Traditionally, autologous bone has been utilized as a foundation for arthrodesis [28]. Various forms of autologous bone available for use in spine procedures include: iliac crest bone graft; local bone harvested during a laminectomy, facetectomy, corpectomy, vertebrectomy, or osteotomy; and bone marrow aspirate. However, each method carries with it its own set of limitations in terms of ability to achieve fusion as well as complication profile.

In response to the aforementioned complications associated with iliac crest bone graft, a significant focus has been placed on the utility and effectiveness of allograft alternatives for successful lumbar fusion. Allograft products, bone harvested from cadaveric donor tissues, encompass extenders and/or substitutes to ICBC. These grafts serve primarily as an osteoconductive matrix with no self-supplied osteogenic or osteoinductive properties. The benefits of intraoperative allograft use include decreased operative time, reduced blood loss, alleviation of donor site morbidity, and elimination of the need to harvest autogenous bone. Allograft is available in three different forms: fresh-frozen, freeze-dried, and demineralized freeze-dried, each imparting differential structural strength [69].

Ceramics are matrices of inorganic, nonmetallic atoms held together by ionic and covalent bonds [85]. Given that the capacity of each biologic is dependent upon its structural, cellular, and biochemical properties, ceramics are prepared to mimic the mineral phase of the bone [86]. Ceramic materials used in spine surgery include calcium sulfate and calcium phosphate, particularly used in the setting of implant coatings and defect fillers. Bone mineral and ceramic matrices display similar crystal structure and molecular compositions as the bone and yield an osteoconductive surface for arthrodesis [87]. For example, Pro-Osteon and Interpore are two frequently employed hydroxyapatite biologics made by application of extreme heat to the calcium phosphate body of a coral, Porites astreoides, which was chosen for its pore size—comparable to the bone. There are several preparations of calcium phosphate and calcium sulfate that display different characteristics as bone graft extenders, with optimal remodeling matching the degradation and remodeling profile of the bone. However, ceramics lack the organic phase of the bone and are therefore brittle with low tensile strength and significantly higher modulus of elasticity than the bone.

Hydroxyapatite and tricalcium phosphate (TCP), the ceramic forms most frequently employed in medicine, are purely osteoconductive and are replaced by the host bone through a process of creeping substitution. They exist in the forms of powders, pellets, putty, and injectable cements. Hydroxyapatite is the most studied calcium phosphate material since the 1970s [88]. Hydroxyapatite directly bonds to the bone, allowing for osteoblast proliferation into its scaffold [89]. TCP has similar biocompatibility of hydroxyapatite formations and comparable tensile and compressive strength to the bone but dissolves more rapidly in situ.

Ceramic-based bone grafts have been widely used in spinal surgical procedures to reduce the complications associated with autograft. However, recent studies in the lumbar spine do not present clear support for its use [90]. For example, Sathira-Angkura et al. in 2011 reported “doubtful fusion” in 22 of 23 patients at a 2-year follow-up when hydroxyapatite was mixed with autogenous bone marrow in posterolateral lumbar fusion [91]. Similarly, Acharya et al. prematurely ended a study after 95 % of the hydroxyapatite group had poor consolidation of the graft after 1 year [92]. However, the study and control groups in examination of hydroxyapatite have generally been of poor quality. According to a meta-analysis by Kaiser et al. in 2014, ceramic bone grafts are demonstrably feasible graft extenders or substitutes [90].

Platelets are activated at sites of injury where they physically limit blood loss and promote generation of thrombin to coagulate blood [93]. Additionally, platelets are also involved in wound healing and aid repair of highly vascularized bone tissue by releasing growth factors that attract mesenchymal cells of the bone marrow [94]. Lowery et al. suggested in 1999 that the application of platelet-rich plasma results in higher bone density 6 months after lumbar spine fusion and that osteoblasts lining a cancellous bone surface survive transplantation and respond to platelet growth factors [95]. And, given usage of hemocomponents like hyper-concentrated platelets gels as a wound sealant, platelet gels have seen the use in combination with autologous bone during lumbar spine fusions [96]. Platelet gels, combinations of concentrated platelets with thrombin, have been used successfully as autologous fusion adjuncts in both animal and human models and are now being marketed to promote bony growth [93, 97–99].

