Introduction

Fracture healing encompasses a succession of dynamic multifactorial metabolic events, which ultimately re-establishes the integrity of the biomechanical properties of the bone. Einhorn et al. highlighted that up to 10% of the fractures occurring annually will need additional surgical procedures because of impaired healing [1,2,3]. Non-union is a significant orthopaedic problem, which is defined as the arrest of progression to union at a fracture site with persisting pain and mobility at the fracture site for a minimum period of six months from injury and no progression on three monthly serial x-rays [4]. Individualized surgical treatment for a fracture is warranted when the surgeon believes the fracture has little or no chance to heal. Non-unions occur in the main due to biological impairment, mechanical factors or both. The cause of a non-union can be attributed to the patient, pharmacological factors, injury and treatment related factors. Patient related factors include old age, poor compliance with rehabilitation, malnutrition, smoking, alcoholism, diabetes, peripheral neuropathy and the immunosuppression. A number of pharmacological agents have been associated with nonunion, these include steroids, cytotoxins, ciprofloxacin, NSAIDS and irradiation. Injury related factors include open fractures, significant soft tissue trauma, soft tissue interposition, infection, pathological fractures, excessive bone loss, segmental injury and comminution. Treatment related factors include excessive distraction of a fracture, inadequate stability with excessive movement and an extensive approach with vascular compromise [1,2,3, 5, 6].

The management of nonunion is a challenge to many orthopaedic surgeons and represents a significant clinical problem. The basic concept behind treatment is to provide both mechanical and biological support to the nonunion site. The aim is to restore mechanical stability with adequate strain [7] in a biologically sound milieu. The biological stimuli for the regeneration of bone involves the interaction of osteoinductive growth signals, stem cells that respond to these signals, an intact vascular bed and a scaffold that supports cellular proliferation and ingrowth [8]. The classic treatment of small defects (<3 cm) involves surgical stabilization and open autologous bone grafting with success rates of 50%–80% reported. However, this has been associated with donor-site morbidity and reduced healing potential in elderly patients [1,2,3, 5]. For larger defects (>3 cm), segmental transport represents the gold standard. However, this method is time consuming and associated with several complications, such as pin-track infections, pseudarthrosis, psychological problems, neurovascular complications and nonunion [5, 6].

There is interest in the development of alternate techniques using mechanical or biological methods, to provide the benefits of bone grafting with lower complication rates, morbidity and improved results [5]. The mechanical method includes the use of mechanical stimulation, electromagnetic fields and low-intensity ultrasound. The biological approach involves the use of osteoconductive biomaterials and osteoinductive factors to support and promote the ingrowth of newly formed bone. Osteoconductive materials include autologous bone graft, demineralized bone matrix (DBM), hydroxyapatite (HA) and tricalcium phosphate (TCP). Licenced osteoinductive factors are currently limited to bone morphogenic proteins, either naturally occurring within bone graft or from an exogenous source [2, 3, 6, 9].

Mesenchymal stem cells (MSC) are believed to have multipotent plasticity with the capability to differentiate along multiple cell lineages such as cartilage, bone, tendon, muscle and nerve [10,11,12]. Such multipotency has the potential to play a prominent therapeutic role in the repair and reconstruction of multiple tissues across a number of orthopaedic specialties [13]. Bone marrow is the most popular source of MSCs [14, 15] and historically many surgeons have utilized unprocessed bone marrow aspirate (BMA) [9, 16,17,18] in an attempt to stimulate healing. Currently, BMA is most commonly obtained from the iliac crest. There is no difference in the total number of cells obtained when harvesting from the anterior compared to the posterior pelvis, however posterior crest provides more connective tissue progenitors. Only a small percentage of mesenchymal stem cells (MSCs) can be obtained through aspiration of the marrow. Approximately 0.01% of the cells in BMA are MSCs, with the total number of viable cells obtained decreasing with age.

