Abstract

Introduction

As the prevalence of adult spinal deformity (ASD) continues to increase with an aging global population (32%-68% of individuals older than 65 years), its debilitating effects on both the physical and mental health of patients are of great concern. Operative treatment for ASD can provide significant improvement in health-related quality of life (HRQoL), yet it is still associated with a high morbidity rate. Despite facing a greater risk of complications, reported as high as 71%, elderly patients seem to benefit from surgery, equally or even potentially more so than their younger counterparts.

Implant-related complications, including rod fractures (RF) and pseudarthrosis, are among the most common causes of postoperative morbidity and reoperation. Roughly one-third of patients undergoing surgery for ASD will experience an implant-related complication, with up to half of these requiring reintervention within 2 years of the original surgery. Rod fracture rates differ widely across the literature and range from 6.8% to 24% depending on the heterogeneity of the patient population, variability of construct design, and patient follow-up. With almost two-thirds of RFs occurring more than 2 years after the primary intervention, the overall RF rate can approach 40%, with 61% of these requiring revision surgery.

Rod fractures in ASD also have a serious economic impact. In a recent investigation assessing the cost of spinopelvic complications after ASD surgery, Zuckerman et al estimated that each revision surgery owing to RF costs approximately $76,695 + $10,000. These significant costs involve not only the actual cost of the rods and grafting material but also the cost of the surgery-anesthesia time, OR time, hospital stay, and postoperative recovery.

Historical Perspective and Nomenclature

With the introduction of pedicle screws, dual rod constructs rapidly became the standard of care in spine deformity surgery. However, achieving optimal results in complex cases is challenging given the high rate of pseudarthrosis and construct failures resulting in correction loss, decreased HRQoL, increased pain, and reintervention rates.

Furthermore, ever since the introduction of pedicle subtraction osteotomies (PSO), and now with a better understanding of spinopelvic goals, 3-column osteotomies have gained widespread popularity, especially in the setting of rigid deformities. However, the inherent destabilizing nature of this procedure results in significant stress at the osteotomy site with high rates of pseudarthrosis and RF. With instrumentation failure reported in up to 30% of patients and a RF rate more than 2-fold higher for patients receiving a PSO, strategies are still being sought to lessen these complications.

The use of supplemental rods in deformity surgery thus began as a means to prevent pseudarthrosis and reduce rod fracture rates across PSOs. Their indications subsequently broadened to improve immediate stiffness and prolong construct durability while bone fusion was achieved at other high-stress areas.

El Dafrawy et al proposed a classification for multiple rod constructs across 3-column osteotomies (3CO). This system is based on defining the different types of rods but does not provide recommendations as to the ideal number or type of rods that should be used and is limited to constructs across 3CO.

Considering that supplemental rods are often used in patients without 3CO, Ramey et al published a more general nomenclature for these constructs that can be used in all areas of the spine regardless of osteotomies. In this classification, the authors describe the primary rod as the longest rod, which typically spans the entire construct and 5 supplementary rod types:

FIGURE 1. The anterior-posterior radiograph demonstrates a dual construct using an alternating pedicle screw pattern. This configuration establishes 4 independent rods, effectively creating 2 separate spinal constructs within each reconstruction. The lateral secondary rods (*) are shorter than the medial primary rods (∆) and are directly connected to both the pedicle screws and the primary rods through lateral connectors. Adapted from Spine Journal, 18 (3), Shen FH, Qureshi R, Tyger R, Lehman R, Singla A, Shimer A, Hassanzadeh H, Use of the “dual construct” for the management of complex spinal reconstructions, 482-490, Copyright (2018), with permission from Elsevier.

Given the simplicity and broad applicability of this lexicon, the authors recommend its use to standardize the description of supplemental-rod constructs.

Biomechanical Analysis

Biomechanical studies have supported the concept that adding rods to a construct will decrease rod and surface strain, thus increasing the lifespan of the rods. Hallager et al used cadaveric models to demonstrate that 4-rod constructs significantly decrease primary rod strain compared with 2-rod configurations (P < .033).

Shekouhi et al used a finite element model to assess the biomechanical effects of lateral vs in-line satellite rods adjacent to a lumbar PSO. They found that satellite rods reduce von Mises stress on primary rods at the PSO compared with the control model. They also found differences between both configurations and reported that lateral satellite rods seem to have increased range of motion (ROM) across the PSO site, which resulted in less von Mises stress and better biomechanical properties as forces are transferred away from the rods and into the anterior vertebral column.

