Spine Interbody Implants: Material Selection and Modification, Functionalization and Bioactivation of Surfaces to Improve Osseointegration

Abstract

The clinical outcome of lumbar spinal fusion is correlated with achievement of bony fusion. Improving interbody implant bone on‐growth and in‐growth may enhance fusion, limiting pseudoarthrosis, stress shielding, subsidence and implant failure. Polyetheretherketone (PEEK) and titanium (Ti) are commonly selected for interbody spacer construction. Although these materials have desirable biocompatibility and mechanical properties, they require further modification to support osseointegration. Reports of extensive research on this topic are available in biomaterial‐centric published reports; however, there are few clinical studies concerning surface modification of interbody spinal implants. The current article focuses on surface modifications aimed at fostering osseointegration from a clinician's point of view. Surface modification of Ti by creating rougher surfaces, modifying its surface topography (macro and nano), physical and chemical treatment and creating a porous material with high interconnectivity can improve its osseointegrative potential and bioactivity. Coating the surface with osteoconductive materials like hydroxyapatite (HA) can improve osseointegration. Because PEEK spacers are relatively inert, creating a composite by adding Ti or osteoconductive materials like HA can improve osseointegration. In addition, PEEK may be coated with Ti, effectively bio‐activating the coating.

Introduction

The clinical outcome of lumbar spinal fusion is correlated with achievement of bony fusion1, 2. Traditionally, posterior stabilization and interbody screw fixation play a major role in limiting motion and load bearing in the immediate postoperative period. Over time, a fusion mass forms and load is transmitted through the newly formed bone. In the presence of pseudoarthrosis, this load is borne by the endplates and implant for the duration of the patient's life. Excluding femoral cortical ring allografts, the most commonly used spacers are metallic (made of titanium [Ti] and its alloys) or polymer (made of polyetheretherketone [PEEK]). Their function is predominantly mechanical: they require the addition of other materials to achieve bony fusion.

Although traditional spacers excel in the area of load‐carrying capacity, native, unprocessed surfaces are generally inert and have limited ability to bond and interlock with surrounding bone; hence, they offer little resistance to interfacial shearing forces3. These implant materials result in a high rate of fusion only when combined with other osteoconductive and osteoinductive agents4, 5, 6, 7, 8. Osteoconductive materials, such as bio‐ceramics, lack substantial load‐carrying capacity and toughness. While osteoinductive agents may also be considered when aiming to enhance fusion rates, there are significant risks that must be considered before use9, 10, 11, 12.

Achieving bone integration with an interbody implant is likely to aid fusion and improve implant longevity by limiting subsidence and stress shielding and associated complications. Surface modification and/or conversion of implant surfaces into bioactive areas is intended to improve in‐growth and on‐growth, thus conferring the previously mentioned benefits. A conceivable, yet extraordinary, result may be a material that is so effective at conducting bone growth that it is able to achieve early biological fixation and support osteoconduction between endplates, obviating the need for additional procedures and materials such as autograft, osteoconductive bone graft substitutes and osteoinductive materials13, 14, 15, 16.

Throughout the last two decades Ti or PEEK devices have dominated; however, in an unaltered state both these materials have limited bioactivity. For an orthopaedic implant to achieve bioactivity, its constituent materials must elicit a specific biological response at the interface of the material, thus facilitating formation of a bond between the tissues and the material17. This is often verified by in vivo tests or soaking in simulated body fluids and investigating surface precipitation of hydroxyapatite (HA)18, 19. Surface treatments are aimed at modifying the interaction with the body to generate a bioactive layer which aids in osseointegration. This in turn creates a strong implant–bone interface, resulting in structural, biochemical and functional stability. While there are several excellent review papers summarizing a wide breadth of biomaterials aspects of bone in‐growth20, 21, 22, and this topic has the potential to fill volumes, this review aims to summarize materials and modifications of those materials for interbody spacers that have recently become available to spine surgeons, or are currently under development. It focuses on clinical and biomaterials reports relating to enhancing the bioactivity of available interbody spacers.