Studies into the efficacy of platelet gels as bony fusion enhancers are limited. Carreon et al. reported a nonunion rate of 25 % in the platelet gel with ICBG group compared to 17 % in the control group with ICBG alone [97]. Castro et al. in 2004 detailed an increase in pre-anesthesia time of 18 min for obtaining the platelet gel and a 19 % lower arthrodesis rate in the platelet gel group [100]. These preliminary reports on platelet gel’s decreased efficacy as a growth factor adjunct have limited further research. To date, there have been no level one evidence studies on the effect of platelet gels in lumbar fusion.

Bone morphogenetic proteins (BMPs) are a family of soluble signaling factors in the transforming growth factor-β (TGF-β) superfamily of growth factors discovered by Marshal Urist in 1965. Several BMP molecules have been identified, though only certain forms demonstrate significant osteogenic properties, including BMP-2 [101–103]. Currently, recombinant human BMP-2 (rhBMP-2) and osteogenic protein-1 (OP-1) are available for commercial use. BMP-2 has been previously reported to induce bone and cartilage formation through osteoblastic differentiation of mesenchymal stem cells [104]. The original US FDA approval of BMP in 2002 was in the setting of anterior lumbar interbody fusion, though BMP-2 has since been implemented with varying success in posterolateral fusion and posterior or transforaminal interbody fusion [20, 105–107]. Currently, it is estimated that the off-label use of these agents exceeds 85 % in primary spine procedures [108]. The reported benefits of BMP centralize on achieving higher fusion rates and decrease donor site morbidity in comparison to autograft [109]. In the operative management of lumbar scoliosis, the use of BMP has proven to be advantageous as a suitable bone graft alternative for multilevel fusion, though reported adverse events may cause a reevaluation of BMP’s efficacy in these procedures. Substantial variation in the literature surrounding BMP’s use, integrated with unclear cost-effectiveness, requires continued evaluation.

BMP use in the surgical treatment of adult lumbar pathologies varies by surgeon preferences and specific pathology. Currently, the only FDA-approved use of BMP in spine surgery is in single-level anterior lumbar interbody fusion (ALIF) with interbody cage [20]. In 2002, Burkus et al., in a multicenter, prospective, randomized study, compared rhBMP-2 on absorbable collagen sponges versus autogenous iliac crest bone graft for interbody fusion in patients with degenerative lumbar disc disease to evaluate fusion progression at 6, 12, and 24 months postoperative. Radiographic fusion assessment was highest (94.5 %) in patients receiving rhBMP-2 compared to ICBG (88.7 %), though new bone formation was identified in all investigational patients. Moreover, the authors reported a 5.9 % rate of adverse events related to the iliac crest graft harvest and a 32 % graft site discomfort rate at 1-year post-op [20]. The summation of these findings was used to highlight the use of BMP as a viable alternative to ICBG in the lumbar spine, given new bone formation in all investigated patients. Multiple studies in subsequent years also underscored the effectiveness of BMP-2 in lumbar spinal fusion. In 2004, Haid et al. evaluated the use of BMP-2 on a collagen sponge carrier in single-level posterior lumbar interbody fusions (PLIF) against an ICBG control in a multicenter, prospective, randomized trial [110]. Clinical and radiographic outcomes were assessed in 6-month intervals through a 2-year follow-up. At 24 months, the authors observed a nonsignificant difference in fusion rates in favor of the investigational cohort (92.3 %) compared to the controls (77.8 %). The authors did not report any device-related adverse events compared to a 6.1 % rate in the control patients. Moreover, patients receiving rhBMP-2 reported superior improvement in Numeric Rating Scale Back Pain scores at 24-months postoperative, while 60 % of controls complained of donor site pain. Kim et al. reported on structural lumbar curve fusions with a posterior approach augmented with structural anterior interbody grafting. These authors used BMP as deemed necessary in concentrations ranging from 24 to 96 mg based on the number of levels fused. BMP was soaked on an absorbable collagen sponge and then wrapped around the cortical bone and placed onto the posterior elements [4]. In total, Kim et al. used rhBMP-2 in 12 cases in anterior column reconstructions and posterior spinal fusions with significantly more BMP use in patients with an anterior apical release.

