Cell Cotransplantation Strategies for Vascularized Craniofacial Bone Tissue Engineering: A Systematic Review and Meta-Analysis of Preclinical In Vivo Studies

Inclusion and exclusion criteria

Search strategy, screening, and study selection

Electronic databases of MEDLINE (via PubMed) and EMBASE were searched for relevant English-language literature up to and including April 2016. Unpublished literature was searched through the Google and Google Scholar search engines. Bibliographies of the selected studies and relevant review articles were checked for cross-references. A specific search strategy was developed for MEDLINE (Table 1) and adapted for other databases. Titles and abstracts of the search-identified studies were screened by two authors (S.S. and A.S.) and full texts of all eligible studies were obtained. Uncertainty in the determination of eligibility was resolved by discussion with the other authors. Two authors (S.S. and A.S.) reviewed the selected full texts independently, and final inclusion was based on the aforementioned inclusion criteria. A summary of the screening process is presented in Figure 1.

OPEN IN VIEWER

FIG. 1.

Flowchart for study screening and selection. N, number of articles.

Data extraction

Data were extracted from full texts of included articles on author(s), study design, animal species, model characteristics, number of animals, number of procedures, inclusion criteria, outcome(s), method(s) of outcome evaluation, main findings, and conclusions. Descriptive summaries of studies were entered into tables. Quantitative data regarding NBF, that is, regenerated bone volume as a percentage, and vessel formation, that is, vessel numbers/density per unit area, were extracted from studies for possible meta-analyses. When data were presented as ratios or volumes (mm³), they were converted to percentages using total defect volume data. Standard errors of means, when reported, were converted to standard deviations for analysis. If data were only expressed graphically, numerical values were requested from the authors, and if a response was not received, digital ruler software was used to measure graphical data (ImageJ; National Institutes of Health, Bethesda, MD). When studies reported outcomes at multiple time points, data from similar time points of different studies were pooled for meta-analysis.

Quality assessment and risk of bias

Reporting quality assessment of all studies was performed based on a modification of the ARRIVE (Animal Research: Reporting In Vivo Experiments) guidelines regarding relevant items.^31,32 Compliance with the guidelines was evaluated using a predefined grading system applied to each of the 20 items ³³ (Supplementary Table S1; Supplementary Data are available online at www.liebertpub.com/teb). Reporting quality was judged as high, moderate, or low. Risk of bias (RoB) assessment was performed using a modification of the SYstematic Review Centre for Laboratory animal Experimentation (SYRCLE) RoB tool for animal studies ³⁴ and judged as high, low, or unclear ³⁵ (Supplementary Table S2). Any disagreement between the reviewers during study selection, data extraction, and quality assessment was resolved by discussion and consensus.

Meta-analysis

Meta-analysis was performed to compare the effectiveness of MSC/EC cotransplantation (experimental [Ex]) and MSC-only transplantation (control [Co]) using histomorphometric and micro-CT data (mean and standard deviation of %NBF and number of animals/defects [n] per group). Studies were pooled based on homogeneity regarding PICO and observation time. Subgroup analyses were performed at the level of animals and observation time(s), using the DerSimonian and Laird random effects model ³⁶ and STATA Statistical Software (StataCorp LP, College Station, TX). In one study, ³⁷ the design included multiple interventions per animal, that is, bilateral calvarial defects, for which a correlation coefficient (r) was calculated for between treatment-group measurements. ³⁸ An r-value of 0.5 was used for this study; for parallel group studies, the r-value was set at 0 and thus both study types were included in the same analysis. ³⁹ Pooled estimates of treatment effect (weighted mean differences [WMDs]) were calculated along with 95% confidence intervals (CIs), and the I ² statistic was used as a measure of inconsistency of results across studies. ³⁶

Search results and study characteristics

Of the 110 search-identified titles, 22 studies reporting quantitative histomorphometric and/or radiographic (CT/micro-CT) outcomes from 5 different species and 474 animals were included in the review (Table 2). A list of excluded studies along with reasons for exclusion is provided in Supplementary Table S5. Small animal models included rats (number of animals [n] = 356), mice (n = 77), and rabbits (n = 27). Large animal models included minipigs (n = 6) and dogs (n = 8; Table 2). Sample sizes in individual studies ranged from 2 to 64, and observation times ranged from 4 to 12 weeks. Fifteen studies used calvarial CSD models, while seven studies used maxillary (two studies) or mandibular CSDs (five studies). No studies included a positive control group treated with gold standard autogenous bone.