Bone marrow aspirate concentrate (BMAC) is an attempt to improve the recovery of the nucleated cells from marrow aspirate, while decreasing the recovery of non-nucleated cells such as RBCs. The exact mechanism of action of BMAC is currently not fully understood. Potentially the MSCs contained within BMAC will provide a direct cell source for repair of the host tissue. Alternatively or in addition to, the nucleated cells may have a paracrine effect by delivering various cytokines and growth factors into the ‘site to orchestrate and direct endogenous bone repair [19,20,21,22]. Through centrifugation the cell concentration can be increased 6–7 fold, the cellular content produces a number of growth factors, with Platelet derived growth factor (PDGF), transforming growth factor- β (TGF-β), and vascular endothelial growth factor (VEGF) likely to be the most important [23]. Fortier et al., compared the constituents in BMA and BMAC. Table 1 demonstrates that there are reduced platelets and raised white blood cells (WBC) in BMAC demonstrating that this is a very different substance to platelet rich plasma (PRP) with a likely different mechanism of action [24].

Table 1 Results of cytological analysis of bone marrow aspirate and bone marrow concentrate [24]
Full size table

The aim of this article is to review the current literature regarding the use of bone marrow aspirate concentrate, its effectiveness and potential complications when used to manage bone defects.

Methods

Eligibility

We have included all published clinical literature investigating the development, techniques and applications of bone marrow aspirate concentrate. Language, design and risk of bias did not deter the initial inclusion of any study. Our search was exclusively limited to studies involving human subjects published in English.

Search strategy

A PRISMA compliant search was carried out as published in 2009 [25]. This included the online databases: PubMed, EMBASE, clinical trial.gov and the Cochrane library from 1960 to the end of May 2015. MeSH terms used included: “bone” AND “marrow” AND “Aspirate” AND “concentrate” AND “bone defects” AND “non-union”.

Study identification

The title and abstract from each study within the results list was reviewed independently by three authors (XX, XX and XX). Any disagreement was resolved by discussion with the senior author. Full text papers of relevant studies were subsequently obtained and reviewed against the eligibility criteria. Then, full texts of the eligible studies were further evaluated and references were checked for more suitable studies.

Critical appraisal

Two authors using the Critical Appraisal Skills Program checklist independently appraised eligible studies. For the purpose of narrative review, relevant studies were included irrespective of methodology or level of evidence.

Results

Of the results yielded by the preliminary search 103 (48 PubMed and 55 EMBASE) were included. Exclusions included three duplicate records, six letters, 17 non-orthopaedics related studies and four records irrelevant to our search topic.

On searching www.ClinicalTrials.gov, we found that five trials were registered. All trials were examining the use of stem cells in bone healing and management of defects (Table 2).

Table 2 Clinical Trials of stem cells and orthopaedic applications registered at ClinicalTrials.gov (June 2016)
Full size table

The CASP appraisal [26] confirmed a satisfactory standard of 35 studies (Fig. 1). They all had clearly defined objectives, were well designed and conducted appropriately to meet them.

Fig. 1
figure 1

Flow diagram of the search results

Full size image

The published studies reported the use of BMAC in non-union and fracture healing (15 studies), bone defects (nine studies), spine fusion (two studies), distraction osteogensis (two studies) and complications related to the use of BMAC (seven studies).

Non-union and delayed union

Four studies have looked at the efficiency of a percutaneous injection of BMA for the management of nonunion [27,28,29,30]. In these studies a total of 301 fractures were managed by BMA, 268 (89%) demonstrated union with an average healing time of 2.5 to eight months. No study documented any adverse systemic effects.

