Advances in Spinal Fusion Strategies in Adult Deformity Surgery

Goals of Correction

In recent decades, there has been increased understanding of the importance of sagittal plane correction in treating patients with adult spinal deformity [12, 43]. Further research has shown the importance of correcting lumbar lordosis (LL) in relation to a patient's native pelvic incidence (PI) [43]. As surgeons began to apply these principles and analyze their outcomes, it became clear that the goals of correction for ADS may vary between patients. The notion of age-adjusted parameters emphasizes that PI–LL mismatch may change substantially for patients over 74 years old [28]. New technology using the latest advancements in machine learning and predictive analytics may offer substantial clinical advantages in determining unique goals of correction to reduce the rates of pseudarthrosis, revision surgery, and proximal junctional failure [41, 42, 45].

Rods

In 2006, Shen et al. described a biomechanical advantage of multirod over dual-rod constructs; with the increased rods, motion at L5–S1 decreased, helping to lessen lumbosacral junction stress [44]. Hyun et al. in 2014 were the first to link fusion rates with this technique, demonstrating fewer pseudarthroses post-operatively when multiple rods were used [19]. In a deformity-focused study, ASD patients with long fusion to the sacrum benefited from multirod constructs, which promote fusion at the lumbosacral junction [33].

As accessory rods were being adopted more frequently, Gupta et al. described their ideal location and placement in ADS [14]. In this series, two rods were used to span the osteotomy to avoid a severe angular bend needed to incorporate the osteotomy site screws with the rest of the construct and avoid creation of a weak biomechanical point in the rod. Since nonunion and rod breakages are known complications of osteotomies, this was a valuable contribution to the multirod evolution. Using four rods allowed for a more controlled osteotomy closure and reduced biomechanical stresses to the rod during rod contouring. Most important, rates of pseudarthrosis and rod failure were lower when four rods were placed. This is especially important for patients requiring higher grade osteotomies, given the substantial biomechanical stress created across osteotomy sites [20, 48]. The authors acknowledge the complex relationship between rod fracture and pseudarthrosis. While rod fracture is associated with pseudarthrosis, there is evidence that rod fractures occur in solid fusion [9]. Further research is required to precisely define the best rod configuration/design to avoid both pseudarthrosis and rod fracture.

Regarding rod type, studies have shown that the use of stiffer rods (cobalt–chromium or steel) can increase strength and resistance to fatigue better than less stiff rods (titanium). Rod motion and rod strain are decreased with stiffer rods, which subsequently are less likely to fracture [15]. That said, stronger and stiffer is not always better; rods should be selected on a case-by-case basis. A stiffer rod may decrease fracture risk but can lead to proximal junctional kyphosis (PJK) more frequently than a titanium rod [16].

Interbody Fusion

The L5–S1 disc space is the most common site of pseudarthrosis in scoliosis correction surgery [24]. For this reason, interbody fusion at L5–S1 is commonly performed in ADS, via an anterior or posterior approach. Interbody placement can potentially increase fusion rates because it completes a circumferential (360°) fusion at that level, decreases stress on the pedicle screws, and increases lordosis. While many studies have examined factors such as foraminal height, disc height, segmental angle, and pelvic parameters with interbody placement, few studies have focused on the differences in fusion rates between anterior and posterior approaches. Tsuchiya et al. showed that in ADS adding interbody support increased L5–S1 fusion rates to 95.9% versus 83.3% without interbody support [47]. Due to the long follow-up (minimum 5 years) and relatively large sample size, this study contributed to the evolution of using interbody fusion more often in deformity surgeries. In a review of failed fusions specific to the lumbosacral junction in ADS, Harimaya et al. strongly recommended including interbody support [17].

However, in some newer studies, adding an interbody to increase fusion rates has not been superior to using bone morphogenetic protein (BMP). Rahman et al. compared 20 mg of BMP and interbody fusion in patients with spinal deformity and found no benefit of interbody fusion over BMP alone [39]. Annis et al. conducted a similar study and concluded that performing interbody fusion was not necessary if BMP was used in addition to sacropelvic fixation [3]. They cited a 3% nonunion rate at L5–S1 at a mean of 38 months without the use of interbody support. Prospective, multicenter, and randomized studies in this realm would help to clarify whether interbody fusions increase fusion rates.