Quality assessment and RoB

Overall, the included studies were of unclear to high RoB and moderate reporting quality (Supplementary Tables S2 and S3). Most studies provided adequate information regarding title, abstract, introduction, study objectives, ethical approval, experimental design, and procedures. However, information regarding experimental animals and their housing and husbandry was generally inadequate. No studies provided information on sample size calculation or baseline characteristics of the animals. In seven studies, animals (or defects) were randomly allocated to treatment groups to minimize selection bias, although no details of the randomization procedure were reported. Four studies reported blinding of outcome assessors to treatment groups to minimize detection bias. All studies reported detailed outcome evaluation, including statistical analyses. Finally, information regarding study limitations and implications for translation to human models was limited. None of the studies reported implications for the 3Rs principle (replacement, refinement, and reduction) of animals in research.^31,40

Characteristics of cell cotransplantation strategies

Qualitative outcomes

Overall, minimal signs of inflammatory response, infection, or adverse reactions were observed across all studies, even after transplantation of allogeneic, xenogeneic, or human-derived cells. Detection of newly formed vessels was performed through immunohistochemical staining for endothelial-specific markers such as CD31, CD34, von Willebrand factor (vWF), VEGF, and so on. One study ⁴⁵ reported PCR analysis of retrieved tissues to identify endothelial marker genes (CD31, vWF, and vascular endothelial-cadherin). A general observation across studies was the detection of vascular structures consistently in areas of NBF. In four studies,^41,42,44,45 detection of human-specific endothelial markers in retrieved tissues confirmed the contribution of implanted ECs in vasculogenesis.

Meta-analyses

Meta-analysis was performed if there were three or more studies in each comparison group. Eleven studies reporting histomorphometric or micro-CT-based %NBF following MSC/EC cotransplantation versus MSC-only transplantation, in calvarial CSDs of rats and mice, were included in two separate meta-analyses (Fig. 2). A comparison was also made between cotransplantation and EC-only transplantation (Fig. 3). Finally, vessel regeneration was compared between cotransplantation and MSC-only groups (Fig. 4). To minimize any confounding influence of exogenous GFs, experimental groups receiving gene-modified cells and/or GFs were excluded. Since there were fewer than 10 studies in each meta-analysis, publication bias through funnel plots or statistical testing was not assessed due to the lack of power to distinguish chance from real asymmetry. ⁶¹

OPEN IN VIEWER

FIG. 2.

Forest plot for the comparison cotransplantation versus MSC-only transplantation for bone formation in rodent CSD models. The forest plot displays relative weight of the individual studies, the estimates of treatment effect (ES, effect size) expressed as WMDs, 95% CIs, and a predictive interval for the outcome %NBF. The diamond indicates the pooled estimate and its 95% CI. μCT, microcomputed tomography-based %NBF; CI, confidence interval; CSD, critical size defect; histo, histomorphometric new bone formation (%NBF); MSC, mesenchymal stem cell; wk, weeks; WMDs, weighted mean differences.

OPEN IN VIEWER

FIG. 3.

Forest plot for the comparison cotransplantation versus EC-only transplantation for bone formation in rat CSD models. The forest plot displays relative weight of the individual studies, the estimates of treatment effect (ES) expressed as WMDs, 95% CIs, and a predictive interval for the outcome %NBF. The diamond indicates the pooled estimate and its 95% CI. EC, endothelial cell.

OPEN IN VIEWER

FIG. 4.

Forest plot for the comparison cotransplantation versus MSC-only transplantation for vessel formation in rodent CSD models. The forest plot displays relative weight of the individual studies, the estimates of treatment effect (ES) expressed as WMDs, 95% CIs, and a predictive interval for the outcome vessel density per unit area. The diamond indicates the pooled estimate and its 95% CI.

The meta-analyses revealed a statistically significant effect in favor of cotransplantation versus MSC-only transplantation for radiographic bone regeneration in rat calvarial CSDs—pooled estimate (WMDs): 7.80% NBF (95% CI 1.39–14.21). Results for histomorphometric bone regeneration and vessel regeneration were inconclusive. Heterogeneity in most cases was very high (I ² > 80%, p < 0.05). Pooled WMDs with 95% CI and measures of heterogeneity for each of the subgroups are presented in Supplementary Table S4, along with an example for interpretation.

