History of Nucleus Replacement

Early attempts at nuclear replacements consisted of injecting polymethylmethacrylate^18,19 or self-curing silicone¹⁹ into the disc space after a nucleotomy. Clinical results were found to be similar to discectomy controls. Flow of the injected material and curing of the injected liquids were difficult to control. These variables led to abandonment of these procedures, but these initial attempts formed the theoretical basis of present-day preformed nuclear replacement systems.

Fernstrom²⁰ was the first to use stainless steel balls as nuclear replacement by inserting them into the central cavity created after a discectomy. He implanted these balls in 125 patients at 191 lumbar levels. In 1966, Fernstrom²⁰ reported that the outcome after this kind of nuclear replacement was better than discectomy alone and was similar to fusion. This procedure was largely abandoned owing to the subsidence of steel balls into the vertebral end plates. In 1995, McKenzie²¹ reported long-term outcomes of the Fernstrom ball. At 17 years of follow-up of 67 patients implanted with Fernstrom balls, he reported an 83% success rate in patients with one or more disc protrusions and a 75% success rate in patients with DDD. Although the Fernstrom ball never found mainstream acceptance as a nuclear replacement device, it did suggest that nuclear replacement can be a viable alternative to either a stand-alone discectomy or a fusion.

In 1981, Edeland²² suggested a design for an elastodynamic disc replacement device that would mimic the native disc biomechanically and biologically. Such a device would allow influx or efflux of water molecules in response to the load applied to it. Finding an elastodynamic material remained the most limiting factor. Attempts at finding such material in the 1980s and 1990s focused on either silicone polymers or a hydrogel nucleus composed of polyvinyl alcohol that was used to make soft contact lenses. Hou et al.²³ designed a solid preformed silicone rubber horseshoe-shaped implant termed the lumbar intervertebral replacement prosthesis. Biomechanical cadaver studies and animal studies in monkeys showed restoration of the disc height and mechanical stability. Human studies with this implant have not been reported.

In 1991, Bao and Hingham patented a hydrogel nucleus made of polyvinyl alcohol that imbibed and extruded water and could replicate the biomechanical viscoelastic properties of the native nucleus. An animal study by Allen et al.²⁴ showed no evidence of systemic or local toxicity but a high incidence of implant extrusion. Despite this setback, proven preclinical benefits with hydrogel acted as a springboard for other nucleus replacement technologies. With its first clinical use in 1996, the prosthetic disc nucleus (PDN) is the most well-known nucleus replacement device made of hydrogel.²⁵ The device is made of hydrogel enclosed in a polyethylene terephthalate (PET) jacket to restrict expansion on hydration. Multiple design and material changes have followed after its first use (Fig. 163-1).

FIGURE 163-1 NeuDisc preformed hydrogel nucleus replacement. Before (left) and after (right) hydration images are shown.

(Courtesy of Dr. James J. Yue.)

Current Nuclear Replacement Technology

Current nuclear replacement devices can be classified as either elastomeric or nonelastomeric. Elastomers can be subdivided into hydrogels and nonhydrogels that are either in situ cured or preformed. Preformed devices may be implanted in their final shapes or may take their final shape on expanding by hydration. Nonelastomers can be subdivided into articulating or nonarticulating based on the presence or absence of articulation as part of the implant design.

At the present time, there are several designs with active clinical or developmental activities. Most clinical activity is occurring outside of the United States. Such devices are elastomers such as Hydraflex, DASCOR, NuCore, NeuDisc, BioDisc, SINUX ANR and SINUX PNR, and Newcleus.

HydraFlex (Raymedica, Bloomington, IN) is the latest version of the PDN; it is inserted in a dehydrated form that rehydrates on implantation (Fig. 163-2). The hydrogel component is a block copolymer of polyacrylamide-polyacrylonitrile (Hypan, Lipo Technologies, Vandalia, OH) formed by solution casting. The jacket is formed of a dense woven circular tube of PET, which is a thermoplastic polymer. The PET jacket is inelastic but flexible, which helps the hydrogel core to maintain its shape when subject to compressive forces. The hydrogel core absorbs water on implantation. Hypan80, which is currently being used, absorbs 80% of its weight by water, which makes it softer and flexible. HydraFlex takes approximately 32 hours to reach hydration equilibrium.