Iliac Fixation

Distal fixation strength is paramount in achieving fusion, although poor bone quality of the sacrum can lead to hardware failure caused by weak purchase of S1 pedicle screws. For this reason, numerous techniques have been employed for lumbosacral and iliac fixation to increase pullout strength, including S2–alar–iliac screws, iliac bolts, S2 screws, and sacral–alar screws [27].

Iliac bolts were introduced by Harrop et al. in 2008 for their improved pullout strength and construct strength, to improve correction and alignment in ADS [18]. They have been shown to decrease rod breakage, correct pelvic obliquity, and decrease implant loosening [37]. Iliac screws were studied in patients undergoing long fusion to the sacrum and were shown to be effective in protecting the sacroiliac screws from failure [47]. In this series with 5-year follow-up, there were no cases of S1 screw failure by breakage, loosening, or pullout. In this early study with iliac screws, 23 of 67 iliac screws needed to be removed for reasons including skin prominence, wound complications, breakage, back out, and haloing. Ebata et al. in 2018 attempted to decrease the rate of iliac screw loosening by placing bilateral dual iliac screws for deformity correction, which showed longer stability for spinal and pelvic fusions in ADS [11]. Because iliac fixation immobilizes the sacroiliac joint, it could be predicted to have a negative toll on a patient's daily activities. One study of the impact of iliac instrumentation on quality of life in ADS patients with more than eight levels of fusion concluded that iliac screws did not negatively influence patients’ quality of life or functional ability [30].

A drawback of the iliac bolt is that it can require an offset connector to connect to more medial screws at other levels. Additionally, placement can require a separate skin or fascial incision, and screw-head prominence can cause pain and require removal. For this reason, S2–alar–iliac screws were introduced with a more medial starting point on the sacrum, with an S2–alar screw that traverses the sacroiliac joint. This was shown to have anatomic feasibility [34], and subsequently, O'Brien et al. showed that S2–alar–iliac screws had equivalent stability to iliac screws biomechanically [35]. Finally, a clinical study showed good outcomes at 2-year follow-up with S2–alar–iliac screws with solid correction of pelvic obliquity and radiographic anchor stability [46], establishing the S2–alar–iliac screw as an alternative method of pelvic fixation in ADS.

Bone Morphogenetic Protein and Iliac Crest Bone Graft

The use of iliac crest bone graft (ICBG) and recombinant human bone morphogenetic protein-2 (rhBMP-2) has been shown to decrease the incidence of pseudarthrosis in many areas of spine surgery.

ICBG was considered the “gold standard” for achieving fusion due to its osteoconductivity, osteoinductivity, and osteogenicity, although ICBG does have drawbacks including donor site pain, increased blood loss, and longer operative times [10].

rhBMP-2 is a recombinant protein that helps with accelerated bone formation to achieve fusion. Fusion rates have shown to improve with rhBMP-2 in the long constructs involved with ADS [17, 29]. Annis et al. showed that successful fusion rates with rhBMP-2 in ADS can be replicated with low-dose rhBMP-2 [3]. Adverse effects include heterotopic ossification, post-operative radiculitis, and seroma formation [31, 32]. Additionally, rhBMP-2 is expensive [13]. However, studies evaluating costs have shown that the cost at 2 years post-operatively is actually less with BMP than ICBG due to decreased revision surgeries, among other factors [1]. In a 2017 meta-analysis of 166 studies evaluating BMP use in ADS, Poorman et al. concluded that BMP provides a strong protective effect against pseudarthrosis resulting in reoperation and is safe and effective for grafting in ADS, with no significant complications other than radiculitis [38]. In a study of BMP and ICBG with 4- to 14-year follow-up, Kim et al. noted that BMP was superior to ICBG in achieving long fusions from the thoracic spine to the sacrum in ADS [26]. Pseudarthrosis was increased in the ICBG group; the fusion rate was 93.5% in the BMP group and 71.9% in the ICBG group.