Animal models

Coculture: translational considerations

MSCs and EPCs represent the most favorable progenitor cell populations for osteogenic and vasculogenic differentiation within a coculture system. Bone marrow has traditionally been considered the gold standard source for MSCs and EPCs (a fraction of the hematopoietic stem cell population); however, the morbidity and procedural burden associated with marrow harvesting have led to identification of relatively less invasive sources of MSCs (adipose tissue, oral tissues, etc.) and EPCs (peripheral blood, umbilical cord blood, endothelial-induced MSCs, etc.). The ex vivo differentiation of MSCs into functional ECs by endothelial induction—as reported by six included studies—represents a particularly attractive strategy since both cell types are obtained from the same primary source, minimizing the need for additional tissue harvesting. ⁸⁹

Predifferentiation or induction of progenitor cells before implantation is considered to enhance their in vivo efficacy. ⁴⁷ The choice of culture media plays an important role in determining the optimal conditions for inducing osteogenic and endothelial differentiation of MSCs and EPC/ECs, respectively, without exerting negative effects on either cell type. ²² In the present review, coculture of MSC-EPC/ECs was performed for 2–14 days before implantation in standard culture medium (SM), osteogenic medium (OM), a 1:1 mixture of SM or OM and endothelial medium (EM), or sequential culture first in EM and then in OM. In one study, ⁵⁸ a transwell indirect coculture system was used, in which MSCs were exposed to the secretome of EPCs before implantation. Although all studies reported favorable in vivo outcomes, no studies correlated in vivo outcomes with different in vitro culture conditions. However, recent evidence suggests that optimal results may be achieved with coculture of MSC-ECs in purely OM in terms of both in vitro and in vivo bone formation and vascularization.^90–92

MSC expansion is commonly performed using basal culture media plus supplements to provide GFs, proteins, and enzymes to support cell growth. Fetal bovine serum (FBS) is commonly used to supplement culture media because the fetal milieu is enriched with growth factors and poor in antibodies. ⁹³ However, for clinical-grade therapy, it is important to substitute animal-derived products because MSCs can internalize xenogeneic proteins and thus there is a risk of infection and immunoreactions. ⁹³ In addition, there are concerns regarding sample-to-sample inconsistency and ethics of animal welfare. Alternatives to FBS include chemically defined media or human blood platelet-derived supplements (e.g., platelet lysate [PL]) to provide the necessary factors for cell growth and function. ⁹³ One study ⁴⁷ reported on the use of human PL for ASC and ASC/HUVEC cultures and that significantly improved in vitro mineralization (osteogenesis) in PL- versus FBS-supplemented cultures was observed. These results are consistent with reports of BMSC/HUVEC cocultures with PL supplementation. ⁹⁴

Only one study ³⁷ reported the use of a dynamic coculture system for rat BMSCs and ECs seeded on copolymer scaffolds, with formation of capillary-like structures in vitro after 1 week, suggesting a possible strategy for prevascularization. ⁹⁵ In context, a recent study ⁷⁹ in a large animal (sheep) model reported bone regeneration in femoral defects following dynamic coculture of autologous OBs and ECs in a perfusion bioreactor system. Bioreactor systems (spinner flasks, perfusion systems, etc.) are reported to enhance the proliferation, homogeneous distribution, and differentiation of MSCs in scaffolds compared with conventional, static two-dimensional (2D) culture by providing the appropriate inductive and mechanical cues in a controlled environment.^96,97 Additionally, bioreactors ensure more homogeneous distribution and perfusion of cells and culture media throughout the scaffold and can reportedly overcome the problems associated with nutrient gradients that develop in static cultures, that is, when cells nearer to the scaffold surface consume more nutrients than cells in the center, resulting in internal cell death. ⁹⁶

Scaffold materials chosen to deliver cells and/or GFs play a critical role in the performance of BTE constructs. Biocompatible and biodegradable scaffolds can support and facilitate the attachment, proliferation, and differentiation of progenitor cells, ingrowth of vascular structures, and mass transfer while also providing structural stability, space maintenance, and osteoconductivity at the regeneration site.^16,22 Commercial bone substitute materials, for example, β-TCP, HA, BBM, and so on, have been shown to be clinically effective scaffolds for MSCs. ⁷ However, alloplastic copolymer scaffolds have gained significant importance since each of their properties integral to performance (architecture, biodegradability, porosity, pore size, etc.) can be optimized during the fabrication process. ⁹⁸ Especially in cotransplantation strategies—where the presence of multiple cell types and complex simultaneous processes demand differential scaffold design, copolymer scaffolds provide several advantages. ⁹⁹ In the present review, a majority of studies reported the use of copolymer composite scaffolds to support MSC monocultures and MSC-EC cocultures with adequate in vitro efficacy and subsequently successful in vivo outcomes.