Hernigou et al. [31] evaluated the outcome of injecting 20 cm3 of BMAC obtained from the iliac crest for the management of atrophic non-union of the tibia in 60 patients. The outcomes included the volume of the callus formed and rate of clinical union. Bone union was achieved in 53/60 (88%) patients at four months following the procedure. In the seven non-united tibias, the concentration (p = 0.001) and the total number (p < 0.01) of progenitors cells injected were significantly lower than in those that united. They reported a positive association between the quantity of hard callus and the number (p = 0.04) and concentration (p = 0.01) of fibroblast colony-forming (FCF) units in the graft. Likewise, they reported the time interval needed to achieve union was negatively correlated with the FCF units’ concentration at the site of the graft (p = 0.04). They concluded that BMAC was a one-step procedure, which may prevent complications related to in vitro cultivation, for example pre-ageing, reduced viability or de-differentiate.

The use of either DBM or recombinant human bone morphogenic protein-2 (rhBMP-2) in combination with BMAC injection in the management of nonunions in long bones has been assessed [32]. This surgical procedure is called the modified Hernigou technique. Desai et al. [32] in a total of 49 patients with nonunion in tibia compared BMAC injection with DBM and/or rhBMP-2. They assessed radiologic healing of the bone gaps which were either less or more than 5 mm. They reported no significant difference in the healing rate (p = 0.81) between patients with defects less than and greater than 5 mm. This study demonstrated that the application of BMAC in combination with either DBM and/or rhBMP-2 is an effective treatment for delayed or nonunion regardless of the fracture gap size or fracture site. The use of BMAC and rhBMP-2 was linked with lower healing rates in comparison to the use of BMAC and DBM (p = 0.036). Patients who had earlier intervention, had higher union rates (p = 0.04).

Kassem [17] assessed the outcomes of percutaneous injection of BMAC in the management of fractures presenting with delayed union or non union after open reduction and internal fixation (ORIF) in 20 patients. The BMAC injection was undertaken at an average of 9.65 (4 to 24) months after ORIF. BMAC was injected percutaneously into the fracture site under fluoroscopy control. They reported an achieved clinical and radiological union rate in 95% of the cases after a mean of 2.95 months.

Hernigou et al. [33] reported the use of an injection of BMAC at the site of non-unions in 86 ankles in diabetic patients. The outcomes were compared to 86 diabetic matched non-unions treated with bone graft harvested from the iliac crest. They found that the application of BMAC resulted in healing in 82.1% of the ankles compared to only 62.3% in the control group. Major complications were observed in the control group, these included amputations (5.8%), AVN (12.7%) and infection (20%). Fewer complications were recorded in the BMAC group. They concluded that the injection of BMAC might be desirable in view of the increased risks of major complications associated with open surgery and iliac bone grafting in this population. Moreover, percutaneous BMAC application is associated with improved healing rates compared with standard iliac bone autograft treatment.

The combination of BMAC and PRP has also been reported to be effective in the management of nonunion [31, 34,35,36]. It is unclear whether BMAC or PRP or both generate this outcome since they were not evaluated independently. However, it would appear that BMAC in isolation is sufficient to promote bone healing in the view of the aforementioned studies [17, 33, 37].

Bone defects and distraction osteogensis

Petri et al. [38] assessed five patients with segmental defects ranging from 3 to 14 cm managed with BMAC on a bovine scaffold. The healing process was evaluated by positron emission tomography–computed tomography (PET-CT) at three months post surgery. PET analysis showed an increased influx of fluoride by a factor of 8.3 ± 6.4 compared with the contralateral side (p < 0.01). Bone density in the cortical area was 75 ± 16% of the contralateral side (p < 0.03). They concluded that BMAC combined with a bovine scaffold could be an alternative option to segmental bone transport in management of large bone defects. However, this statement needs to be supported by further studies to prospectively compare this procedure to autologous bone grafting and segmental transport.

Sauerbier et al. [39] and Rickert et al. [40] carried out two randomized controlled clinical trials to examine the establishment of new bone in patients with severe atrophy of the maxillary sinus. Patients obtained an augmentation of the sinus using bovine bone either with BMAC or autologous bone. Sauerbier et al. [39] studied 45 severely atrophied maxillary sinus in 26 patients in a partial cross-over design. Thirty-four sinuses in 25 patients were augmented with bovine bone mineral (BBM) and BMAC. Eleven control sinuses in 11 patients were augmented with a mixture of 70% BBM and 30% autogenous bone (AB). Biopsies were obtained after 12 to 16 weeks. The authors found comparable new bone formation in both groups three to four months after surgery, which was similar to the results reported by Rickert et al. [40].