FIGURE 163-2 PDN replacement.

(From Berlemann U, Schwarzenbach O: An injectable nucleus replacement as an adjunct to microdiscectomy: 2 year follow-up in a pilot clinical study. Eur Spine J 18:1706–1712, 2009.)

DASCOR (Disc Dynamics, Eden Prairie, MN) is a two-part in situ curable polyurethane. It consists of a polyether polyurethane core encapsulated and adhered to a polycarbonate polyurethane balloon, which has cavity expansion and conforming capabilities. The device is fabricated by mixing a two-part, liquid, pre–polyether polyurethane reactive system and injecting a liquid mixture in a polycarbonate polyurethane balloon delivery catheter placed within the nucleus cavity (Fig. 163-3). A custom electromechanical injection system is used to apply a computer-controlled pressure profile designed to deliver the liquid polymer.²⁶

FIGURE 163-3 DASCOR nuclear replacement.

Implant placement (A) and deployment (B). (From Ahrens M, Tsantrizos A, Donkersloot P, et al: Nucleus replacement with the DASCOR disc arthroplasty device: interim two-year efficacy and safety results from two prospective, non-randomized multicenter European studies. Spine [Phila Pa 1976] 34:1376–1384, 2009.)

NuCore Injectable Disc Nucleus (Spine Wave, Shelton, CT) is an in situ curing protein polymer hydrogel that mimics the properties of the natural nucleus. The polymer chain is composed of silk and elastin components designed for elasticity and toughness. The polymer (P27K) is mixed with a cross-linking agent at the time of implantation and is injected as a fluid through the anular defect where it adheres to the surrounding intradiscal tissue as it cures. There is no measurable temperature increase produced during the cure process. NuCore has been studied as an adjunct to discectomy²⁷ and is in the process of being studied for treatment of DDD (Fig. 163-4).

FIGURE 163-4 NuCore Injectable Disc Nucleus (red), shown interdigitating with normal disc after injection into nucleotomy defect.

(Used with permission of Spine Wave, Shelton, CT.)

The NeuDisc device (Replication Medical, Cranbury, NJ) is composed of a material that is a hydrolyzed polyacrylonitrile hydrogel (Aquacryl). In the absence of mechanical restrictions, the hydrogel imbibes 90% volume as a liquid. When implanted in a dehydrated state, this hydrogel implant is substantially smaller than the volume of resected nucleus and is easily placed though an incision in the anulus. After hydration, it expands anisotropically, in the axial direction principally, and becomes substantially larger than the incision. The axial expansion is not constrained by a jacket and results from its structure of polymer layers tied together with an internal substructure²⁸ (Fig. 163-5; see also Fig. 163-1).

FIGURE 163-5 NeuDisc device shown clinically (preoperative and postoperative) (A) and on MRI (B).

(From Bertagnoli R, Karg A, Voigt S: Lumbar partial disc replacement. Orthop Clin North Am 36:341–347, 2005.)

BioDisc is based on the CryoLife (Kennesaw, GA) BioGlue surgical adhesive with a formulation for spinal application. It comprises an in vivo mixture of serum albumin with gluteraldehyde as a cross-linking agent. Both substances mix and completely solidify within 2 minutes. SINUX ANR (DePuy Spine, Raynham, MA) is a liquid polymethylsiloxone (silicone) injectable polymer that cures within 15 minutes. PNR is an inflatable silicone device between threaded studs. The device is threaded into the vertebral bodies and inflated with silicone. It is implanted through the presacral approach and is the only nucleus replacement that is fixed to the vertebral end plates. Newcleus (Zimmer Spine, Warsaw, IN) is a memory coiling polycarbonate urethane that is coiled within the disc space. It absorbs 3% to 5% of its weight by water but does not have osmotic capabilities.