Ti and Ti Alloys

Titanium and its alloys were introduced into the orthopaedic world in the 1940s and into the spine world a few decades later (Table 1). Following Stadelmann et al.'s discovery of the osseointegration phenomenon associated with Ti implants, exploration of their applications in dental and surgical fields began23. Ti is well suited to the spine because of its biocompatibility, robust repassivation that is attributable to TiO₂ formation, which provides excellent resistance to corrosion, and low density (approximately 4700 kg/m³)24 (Fig. 1).

Table 1.

Timeline of introduction of spinal interbody materials

Timeline	Introduction of spinal interbody implant material
1930	Autograft
1950	Tri‐corticate autograft
1960	Stainless steel
1980	Titanium and alloys
1990	Carbon‐fiber‐PEEK, silicon nitride
2000	Nitinol, tantalum, cobalt–chromium–molybdenum

Figure 1.

Elastic modulus of available spinal implant materials. CFR, carbon fiber‐reinforced.

A potential disadvantage of Ti is the mismatch in elastic modulus (110 GPa) compared with that of bone (10–30 GPa)21, potentially leading to stress shielding around the implant and, together with local inflammation, causing bone atrophy, subsidence and implant failure25, 26, 27. Another disadvantage, not related to bone integration, is difficulty in confirming fusion status because of its high radiodensity. Although Ti spacers in a raw state are capable of supporting bone growth, they are not sufficient for strong implant–bone integration. Various modifications are available to improve this.

Surface Modification of Ti and Its Alloys

Titanium can be modified to improve both on‐growth and in‐growth. On‐growth of bone is the direct apposition of bone to the surface, whereas ingrowth involves the interlocking or bone growth “into” the surface of a material, requiring a three‐dimensional (3D) structure with pores open to the outside. These modifications are aimed at influencing the way tissues incorporate with the implant material (Table 2).

Table 2.

Bioactive treatments available for Ti and PEEK

Implant material	Treatment to convert into bioactive material
Titanium	Rough surface Modification of surface topography Heat treatment Alkali treatment Removal of Na ions Porous material conversion HA coating
PEEK	Ti composite HA composite Calcium silicate composite Bioglass composite β‐TCP composite

Cells in contact with a material will first attach, adhere and then spread. Much of the body's reaction to implants is dictated by the proteins that adsorb during initial contact with bodily fluids. This attachment is preferential and influenced by a very wide range of factors28. Some of these, such as surface roughness, can be manipulated to influence a desired protein family, further influencing the types of cells that attach. This effect occurs with metallic surfaces29, PEEK30 and HA31. It should be noted that surface roughness also influences initial fixation simply by increasing friction32, 33, aiding in initial fixation and limiting micromotion. This is important because micromotion influences the type of tissue that can form against a surface. Findings vary between different studies; however, motion of up to 50 μm is generally reported as supportive of bone apposition34, 35, 36, 37. A range of 40–150 μm is associated predominantly with fibrous integration34, 35, 36, 37, suggesting that the range of 40–50 μm indicates a “cutoff” above which bone on‐/in‐growth is unlikely.

The creation of micro scale roughness on Ti increases the amount of total protein and alkaline phosphatase in cell culture, indicating support for osteogenic cell differentiation27. Such architecture can be created by various methods including plasma spraying (Fig. 2) and electron beam melting techniques38. On a smaller scale, roughening the surface by nano‐modification, with surface features of tens of nanometers, mimics the architecture of natural tissues39. Such surface modification and patterning result in increased bone cell adhesion, growth and differentiation, as well as phenotypic maturation39, 40. Stand‐alone plasma‐sprayed Ti implants without additional bone graft material (Fig. 2) have been used in cervical41 and lumbar fusion42, successful fusion having been assessed indirectly by absence of segmental motion on dynamic X‐rays and by presence of bone bridges surrounding the cages43.