Efficacy of human-derived cells

The main translational implication of the in vitro findings is that cotransplantation of previously cocultured osteogenic and vasculogenic cells could enhance in vivo bone regeneration either by directly inducing osteogenic differentiation of MSCs or by enhancing neovascularization and cell survival within the constructs.^47,86 In the present review, significantly better in vitro cell survival, metabolic activity, and differentiation were observed in human-derived MSC and EC cocultures compared with MSC or EC monocultures. This is consistent with recent reports of the contribution of ECs to enhanced osteogenic differentiation of MSCs.^92,100 Cotransplantation of human BMSCs or ASCs with EPC/ECs (with or without prior coculture) in rodent calvarial CSDs yielded significantly greater bone formation in two studies,^41,45 while others^42–44 observed no differences between cotransplantation and MSC-only groups. In one study, ⁴⁷ human ASC monocultures demonstrated superior in vitro activity and in vivo bone formation than 1-week ASC-HUVEC cocultures on titanium mesh scaffolds. Therefore, the evidence for bone regeneration by cotransplantation of human MSCs and ECs is conflicting. Interestingly, one study, ¹⁰¹ excluded from the current review due to lack of quantitative outcomes, reported a codifferentiation strategy where human ASCs cultured in a combination of osteogenic and vasculogenic induction media acquired both phenotypes—positive for CD31 and osteopontin. These codifferentiated ASCs, when seeded on HA scaffolds and implanted into rat calvarial CSDs, resulted in superior (qualitative) bone and vessel formation than scaffold-only controls. ¹⁰¹

It is of relevance to consider the sources and types of vasculogenic cells used in a coculture/cotransplantation system. EPCs, under physiological conditions, are largely restricted to bone marrow and peripheral blood and consist of a heterogeneous population of cells.^102,103 Recent evidence suggests that EPCs may also be isolated from umbilical cord blood. ¹⁰⁴ ECs are an EPC subtype committed to EC differentiation. ¹⁰² In the included studies, relatively less variation in human EPC/EC sources was observed compared with osteogenic cell sources (Table 2). EPCs were derived from bone marrow, peripheral blood, or umbilical cord blood, while ECs were most often derived through endothelial differentiation of MSCs. The contribution of implanted human EPCs or ECs to vessel formation was confirmed by identification of human-specific endothelial markers in regenerating tissues.^41,42,44,45 Vessel counting was generally performed on stained histological sections of the central region of tissue explants and therefore represents vasculogenesis in the central part of the construct. This finding is of clinical relevance since a limitation of traditional BTE constructs is the lack of vascularization within the central regions.^16,22

Vascularization in BTE

Implanted EPC/ECs are reported to contribute to neovascularization through direct endothelial differentiation (vasculogenesis) and/or stimulation of the host regenerative response, for example, by release of trophic factors to recruit host EPC/ECs. ⁸⁶ Recent evidence suggests that exogenous EPCs are also involved in homing of MSCs to defect sites, ¹⁰⁵ while MSCs, in turn, may stimulate EPC differentiation through release of trophic factors. ¹⁰⁶ Moreover, a fraction of MSCs exposed to endothelial factors reportedly differentiate into pericytes, which line and stabilize newly forming vessels.^92,95 One study reported differentiation of human BMSCs into smooth muscle cells (SMCs) before coculture and cotransplantation with EPCs—the hypothesis was that SMCs would support EPC survival and stabilize the regenerating vessels, thus improving vascularization. ⁴⁴ However, no significant additional benefit of cotransplantation versus MSC-SMC-only or EPC-only transplantation was observed in terms of bone and vessel regeneration. ⁴⁴

Initiating prevascularization within BTE constructs before implantation is another emerging strategy. ¹⁰⁷ One included study ³⁷ reported the identification of nascent microvascular networks in vitro following 1-week dynamic coculture of rat BMSCs and ECs on copolymer scaffolds. Upon implantation into rat CSDs for 8 weeks, slightly higher amount of vessel formation was observed in the regenerating tissues of cocultures than MSC-only implants. ³⁷ In context, a recent study ⁷⁹ in a sheep model investigated bone regeneration in femoral defects implanted with prevascularized (MSC+EC) constructs dynamically cultured for 14 days in a perfusion bioreactor. Histological analysis of the prevascularized constructs at 16 weeks revealed microvessel-like structures containing red blood cells, suggesting a functional connection of primitive networks with host vasculature. ⁷⁹ The ability of primitive vessels in prevascularized constructs to anastomose with host vasculature has previously been demonstrated.^20,108–110

