm Primaria dishes at a concentration of 0.5 × 106 cells/9.6 cm2 then incubated	n = 23 BMAC harvested through the anchor tunnel of the humeral head during arthroscopy. n = 23 matched controls. Mean time to follow-up was 10.6 ± 6.7 mo in the aspirate group and 10.0 ± 6.2 mo in the control group	Reverse transcription polymerase chain reaction analysis, Single Assessment Numeric Evaluation score	Reverse transcription polymerase chain reaction analysis and cellular staining confirmed the osteogenic potential of the connective tissue progenitor cells. There was no statistically significant difference in the Single Assessment Numeric Evaluation score, range of motion measures or post-operative strength measures between groups	III
Stein et al[38]	Achilles	30 to 60 mL of BMA, combined with a standardized mixture of anticoagulant citrate dextrose solution A and separated by centrifugation at 3200 rpm for 15 min. The aspirate was concentrated to yield a volume of 6-9 mL of BMAC	NS	n = 28 open repairs with BMAC. Mean follow up: 29.7 mo. Patients were followed postoperatively at two weeks, six weeks, three months, six months, one year and annually thereafter	Calf atrophy, maximum dorsi- and plantarflexion, and fatigue limit during single-limb heel raise. Functional and activity status was measured in terms of time to walking, light activity (such as cycling or jogging) and return to sport, as with the validated Achilles Total Rupture Score. Self-reported functional status, activity level and ATRS	All patients achieved good or excellent outcomes postoperatively by attaining functional use or return to sport. At final follow-up of 29.7 ± 6.1 mo, mean calf circumference for paired operative and nonoperative extremities was 37.7 ± 2.0 and 38.2 ± 2.0 (difference - 0.5 ± 1.3) cm, respectively, for the 26 patients with single Achilles tendon repair. Walking without a boot was at 1.8 ± 0.7 mo, and participation in light activity was at 3.4 ± 1.8 mo. Overall, 92% (25 of 27) patients returned to their preferred sport successfully at 5.9 ± 1.8 mo. Mean ATRS at final follow-up was 91 (range 72-100) points, with no single mean item score below 8 points. All patients were able to achieve a ROM of neutral dorsiflexion or greater and were able to successfully perform a single-limb heel raise at final follow-up	IV

NS: Not significant; MSC: Mesenchymal stem cell; BMA: Bone marrow aspirate.