Figure 2.

Spinal implants with plasma‐sprayed titanium surfaces. (A) Cervical interbody implant. (B) Anterior lumbar interbody fusion implant. (C) Transforaminal lumbar interbody fusion implant. (D) Posterior lumbar interbody fusion implant (Melsungen, Germany).

Thermal and chemical processing can also be applied to Ti implants. In a leporine mode, Nishiguchi et al. demonstrated surface modification by heat and alkali allows a direct Ti‐bone bond44. Further treatment with removal of sodium (Na) ions enhances this phenomenon45.

Porosity

The high porosity (60%) and high interconnectivity of Ti (Fig. 3) allow bone in‐growth and tissue differentiation46, 47. As mentioned above, further modification by combined chemical and thermal treatment has resulted in improved preclinical fusion rates43. Its porous structure results in a bulk modulus of 4.2 GPa, which is similar to that of native or “neat” PEEK. The porous structure and rigid frame can survive a cyclic load of 10,000 N for 1,000,000 cycles. This relates to a predicted interbody cage loading requirement of at least 5000 N48. This type of implant has been evaluated in a prospective human trial of five patients with unstable lumbar spine disease treated with transforaminal lumbar interbody fusion (TLIF)3. The mean duration of follow up was 15.2 months; bony fusion was seen in 4/5 cases by 3 months and 5/5 cases by 6 months. The authors noted low elastic modulus (less chance of subsidence) and osseointegration due to surface modification as advantageous and contributors to the clinical success rate.

Figure 3.

Porous titanium lumbar interbody implant (Kasios, L'union France). (A and B) TLIF implants. (C) Scanning electron microscopy (SEM) at a magnification of 50. (D) SEM at a magnification of 500. TLIF, transforaminal lumbar interbody fusion.

HA Coating

It is well documented that bone shows a strong affinity to implants made of sintered HA49, 50. Although HA interbody spacers have been used in the past, the mechanical properties of HA alone are not well suited to this use51. Although a compressive strength of 600–700 MPa and tensile strength of 200–250 MPa can be achieved, resistance to fatigue failure is very low. It is possible to use techniques such as plasma spraying to bring HA and Ti together at high temperatures to form a chemical bond. De Groot et al. have shown that HA‐coated Ti (50 μm) provides the best of both worlds, having the mechanical strength of Ti and the bone bonding properties of sintered HA52. Although such micro scale coatings are effective, the use of HA nanoparticles suggests that early osseointegration may be more sensitively regulated by nanoscale modification46.

These coating have demonstrated considerable improvements in bone apposition and stability for pedicle screws. In an in vivo study in dogs, Hasegawa et al. found that HA coating of Ti pedicles increased bone apposition as early as 10 days post‐operation and increased resistance to pull out forces47. In a randomized clinical study, Sandén et al. found higher extraction torque for HA‐coated stainless steel pedicle screws and fewer radiolucent zones around screws53. Although HA coating has shown to be beneficial to pedicle screw removal forces, the same phenomenon has not yet been shown for interbody spacers.