The utilization of BMP in adult spinal deformity (ASD), where pseudarthrosis is a common postoperative major complication, is driven by different surgical and radiographic indications [24, 111, 112]. Interbody support though has shown consistent effectiveness in stabilizing long fusions for lumbar deformity. Surgeons are recently trending away from the standalone use of ICBG and using instead a combination of locally harvested autogenous bone graft and allograft to stimulate fusion. Crandall et al. evaluated the use of TLIF with rhBMP-2 among 509 patients, of which 123 were diagnosed with lumbar deformity (including adult idiopathic scoliosis and degenerative lumbar scoliosis) [113]. The arthrodesis rate was 98.4 %, and, of the eight patients that developed nonunions at TLIF levels, five were long fusions for deformity. These lumbar scoliosis patients had significantly lower preoperative visual analog scale (VAS) functional scores, though they also displayed significant improvements at a 2-year follow-up. Comparably, Maeda et al. reported on long fusions to the sacrum and found that of the 23 patients receiving rhBMP-2, only 1 (4.3 %) developed a pseudarthrosis in contrast to a rate of 28.1 % in the ICBG cohort. However, the BMP group was limited by a shorter follow-up interval—2.7 years versus. 4.9 years in the ICBG group [114]. As previous usage recommendations were established largely based on trials studying single- and double-level fusions, which represents 85 % of rhBMP-2 use, the role of BMP in the context of ASD and lumbar scoliosis surgery in particular is continually being refined [115–117]. For example, Bess et al. in 2014 found no increased risk of perioperative complications using BMP versus ICBG in long fusions for ASD [118]. Future research on BMP in ASD should focus on a dose effect and correlations with longer-term outcomes to provide meaningful recommendation for use.

Since its initial introduction, the use of BMP-2 in the lumbar spine has been associated with several adverse events and complications. Contraindications for BMP use include active malignancy, pregnancy, active infection at the operative site, and hypersensitivity, among others. At the forefront of discussion is the impact of BMP on the growth and invasiveness of malignancy, given BMP’s properties as a growth factor. Despite preclinical safety, data regarding BMP-2 effects on cancer cell proliferation failed to unveil any mutagenic associations, and high expression of BMP surface receptors have been observed in certain tumors [115, 119, 120]. Carragee et al. evaluated the risk of new malignancy in patients receiving a high dose (40 mg) of rhBMP-2 in a compression-resistant matrix in single-level posterolateral arthrodesis for degenerative lumbar spine conditions compared to autogenous bone control [121]. At 2-year follow-up, the author identified 15 distinct cancer events in the rhBMP-2 group with an incidence rate of 3.37 (95 % confidence interval, 1.89–5.56) compared to two cancer events in the control arm. This observed risk was sustained in a retrospective cohort study of Malham et al. of lumbar fusion (anterior, lateral, posterior, and posterolateral) with rhBMP-2 [122]. Twenty-seven of 527 patients were diagnosed with invasive cancer following treatment. Despite support in the literature, there remains as of yet no definitive or causative link between BMP use and tumorigenesis. Importantly, the Yale University Open Data Access (YODA) Project meta-analyses displayed no clinical advantage of BMP over bone graft, further confounding direct indications for BMP use.