In the present review, one study ⁵⁹ reported the use of a radiopaque contrast agent to quantify vessel regeneration within rat calvarial CSDs through micro-CT. At 4 weeks, infiltration of the contrast agent revealed vessel formation and/or infiltration (from the periphery) throughout the defect—no significant differences in vessel quantity (%) were observed between the central and peripheral regions. ⁵⁹ Such novel imaging techniques, which can assist in tracking the in vivo behavior of primitive vessels in prevascularized constructs, should be utilized in future studies of vascularized BTE in orthotopic defect models.

The primary rationale for cell cotransplantation is that ECs can form primitive microvascular networks and improve cell survival in BTE constructs. ¹¹¹ Although the presence of ECs may enhance vascularization in CSDs, transplantation of solely ECs (vs. solely MSCs) does not appear to enhance bone formation; indeed, the presence of an osteogenic cell population within the construct is reportedly necessary for osteogenesis. ¹¹² In the present review, three studies^42,43,45 included a control group of solely human EC transplantation and observed significantly lower bone formation compared with MSC-only or MSC+EC groups, while one study ⁴⁴ reported no differences between groups. A similar trend was observed for vessel formation. However, in one study, ⁵⁴ transplantation of rat ASC-derived ECs yielded significantly greater bone and vessel formation than MSC-only or MSC+EC transplantation. Therefore, the combination of MSCs with ECs, at least for human-derived cells, appears to be the most optimal strategy for vascularized BTE than any one cell type alone.

Alternative strategies

Alternative strategies to enhance vascularization in BTE constructs have been proposed, of which local delivery of exogenous osteogenic and/or angiogenic GFs has been extensively documented. ¹⁶ GFs are molecular signals that coordinate wound healing and regeneration. Local delivery of exogenous (recombinant) or autologous GFs can direct cellular behavior by initiating specific osteogenic cell signaling cascades and mobilizing host progenitor cells (MSC, EPC) to the regeneration site.^113,114 Delivery of pooled (e.g., autologous platelet concentrates) or single GF (e.g., recombinant human bone morphogenetic protein-2 [rhBMP-2] or growth and differentiation factor-5 [rhGDF-5]) combined with biomaterial scaffolds has been shown to significantly enhance bone regeneration in clinical studies of contained or limited bone defects, for example, sinus augmentation, saddle defects, and extraction sockets.^114,115 However, for larger defects, combined delivery of osteogenic and angiogenic GFs may be more beneficial. ¹¹⁶ VEGF plays a critical role in angiogenesis during bone formation and is an ideal candidate for combination GF delivery strategies. ¹⁴ Recent preclinical studies have reported conflicting results of bone and vessel regeneration, following implantation of scaffolds loaded with BMP-2 or VEGF either alone or in combination, in calvarial CSD models.^117–119 In the present review, delivery of VEGF through β-TCP-PLGA scaffolds resulted in significantly lower %NBF than delivery of VEGFs with MSCs and/or EPCs in canine mandibular CSDs, ⁶⁰ suggesting that GF delivery may compliment, but not substitute, the regenerative action of implanted cells. Recent studies, however, have shown that by using an appropriate delivery mode, it could be possible to lower the dose and delivery of GFs and minimize adverse reactions.^120,121 Determination of the optimal dosage, delivery method, temporo-spatial release profile, and reliance on host cell responses are some of the reported limitations of current GF delivery strategies.^16,22

Another proposed strategy is the implantation of genetically modified MSCs encoding for angiogenic and/or osteogenic GFs. ¹²² MSCs are modified ex vivo, most commonly through adenoviral transduction, to express favorable GFs (e.g., BMP-2, VEGF) or transcription factors (e.g., HIF-1α) and have been shown to enhance bone regeneration in craniofacial CSDs compared with unmodified MSCs.^123–126 In the present review, one study ⁵² reported significantly greater bone and vessel formation in rat calvarial CSDs following coculture and cotransplantation of BMP-2-transduced MSCs and ECs compared with unmodified cocultures or unmodified MSCs after 5 weeks. However, additional sensitive in vitro procedures relating to GF gene transduction must be considered in these strategies.