The anterior transfer of forces theoretically increases osseous load, which according to the Wolff law is key to promoting bone fusion. This was also observed in a study by La Barbera et al in which the addition of supplementary rods across PSO models significantly decreased primary rod strain in flexion (50%), extension (40%), and axial rotation (40%) compared with 2-rod constructs. However, and despite increased posterior stiffness, the ventral spine was not shielded from compressive loads, which suggests that the use of supplementary rods not only adds robustness to the construct but could also potentially improve bone fusion.

FIGURE 3. This is a 75-year-old woman with a history of symptomatic adult scoliosis who underwent a T9-Pelvis fusion with L4-L5 and L5-S1 posterior column osteotomies and transforaminal lumbar interbody fusions. Notice a left medial accessory rod (*) attached to the main rod through side-to-side connectors and a right iliac accessory rod (∆) attached to an independent iliac bolt.

Berjano et al analyzed the effect of delta-rods on stiffness and primary rod stress reduction using an infinite element model. The authors found that delta rods had a superior reduction of von Mises stress and ROM across the PSO than conventional accessory and satellite rods in all loading conditions. They attributed these findings to the delta rods providing a straighter axial support and by bridging the PSO more posteriorly, as opposed to the highly angulated and more anteriorly seated accessory and satellite rods.

Apart from reducing the stress associated with cyclic biomechanical loading, supplemental rods can also have a beneficial effect over primary rods by reducing the need of extreme in situ contouring. Gupta et al reported that having satellite rods at sites with severe angular corrections (3COs) allows for primary rods to be under-contoured, which improves their fatigue life. Furthermore, Tang et al also found that contouring rods from a 20-degree PSO to either 40 or 60° significantly decreased their fatigue life (Hazard Ratio (HR) = 7863.6, P = .0144). In fact, Gelb et al found that using in-line satellite rods across 3CO would allow for under-contouring primary rods by 14.2°, thus potentially protecting the primary construct.

Considering the biomechanical advantages provided by supplemental rods, it could be assumed that adding even more rods would theoretically lower the rate of complications. However, that does not seem to be the case. Shekouhi et al used a finite element model to assess the effect of 4-, 5-, and 6-supplemental-rod constructs. As expected, they found that the ROM across the PSO was inversely proportional to the number of rods. Initially, this translated to a better force distribution across the PSO site going from 336N with 2-rods to 348.6N with 4-rods. However, when adding more rods, the forces across the anterior column of the PSO decreased to 343.2N with 5- and to 324.2N with 6-rods. Even though 6-rods do provide the most rigid posterior construct, they offload considerably the anterior column, which could result in a greater risk of pseudarthrosis.

Clinical Evidence

Three-column osteotomies represent highly unstable areas with an inherently greater rod strain, high risk of pseudarthrosis, and rod failure. A multicenter International Spine Study Group cohort of patients with ASD undergoing 3CO with a minimum 2-year follow-up reported RF to be the most common complication (32%). These high rates of mechanical failure have resulted in a greater acceptance of supplemental rods as a means to help reduce primary rod strain.

Gupta et al first introduced the concept of satellite rods for PSOs to improve osteotomy closure and stiffness. The authors reported a statistically significant difference in nonunion (3.4% vs 25%, P = .035) and RF rates (0% vs 25%, P = .008) between the 4- and 2-rod construct groups at a 3-year minimum follow-up. However, there were confounding variables including differences in rod material, anterior column support, and use of rh-BMP-2, which warranted further research.

FIGURE 5. Delta-rod configurations: posterior intraoperative image of delta configuration; lateral radiograph of delta configuration. Adapted with permission from Berjano P, Xu M, Damilano M, Scholl T, Lamartina C, Jekir M and Galbusera, Supplementary delta-rod configurations provide superior stiffness and reduced rod stress compared with traditional multiple-rod configurations after pedicle subtraction osteotomy: a finite element study, European Spine Journal, 28 (9), 2198-2207, 2019, Springer Nature.

To reduce confounders, Hyun et al compared matched cohorts of patients who underwent 3CO. They found significant differences in rod failure (11 RF vs 2 RF, P = .002) and revision surgery for pseudarthrosis at the osteotomy site (6 vs 0, P = .011) between the 2-rod and supplemental-rod construct cohorts. However, the number of rods used was variable (3- to 5-rods), and there was no clear description of the type of supplemental rods used.

Similarly, in another cohort of 264 patients with ASD treated with at least 1 lumbar PSO and a minimum 2-year follow-up, implant failure between 2-rod and supplemental-rod constructs was assessed. This study included 190 patients with 2 rods, 36 patients with 3 rods, and 38 patients with 4-rods using both accessory and satellite configurations. When grouped together, patients with supplemental rods had a lower revision rate for RF than the 2-rod cohort (15% vs 26%, P = .035). Interestingly, and differing from data observed in biomechanical studies, satellite rods seemed to provide better support than accessory rods, having lower rates of failure at the PSO site (10% vs 31%, P = .034) and fewer revisions for pseudarthrosis (0% vs 23%, P = .009).