PEEK

PEEK spinal cages were introduced by AcroMed in 1990s (now DePuy Spine, Raynham, MA). Carl McMillin was the polymer engineer who recommended PEEK and initiated the commercial success of this cage, which came to be known as the Brantigan cage54. It is biocompatible21 and, when constructed of neat PEEK, the elastic modulus is less than that of cortical bone. Addition of carbon fiber reinforcement brings the modulus closer to that of cortical bone24. By manipulating the quantity and direction of the fiber element, the desired material properties can be achieved55. This provides the potential to minimize stress shielding as compared with solid Ti implants; however, when compliance is engineered into the design this becomes less of an issue. In a study comparing PEEK and titanium ALIF implants in sheep, Pelletier et al. showed no difference in initial biomechanics, mechanical properties or fusion rates when similar amounts of bone graft were used56. This is analogous to the manipulation of carbon fiber in PEEK implants to more closely mimic bone. Another advantage of PEEK implants is radiolucency, which permits easy assessment of radiographic fusion. There is some debate as to the interaction between PEEK and osteoblasts. Sagomonyants et al. have shown that PEEK and rough Ti have comparable in vitro bone forming capacity30, whereas Olivares‐Navarrete et al. found that osteoblasts differentiate to a lesser degree when cultured on PEEK versus Ti surfaces, suggesting that the former has a lower level of support for osteogenic tissues40. This is likely due to differences in chemistry and surface energy, which are important factors for protein adsorption for both polymeric and metallic surfaces57, 58, 59. This lower level of bone integration has been demonstrated in both animal models56 and clinically. Figure 4 shows an example of this, the lucency at the bone–implant surface indicating a lack of implant integration. As with Ti, PEEK surface modification is possible. Schroeder et al. have shown that plasma processing can functionalize the surface of PEEK and enhance osteoblast cell adhesion, as it does with Ti60.

Figure 4.

Radiolucency at PEEK implant‐bone fusion interface persisting even at 12 months.

HA–PEEK Composites and Coatings

Because natural bone is a composite of fine HA reinforced by collagen, bioactive particles capable of supporting bone growth combined with a polymer would be analogous to this environment61. HA was originally added to PEEK with the intent of more closely mimicing bone62. More recently, Wong et al. developed strontium‐containing HA (15%–30% vol)/PEEK composite with a similar bending modulus to that of cortical bone (9.6–10.6 GPa)63. This composite shows enhanced bioactivity in vitro. The addition of HA whiskers has also been described64, 65. Other potential composites with PEEK include calcium silicate, bioglass and β‐tricalcium phosphate (β‐TCP)61, 62, 63. Coating of PEEK with HA has also been attempted. An in vivo study on femurs of rabbits found nanocrystalline HA‐coated PEEK surfaces showed better osseointegration than uncoated implants66. This study showed that nano HA‐coated PEEK had enhanced bone to implant contact and more bone area than uncoated PEEK. Published reports of preclinical and clinical trials using interbody implants manufactured of these materials are still lacking.

Ti‐PEEK Composites and Coatings

The addition of Ti modifies not only the mechanical properties of the composite but also the potential for osseointegration. Wu et al. fabricated an n‐TiO2/PEEK composite that enhanced cell attachment compared to PEEK alone. In vivo comparison using microcomputed tomography and histological analysis demonstrated twice the bone volume/tissue volume with n‐TiO2/PEEK compared with neat PEEK67. Raines et al. created a PEEK–Ti composite with Ti endplate inserts and a PEEK central portion (Fig. 5) that showed increased expression of all measured growth factors involved in bone formation and maturation. Figure 6 shows a PEEK implant with titanium endplates to augment bone‐implant fusion. More recently, Han et al. have used electron beam deposition to apply a coating of Ti to PEEK, producing improved cell proliferation as well as greater bone contact following implantation68. Other authors have described the application of plasma‐sprayed HA on top of Ti‐coated PEEK69. This demonstrates that treatments previously only applicable to Ti are also applicable to Ti‐coated PEEK, such treatments include NaOH bioactivation of Ti surfaces70.

Figure 5.

Ti composite interbody implant in which the bone‐implant surface is Ti whereas the core of the implant is PEEK (A‐Spine ASIA, Taipei, Taiwan). (A) Cervical interbody implant. (B) TLIF implant. (C) Scanning electron microscopy (SEM) of the junction of PEEK and Ti at a magnification of 50. (D) SEM of the Ti surface at a magnification of 500.

Figure 6.

PEEK with Ti endplate composite cervical interbody implant: The CONSTRUX Mini PTC System (Orthofix, Lewisville, TX, USA) is composed of two endplates made of 3D porous Ti scaffold (gray) with a PEEK interior core (tan).Photographs courtesy Orthofix.