Distraction osteogenesis for segmental bone defect reconstruction involves long duration of treatment with external fixation, which can cause considerable morbidity and high complication rates. Augmentation with percutaneously applied adjuvants to reduce consolidation time has been designated as one of the major goals for future research in this field.

Kitoh et al. [41] reported the clinical outcomes of distraction osteogenesis with BMAC and PRP in three femurs and two tibias in three patients. BMAC and autologous PRP was applied into the distracted callus. The target lengths were obtained in 100% of bones without major complications with a mean healing index of 23.0 days/cm (18.8–26.9 days/cm). Although these results were preliminary, it was concluded that the use of BMAC combined with PRP was shown to be a safe and minimally invasive therapy, which reduces the treatment period by hastening bone regeneration during distraction osteogenesis.

Lee et al. [19] demonstrated in a randomized trial superior bone healing when using BMAC during distraction osteogenesis of the tibia. Twenty patients (40 tibias) undergoing bilateral tibial lengthening were enrolled in this study. They compared patients receiving an osteotomy site injection of BMAC and PRP (treatment group) with those not receiving such an injection (control group). Twenty segments (10 patients) were included in the treatment group and 20 segments (10 patients) received no injection with a reported minimum follow-up of 24 (24–34) months. All patients undertook lengthening for familial short stature, utilizing the lengthening over an IM nail. They reported similar average distraction rates between the two groups. The mean distraction rate was 0.75 mm/day in the treatment group compared to 0.72 mm/day in the control group (p = 0.24). The mean cortical healing indexes were significantly higher in the treatment group when compared to the control group (p < 0.001), presenting quicker healing in the treatment group at each cortex. Although the callus profile and type were not different between the two groups, full weight bearing was allowed earlier in the treatment group than in the control group (index: 0.99 months/cm and 1.38 months/cm, respectively, p < 0.001).

Jager et al. reported the clinical outcomes and radiological evaluation of ten patients with bone defects treated with BMAC. Additionally, they reported the in vitro data of BMAC cultivated onto a collagen scaffold. They demonstrated that there is a rationale for a clinical application of BMAC in the treatment of osseous defects. The intra-operative harvest procedure is a safe method and does not significantly prolong the time of surgery. In addition, MSC’s isolated from the aspirate was able to adhere and proliferate onto a collagen scaffold in significant numbers after a 15 minute incubation period. These cells were then able to undergo osteogenic differentiation in vitro without any osteogenic stimuli [36]. Similar satisfactory results were reported by Hendrich et al. [42] who evaluated new bone formation after the application of BMAC in 20 patients presenting with osseous defects. They reported no adverse events in those patients.

Although Kitoh et al. [41] and Lee et al. [19] established that the combination of BMAC and PRP enriched the healing of bone defects; it is also unclear in this scenario whether BMAC or PRP or both generate this outcome since they were not evaluated independently.

Other applications

The use of BMAC has been described in the treatment of simple or unicameral bone cysts (UBC). Different treatment options have been suggested but there is no consensus regarding the best procedure [43]. Di Bella et al. [44] in a level III therapeutic study compared the healing rates and failures of multiple injections of corticosteroid versus a single injection of DBM in association with BMAC in UBC with a minimum 12 months follow-up. They retrospectively reviewed 184 patients who had received either of these two methods. They observed a healing rate of 21% in the steroid group compared with a healing rate of 58% in the BMAC group. The rate of failure observed in the steroids group was 63% compared with 24% in the BMAC group. There were no differences observed in fracture rates between both groups.