Meta-analyses and heterogeneity

A random effects model was chosen for the present meta-analysis to account for the expected between-study variance. ¹²⁷ Heterogeneity (I ² ) was found to be very high within all categories in the meta-analyses. This could be due to biological factors related to the animals and/or methodological differences between the studies. Biological factors may include the animals' species, gender, age, immunological status, or the sources of MSCs and ECs used, and so on. In particular, sex and age of animals are reported to affect morphological, physiological, immunological, and behavioral parameters and thereby influence the outcomes of experiments. ¹²⁸ Methodological differences in the study design (e.g., sample sizes, randomization), nature of interventions (e.g., use of autologous vs. allogeneic or human cells, osteogenic induction of cells before implantation), and outcome evaluation (e.g., methods of bone histomorphometry or micro-CT) may have also contributed to heterogeneity.

Although histomorphometry is considered the gold standard method for evaluation of 2D bone structure, micro-CT is emerging as a reliable alternative for assessing 3D microarchitecture with high resolution and accuracy in a fast and nondestructive manner. Several studies have reported high correlation between micro-CT and histomorphometry.^129–131 However, to minimize heterogeneity in the present meta-analyses, results of micro-CT and histomorphometry studies were analyzed separately.

Quality of reporting

The reliability of results of meta-analyses directly depends on the quality of the primary studies. ¹²⁷ As assessed by compliance with the ARRIVE guidelines and SYRCLE RoB tool, the included studies were judged to be of moderate reporting quality, but an unclear to high RoB. The ARRIVE guidelines aim to improve the reporting quality of animal studies and have been widely used for assessment of preclinical research in implant dentistry, ³² while the SYRCLE tool addresses particular aspects of bias that play a role in animal experimental studies. ³⁴ Nevertheless, a clear need for more standardized reporting of animal studies was identified herein to allow reliable future reproduction and synthesis.

Clinical relevance

Although no clinical studies have reported direct coculture or cotransplantation of cells for bone regeneration in humans, one study ¹³² evaluated the efficacy of bone marrow aspirate containing populations of both hematopoietic (CD34⁺ HSC) and mesenchymal (CD44⁺, CD90⁺, CD105⁺ MSC) progenitor cells for mandibular reconstruction. Patients (n = 20 per group) received bone marrow grafts with either standard (54 ± 38 cells/mL) or high (1012 ± 752 cells/mL) concentrations of CD34⁺ HSCs and similar concentrations of MSCs (15 × 10⁶ cells) together with rhBMP-2 in an allograft scaffold. Significantly higher %NBF was observed in patients receiving higher concentrations of CD34⁺ HSCs (67% ± 13% vs. 36% ± 10%) after 6 months, suggesting an important and complimentary role of HSC to MSC in alveolar bone regeneration and the benefits of implanting heterogeneous cell populations at regeneration sites.^132,133

Due to the large inherent variation in animal studies, preclinical meta-analyses aim to disclose the direction rather than size of the effect of an intervention.^127,134 Fundamental limitations of animal models that must be considered when interpreting the findings are (1) the underestimation of clinical variation, with regard to both local (defect size, morphology, mass transfer, etc.) and systemic (age, comorbidities, etc.) aspects; and (2) the overestimation of clinical performance since the CSDs utilized are uniform surgically created defects with sound surrounding tissues and generally uncompromised blood supply and most often involve healthy animals, which is often not the case in clinical scenarios. ⁷⁵ Moreover, the dimensional differences between experimental defects in small animals and the clinical situation must be considered, especially with regard to vasculogenesis. Although it may be inferred that bone and vessel regeneration observed within calvarial CSDs in the included studies was largely due to the direct and/or indirect regenerative effects of exogenous cells functioning in a truly demanding environment (due to the highly cortical structure [low bone marrow] and poor local vascularity [absence of a primary nutrient artery] of the calvarium^28,135,136), the numerical values or magnitude of response from the present meta-analysis cannot be directly translated to the human situation. Thus, the ∼8% additional NBF observed in defects augmented with cotransplanted cells would not translate to 8% (or x-times 8%) more bone if the same procedure was performed in humans. ²⁷ Moreover, considering the magnitude of estimates of treatment effect, the clinical relevance (benefit) of cotransplantation procedures for oral and maxillofacial indications may be questioned.