As data emerged and evidence started to support the use of supplemental rods to improve mechanical outcomes in patients undergoing 3CO, surgeons began to consider them for other high-risk situations as well. Patients undergoing ASD surgery, even without 3CO, still have high RF rates, nearly 40%. However, instrumentation failure in patients with 3CO seems biomechanically different from those without a 3CO. While the former occurs earlier after surgery and is secondary to high-stress factors at the osteotomy site, the latter usually occurs years later and appears to be dependent on multiple factors, which ultimately translate to a greater pseudarthrosis risk.

Various articles across the literature have documented particularly high rates of pseudarthrosis and RF across the lumbosacral junction with rates as high as 77%. Considering that nearly 80% of RFs develop at or below L3, there is clear evidence that the lumbosacral junction is an area of high strain and, as such, could benefit from additional rods to redistribute the loads. In a retrospective analysis of patients with ASD, Rabinovich et al found that most RFs (81.3%) occurred between L4-S1 and that dual-rod constructs were more likely than supplemental-rod constructs to present with RF (21% vs 5.8%, P = .012). In addition, their multivariate analysis also demonstrated that accessory rods were protective against instrumentation failure (OR 0.231, 95% CI 0.051-0.770, P = .029).

Similarly, in a small study comprised of 31 patients with ASD, Merrill et al compared dual-rod vs supplemental rod constructs and reported that lumbosacral RF was significantly reduced when using more rods (0/16 vs 6/15, P = .007). Interestingly, the authors reported that both CT and surgical findings revealed hypertrophic nonunions in the dual-rod cohort, which suggests that pseudarthrosis may have been due to mechanical instability instead of biology. This article once again supports that supplemental rods not only add stiffness to the instrumentation but also potentially contribute to the mechanical loading needed for fusion, which is particularly important across the lumbosacral junction.

In a meta-analysis by Yang et al that included 10 studies with nearly 800 patients with ASD (399 supplemental-rod and 398 dual-rod patients), supplemental rods were found to be protective of RFs (Relative Risk (RR), 0.43; 95% CI 0.33-0.57, P < .01), pseudarthrosis (RR, 0.38; 95% CI 0.28-0.53, P < .01), and need for revision surgery (RR, 0.44; 95% CI 0.33-0.58, P < .01). Pooled analysis found that pseudarthrosis was significantly reduced in the supplemental-rod group (14.2% vs 35%, P < .01), which suggest that the additional support potentially offers a better environment to promote bone fusion, thus reflected in a lower RF rate as compared with dual-rod patients (15.8% vs 32.9%, P < .01).

Interestingly, the authors also noted that at the last follow-up, there was a statistically significant improvement in sagittal vertical axis, lumbar lordosis – pelvic incidence missmatch, Oswestry Disability Index, Scoliosis Research Society-22score, and visual analog scale-back pain scores in the supplemental-rod construct population. These findings could be related to the reduction in mechanical complications, which are known to be associated with loss of correction and decreased HRQoL outcome scores.

Kickstand Rods

Makhni et al reported for the first time the use of a “kickstand rod” in a patient with severe coronal malalignment. This technique was conceived for a supplemental rod to be attached to an independent iliac bolt and serve as a buttress, thus allowing distraction of the coronal deformity while redistributing the loads.

Given its recent description, there are a few reports in the literature that thoroughly assess the use of kickstand rods. In a retrospective evaluation of a prospectively collected multicenter database, Mundis et al analyzed a 1:2 matched cohort of patients with and without kickstand rods and found that the former group had improved coronal correction compared with the control group (18 vs 35 mm, P < .01). Though there were no statistically significant differences in RFs between groups, follow-up was insufficient to properly assess mechanical complications.

Two other studies by Redaelli et al and Buell et al also found kickstand rods to be a safe adjuvant technique for coronal realignment, but their results were limited by small samples, no control groups, and short follow-up. Bearing in mind that the degree of coronal correction has been associated with increased odds of RF (1.03, 95% CI 1.01-1.07, P = .044) and even more so if the correction is greater than 30 mm (OR 7.72, 95% CI 1.17-51.10, P = .034), the use of a kickstand rod may provide additional mechanical strength in a select group of patients. However, there are no formal biomechanical studies testing these rods or any clinical evidence supporting their long-term durability.

Where Should Supplemental Rods Be Used?

It is widely accepted that the extra support should be used across high-risk pseudarthrosis areas like osteotomies (3CO and posterior column osteotomies) and at hypermobile segments like the lumbosacral junction. However, there is no consensus in the literature on how far above and below the osteotomies and where the proximal and distal fixation points of the supplemental rods should be.