Although bioactive treatments have potential advantages (Table 3) many are not readily available for clinical application because of difficulty in manufacturing, cost, altered physical properties or because they are still under development (Table 4).

Table 3.

Potential advantages of bioactive conversion of spinal interbody materials

	Advantages of bioactive conversion of implants	Comment
●	Rapid osseointegration of implant	Prevents micromotion, stress shielding, subsidence and implant failure
●	Avoids autologous graft harvest	Avoids: Graft site pain Limited availability of anatomical sites Increased operative time Blood loss Additional incision
●	Avoids allograft	Avoids: Potential transmission of disease Immune responses
●	Avoids bone morphogenetic protein	Avoids complications of: Osteolysis, subsidence and implant failure Massive bleeding Soft tissue swelling Infection Retrograde ejaculation Radiculitis Ectopic bone formation
●	Reduces operating time	Because there is no need for autologous bone harvest or preparation of implant with bone graft substitutes and growth factors

Table 4.

Disadvantages of bioactive conversion of spinal interbody materials

	Disadvantages of bioactive conversion	Comment
1	Alters mechanical property	The native implant material may lose some mechanical properties
2	Manufacturing issues	These conversions can be difficult to manufacture both in terms of cost and process
3	Biochemical issues	Conversions may result in potentially harmful materials

Silicon Nitride

Silicon nitride is a non‐oxide ceramic with osteoconductive properties similar to those of porous Ti, high mechanical and wear properties, partial radiolucency and high fracture resistance25, 71, 72. Silicon nitride is also supposed to have anti‐infective properties73, 74. Although implants made of this material are currently available, there are no published reports concerning their performance as interbody cages.

Tantalum

Tantalum is a transition metal with high compressive strength, Young's modulus in the range of that of cancellous bone and biocompatibility75. Although a less traditional material, porous tantalum (Fig. 7) has demonstrated good osseointegration; pre‐treatment with alkali and heat reportedly further improves bone bonding76. Sinclair et al. demonstrated in goats that porous tantalum cervical interbody cages show significantly greater volumes of bone at the bone–implant interface; in addition, more animals with tantalum implants had connection between autograft bone and both vertebrae than did animals with PEEK implants77. When evaluated at an early three month time point, porous tantalum has been shown to be similar to carbon fiber‐reinforced PEEK in pigs78. Stand‐alone trabecular tantalum cervical implants (without graft) and tricortical autograft with plating have been compared in a randomized human trial75. The authors found 82.1% radiographic fusion at 6 months and 89.3% at 12 and 24 months in trabecular tantalum patients as compared with 78.7% and 84.8%, respectively, for tricortical autografts.

Figure 7.

Trabecular tantalum implants with porosity up to 80% and average pore size of 550 μm. (A) Cervical interbody implant. (B and C) Posterior lumbar interbody fusion implants. (D) Porous architecture of trabecular metal tantalum. (E) Porous architecture of bone. Images courtesy Zimmer Spine, Minneapolis, MN, USA.

This discussion would be incomplete without mention of other potential implant materials. Nitinol is an alloy containing 50% nickel and 50% Ti with shape memory and superelastic properties: these may aid in application of constant load irrespective of patient position as well as preservation of some degree of motion79. Bio‐absorbable cages may have the theoretical advantages of low stress shielding, avoidance of artefacts on scans and no long term toxicity80 (Table 1).

Conclusion

Various modifications for improving the osseointegration of native materials used for spinal interbody spacers are available. These are predominantly under investigation by ongoing biomaterial studies and in vitro and in vivo animal studies. Surface modification of Ti to improve its bioactivity can be achieved by modification of the surface topography, physical and chemical treatment and creating a porous material with high interconnectivity. Coating the surface with osteoconductive materials like HA can improve osseointegration. Creating a composite by adding Ti or other known osteoconductive materials such as HA to PEEK can improve osseointegration. This can be achieved by bulk incorporation or coating.