Among nonelastomeric devices, Regain (Biomet, Warsaw, IN) is a modified form of the Fernstrom ball with a larger radius for larger end-plate contact surface that is manufactured from pyrolytic carbon. NUBAC (Pioneer Surgical Technology, Marquette, MI) is the only articulating nonelastomeric implant and is made of PEEK-OPTIMA (Invibio, Greenville, NC), which is a proven biostable and biocompatible polymer widely used for spinal fusion and shown to generate minimal wear. NUBAC comprises two plates with a large contact surface area in an anatomic shape similar to the native nucleus with an inner ball-and-socket articulation²⁹ (Fig. 163-6).

FIGURE 163-6 NUBAC two-piece mechanical nucleus. This is an articulated polyetheretherketone (PEEK)-on-PEEK device.

(Used with permission from Pioneer Surgical, Marquette, MI.)

Indications for Nucleus Replacement

The indications and contraindications for nucleus replacement are largely determined by the intended objectives and the unique design of the nucleus replacement devices. Although the indications and contraindications for the initial clinical studies were determined a priori, a better idea of these criteria cannot be established until at least limited clinical follow-ups are performed. The criteria will then be based on the outcomes of such studies and the risk-to-benefit and cost-to-benefit ratios.

The primary indication for nucleus replacement is single-level discogenic pain caused by DDD with or without leg pain. Disc degeneration should be mild to moderate, and so disc height of 5 mm or more, or at least 50% of the adjacent level, is considered necessary. In the case of late DDD, the anulus may become biomechanically incompetent on distraction owing to secondary structural changes.

Nucleus replacement devices rely on the anulus and the end plates for biomechanical restraint and load sharing. It may be difficult for the nucleus replacement device to correct spinal deformities such as spondylolisthesis, scoliosis, or kyphosis because of morphologic and structural changes in the anulus that can lead to inappropriate load sharing and increased extrusion risk. Similarly, severe end-plate degeneration and the presence of Schmorl nodes would predispose to implant subsidence. End-plate geometry may have a further impact on clinical outcomes: Hyperconcave end plates pose the risk of scoliosis, whereas truly convex end plates pose the risk of implant extrusion.

Surgical Technique

The nucleus replacement device can be implanted via either a retroperitoneal approach or a posterior approach. The retroperitoneal approach is used more commonly. Although this approach is similar to the approach used for anterior fusion or a disc replacement, the amount of anular exposure required for nucleus replacement is less because of the smaller dimensions of the nucleus replacement implants. Also, because the size of the nucleus replacement device is smaller than a disc replacement prosthesis, it is possible to use the posterior or a direct lateral approach for placement of the device.

Careful surgical planning is required for effective placement of the device. Some general but critical surgical pearls are important. The size of the anulotomy window differs based on the approach and the size of the device to be inserted. The anulotomy window needs to be large enough for removal of the nucleus but small enough to reduce the risk of device extrusion.

With a dorsal approach, the anulus is thin or deficient if there is a disc herniation, so the approach should be from the side of disc herniation. Laminectomy and facetectomy should be done as required while trying to preserve as much of the posterior elements as possible. The anular defect caused by the disc herniation can be used, or a small anular slit or a box cut can be made if the anulus is intact. In the case of a larger implant, either a larger box cut is required or the window can be dilated to receive the implant. When performing a ventrolateral or direct lateral approach, because the anular wall is thick, a box cut or a circular cut by a trephine is necessary for nucleotomy. A ventrolateral anulotomy is usually better than a direct ventral anulotomy to reduce the risk of implant extrusion. The initial cut can be either dilated or enlarged to device dimension. Alternatively, an anular flap can be made for device insertion and can be sutured back later. A direct lateral approach is becoming more popular among spine surgeons and may become a preferred approach in the future for nucleus replacement.

The nucleotomy required for nucleus replacement involves removal of the entire nucleus to create a symmetrical cavity lined by the anulus and the end plates. The implant should be positioned in the center of this cavity. It is crucial to preserve the integrity of the end plates and the anulus.