FIGURE 6. This is a 67-year-old woman with a history of 5 previous spine surgeries and osteoporosis who underwent surgical treatment with T4-Pelvis fusion and an L4 PSO to correct a flatback deformity. Lateral iliac accessory rods (*) were used to supplement the PSO extending from independent iliac bolts to the primary rods between T12/L1 using side-to-side connectors. PSO, pedicle subtraction osteotomies.

In a study that included 253 patients with ASD, Lee et al assessed different protective strategies to prevent pelvic fixation failures (PFF) and reported that a higher number of rods crossing the lumbopelvic junction were protective for failure. Multivariate analysis revealed that anchoring the accessory rods at the S2/ilium junction vs S1 (OR 0.2, P = .004) and a higher number of rods crossing the junction (OR 0.15, P = .002) protected patients from PFF. This is in line with biomechanical studies that have shown enhanced construct stability and lower primary-rod and screw strain at the lumbopelvic junction when supplemental rods are used.

Berlin et al recently published a series of 82 patients with ASD who underwent deformity correction using a novel quad-rod technique. Early results, with a median follow-up of 2 years, showed only 1 patient with an incidental asymptomatic RF (proximal to the coverage of the accessory rod) and 1 case of PFF. Similarly, Uotani et al found that patients with dual bilateral sacro-pelvic fixation had a lower risk of S2AI screw loosening than patients with bilateral single screws (23% vs 65%, P = .011).

Considering that the most accepted mechanism behind PFF is pseudarthrosis, that translates to screw loosening or fracture or RF, the additional support of supplemental rods crossing the lumbosacral junction and specifically distal to S1 seem to provide a solid foundation that can potentially reduce the risk of RF and PFF improving the odds for proper osseous fusion.

The proximal extent of the supplemental rods is a subject that has not been well described either. Berlin et al reported that most proximal connections in their study were between T12 and L2 (65.8%); however, their cohort was heterogeneous and included patients with upper instrumented vertebrae at the upper-, mid-, and lower-thoracic spine, which could affect the optimal landing spot.

Considering most RFs occur between L3 and S1, which are also the segments where most osteotomies are performed to correct the sagittal plane in ASD, these segments may benefit the most from additional rods. However, theoretically a “soft-landing” near the upper instrumented vertebrae should help prevent the risk for proximal junction complications; thus, the added stiffness of supplemental rods should be avoided near the junction. Nevertheless, there are no clear recommendations or evidence as to where the optimal attachment should be.

Are Supplemental Rod Constructs Safe?

The use of multiple rods has been linked to some drawbacks, including increased risk of proximal junctional kyphosis (PJK)/proximal junctional failure (PJF), delayed pseudarthrosis, and increased intraoperative time. Regarding proximal junction complications, a study by Han et al raised concerns after their results revealed that stiffer constructs (CoCr rods and multiple rods) would lead to a higher incidence of PJK/PJF. The authors found that increasing rod stiffness reduced the risk of RF (0% vs 32.4%) but also adversely affected the occurrence of PJK/PJF (60% vs 26.5%, P = .015). However, the rod material was a confounding variable because they compared patients with supplemental-rod constructs composed of CoCr rods vs 2-rod constructs composed of titanium.

In a study by Ye et al that included 1300 ASD prospectively enrolled patients in an International Spine Study Group database, the authors analyzed the impact of supplemental rods on PJK/PJF. At last follow-up, patients with supplemental rods had similar incidence of PJK (58.6% vs 58.1%) and revision surgery (13% vs 17.7%) compared with patients who had a 2-rod constructs. After controlling for demographic and radiographic parameters, the PJK-free survival analysis demonstrated similar durations among both groups (HR 0.889, 95% CI 0.745-1.062, P = .195).

Furthermore, 2 meta-analyses, one by Yang et al and 1 by Moniz-Garcia et al, revealed that PJK rates were similar between the 2-rod and supplemental-rod construct groups, with low heterogeneity observed in both studies. Together these findings seem to support that supplemental-rod constructs do not increase the risk for PJK/PJF despite initial concerns. This may be due to most surgeons adding the supplemental rods at the areas of major intervertebral rod stress, which are usually away from the uppermost instrumented vertebrae, thus adding stiffness low in the construct without affecting the junction itself.

Finally, after analyzing reported complications in multiple studies and 2 meta-analyses, evidence suggests that supplemental-rod constructs do not seem to significantly affect estimated blood loss, operative time, overall surgical complication rates, neurological deficit, dural tears, wound-related complications, and systemic complications when compared with 2-rod constructs.

Conclusion