Gan et al. [45] utilized BMAC with porous beta-tricalcium phosphate (beta-TCP) to augment spinal fusion in 41 subjects. A hard fusion was attained in more than 95% of patients studied. BMAC was combined with porous beta-TCP granules using negative pressure followed by a short-time incubation. In less than three years, 95% of the cohort of patients had satisfactory spinal fusion results. Of patients 4/41 (9.7%) had exudation or moderate swelling in their wounds, and all of them were treated successfully with conservative management.

Potential risks of BMAC therapy

Although dangers are acknowledged and anticipated, bone marrow aspiration is considered to be a safe procedure, however, adverse events have been reported [46,47,48,49]. These can be categorized into risks associated with the harvest and those associated with the administration of BMAC.

Harvest

Hernigou et al. [50] has defined the sector rule for aspiration of marrow from the iliac crest, which is based on safety zones. They divided the iliac crest into six equal sectors from anterior to posterior direction. The authors studied 480 trocar entry points undertaken by six surgeons in 120 patients. They demonstrated that the sector system consistently envisaged safe and unsafe zones for placing the trocar in the iliac crest. They observed increased risk of breaches in obese patients and this risk is decreased in more experienced surgeons. Ninety-four breaches out of 480 entry points occurred with increased risks observed in the thinner sectors in the iliac crest. Additionally, there is increased risk of injuring the external iliac artery in the four most anterior sectors (1 to 4) especially in females. On the other hand, posterior sectors were associated with increased risk of sciatic nerve and gluteal vessel injury when the trocar was inserted more than 6 cm into the posterior iliac crest. They concluded that the sector rule is a reliable system to use for BMA aspiration.

In 2001, Bain surveyed 19,259 procedures from 63 hospitals over three consecutive years in the largest study looking at potential risks of the harvesting procedure. Only 11 (0.0005%) of almost 20,000 patients experienced significant haemorrhage. Recent anticoagulation therapy appears to be the most significant risk factor associated with haemorrhage. Infection has been reported only in two subjects, both were superficial infections which were successfully treated with antibiotics [46]. Whilst dependent on the site of harvest, chronic pain can be a potential concern. The pain might not spontaneously resolve, and may require treatment using neuropathic pain medications [47, 49]. Pathological fractures are a potential concern especially in the presence of osteoporosis or osteomalacia [46,47,48,49]. Finally, a drastic adverse event of death has been reported in one patient in 2001, as a result of the formation of a retroperitoneal hematoma after an aspirate from the posterior iliac crest [46].

Administration

Infection is a potential concern at the administration site, although antibiotic prophylaxis is, standard practice and overall the general risk is low, particularly in percutaneous administration. When used intra-osseously, due to bone being permeable to liquefied material, fat embolisation is a potential risk [51]. Animal studies have shown fat globules in dogs’ lungs post mortem, however in human trials, adverse clinical outcomes in the form of respiratory complications or decreased oxygen saturation have not been reported [52]. Those subjects at a greater risk of embolization, such as those with cardiac shunts should be considered as to their suitability to receive intra-osseous BMAC. In all cases of intra osseous administration, patients should be monitored for the clinical signs of fat embolism [20].

Conclusions

In conclusion, MSCs in BMAC have the potential to self-renew, undertake clonal expansion and differentiate into different musculoskeletal tissues. These include osteoblasts, chondrocytes, fibroblasts and adipocytes. MSCs are also known to regulate the immune system and have a potential positive paracrine effect. BMAC has been mostly used to encourage bone formation and treat AVN of the femoral head, with encouraging results.

Further work is needed to determine whether one preparations of BMAC performs better than others with regard to bone formation, as each system will produce an end product that will vary in cell concentration, platelet number and haematocrit.

Along side well designed clinical trials, further basic science work is required to investigate the therapeutic action of BMAC. The commercial processing of BMAC needs to be optimized in order to achieve a consistent end product, which will provide predicable and translatable results. The future potential of cell characterization in order to determine the optimum cell for repair/regeneration of bone also needs to be explored.