Outcomes

Outcomes data for nucleus replacement technology are limited by the lack of large randomized prospective studies. Most of these devices have been used in limited Investigational Device Exemptions studies with short-term follow-up reported. None of these devices are approved for use in the United States by the Food and Drug Administration (FDA).

To date, the most widely studied device with the longest history is the PDN. The PDN has shown favorable biologic compatibility and biomechanical stability. Fatigue testing performed with loads of 200 N to 600 N for 50 million cycles showed that the device continues to function with maintained disc height and structural integrity. It has been shown not to form polyethylene wear debris.³⁰ Eysel et al.³¹ analyzed the effect of the PDN on restoration of segmental mobility after a nucleotomy in human cadaver spines. These investigators found that there was a statistically significant (P < .05) increase in segmental mobility in all directions after nucleotomy with an increased mobility of the segment of 38% to 100%. After the introduction of two PDN implants, there was a restoration of segmental mobility for all movement directions with no statistically significant difference compared with the intact segment before nucleotomy.

In clinical studies, the major concerns with the PDN device have been the risk of implant migration and the potential for end-plate changes and subsidence. Schonmayr et al.³² reported early clinical results with the PDN implant with a minimum 2-year follow-up in 10 patients with symptomatic lumbar DDD. They showed good functional outcomes with restoration of disc height and motion. However, three implants migrated, and one patient required a revision surgery for migration.

Multiple phases of implant trials following these early results involved changes made to the device shapes and to the surgical protocol to facilitate implantation and eliminate the high device migration rates. After these modifications, the success rate for the device improved significantly. Clinical data show excellent results with marked improvements in Oswestry Disability Index (ODI) and Prolo scores with disc height measurements being well maintained within normal physiologic ranges. Klara and Ray²⁵ reported a cumulative 90% success rate with 10% of the patients requiring explantation. Following these early results, Bertagnoli and Vazquez³³ described implanting the PDN device through an anterolateral transpsoatic approach to reduce the risk of implant migration. They reported the results in eight patients after using this technique. Four of the first five patients developed a transient psoas neurapraxia. Ventral migration of the PDN was detected in three patients and was thought to be due to preexisting defects in the anterior anulus unrelated to surgery. Because the patients were asymptomatic, no revision was attempted. One patient required a revision surgery with fusion for implant subsidence. Of the remaining seven patients, significant improvements were noted in the mean ODI and Prolo scores at 6 months and 12 months with maintained disc heights.

In earlier studies, two PDN implants were used per disc level. It was suggested that using two implants per level may lead to overstuffing of the disc space and increased risk of implant migration. Manufacturers have advocated the use of a single larger implant sufficient to occupy the cavity after a nucleotomy. In 2002, 45 patients were implanted with a single PDN implant for lumbar disc herniation. Follow-up was reported by Jin et al.³⁴ for 33 patients at 6 months. Statistically significant improvements were noted in the ODI and Prolo scores with significant improvements in spinal mobility and disc heights. There were no cases of implant migration or failure in 6 months. Although Jin et al.³⁴ did not report any end-plate changes and subsidence at 6 months, these have been reported on longer follow-up after PDN implantation.

Shim et al.³⁵ followed 46 patients after PDN implantation for more than 6 months. Four patients (8.7%) showed device extrusion that needed revision surgery, and one patient had an infection. The investigators reported a 9.4% increase in the disc height at 1-year follow-up; 19.6% of patients showed some subsidence, 60.9% of patients showed end-plate sclerosis, and 82.8% of patients showed aggravation of Modic changes on MRI. Clinical results were excellent in 5 patients (10.9%), good in 31 (67.4%), fair in 3 (6.5%), and poor in 7 (15.1%), with an overall success rate of 78.3%.

Ma et al.³⁶ reported on the midterm to long-term follow-up of PDN. From March 2002 to October 2003, 34 patients who underwent PDN replacement were followed for an average of 52.6 months (range 48 to 66 months). Although initial clinical results were encouraging, the patients did not do as well over the longer term. At 12 months after operation, a significant proportion of patients recovered from low back pain or leg pain, with an ODI score that decreased an average of 18.2% and pain on the visual analogue scale (VAS) that decreased to 1.8. The average increase of the postoperative disc height was 17.6%, with a range of motion of 9.2 degrees. At the final follow-up, the ODI increased from 18.2% 12 months after operation to 31.2%, VAS increased from 1.8 to 3.1, disc height improvement averaged to 13.5%, and range of motion decreased to 6.8 degrees. The rate of degeneration or breakages of the end plates was 64.7% (22 of 34), and implant device migrations were observed in 25 patients.

Despite these complications, early clinical results with the PDN have been encouraging. With continuing changes in the implant structure and the use of suitable approaches such as the anterolateral transpsoatic approach and direct lateral approach, reduction in the complication rates can be expected.

Published clinical data are limited for other devices. For the DASCOR device, 2-year follow-up data have been reported from European centers for 85 patients implanted with the device for single-level DDD between February 2003 and July 31, 2007. Mean VAS and ODI scores improved significantly after 6 weeks and throughout the 2-year follow-up. Radiographic results showed, at a minimum, maintenance of disc height with no device expulsion and, despite Modic type 1 changes, no subsidence. Analgesic medication use decreased 84.6% by 24 months from the preoperative average. Seven patients required explants for various reasons, but the device was in its intradiscal implant position in all patients, and no expulsions of the device had been observed.²⁶

Berlemann and Schwarzenbach²⁷ reported on the use of NuCore as an adjunct to discectomy to replace the loss of disc tissue after herniation and microdiscectomy. After a standard microdiscectomy procedure in 14 patients, the hydrogel material was injected into the nuclear void to replace what tissue had been lost to the herniation and surgery. Results showed significant improvement for leg and back pain and functional scores. No complications or device-related adverse events were observed. MRI controls confirmed stable position of the implants with no repeat herniations. Radiographic measurements indicated better maintenance of disc height compared with literature data on microdiscectomy alone.

A short follow-up has been reported after implantation of the NUBAC device. Ten patients in Mexico were implanted with NUBAC after discectomy and followed for 3 months. VAS score improved from 8.1 to 2.5 (P < .05), and the ODI improved from 58.2% to 24.2% (P < .05). Disc height before surgery was 9.4 mm and 3 months postoperatively was 12.5 mm with no complications, migration, or subsidence.²⁹

Summary

As the population ages, the prevalence of DDD and its economic and social burden are expected to continue to increase. Current treatment modalities of discectomy and fusion do not slow down the degenerative cascade. It is hoped that nuclear replacement may bridge the gap between discectomy and fusion or a disc replacement procedure, while attempting to slow down the degenerative process. Nuclear replacement can be a minimally invasive, motion-preserving option with minimal risk of neurologic and vascular complications in contrast to a fusion or a disc replacement. Nuclear replacement strategies preserve the anulus and end plate, making it easier to be revised to a fusion or a disc replacement if required.

Over the years, attempts at nuclear replacement were limited by the lack of material technology and a suitable design. With more recent advances, it has been possible to use materials that replicate the native nucleus properties and are able to imbibe and hold water. Multiple designs are in the process of development and in clinical studies. Current outcome data have been encouraging in terms of improved VAS and ODI scores. However, owing to lack of randomized controlled trials, it is unclear if these results are better and more durable than stand-alone discectomy or fusion procedures. Longer-term studies are required to assess the effects of nuclear replacement on the progress of degenerative cascade.

Major clinical concerns with nuclear replacement have been the risk of implant extrusion and end-plate changes with subsidence. It is hoped that continuing improvements in material selection, design properties, and surgical techniques will address the issue of implant extrusion. Implant extrusion rates have been declining over the years with changes in technology. End-plate changes are thought of as precursors for subsidence. However, these are also common after most types of surgical procedures involving the disc, such as decompression, discectomy, and fusion. Subsidence seems to stabilize over time and, in most cases, leads to no clinical sequelae. Attempts at dorsal motion preservation systems such as dynamic rods and facet replacements in combination with a nuclear replacement are also ongoing. These technologies may provide options for 360 degrees of motion preservation.