14.1 Introduction

Anterior release of the spine involves removal of the discoligamentous complex of the intervertebral segment with its associated cartilaginous endplates. This can be achieved through either an open approach or thoracoscopically. The primary indications typically have been to improve curve flexibility in large, rigid curves; prevent crankshaft deformity; increase fusion rates; and restore kyphosis by shortening the anterior column. Over time, the indications for anterior release have been refined based on studies regarding the cost, effectiveness, comparative outcomes, and complications of the procedure. In addition, advances in posterior surgical techniques and posterior-based spinal implants have resulted in further redefining the role of anterior release and fusion for adolescent idiopathic scoliosis (AIS) in the modern era. This chapter will focus on the current indications of anterior release and fusion in the setting of AIS surgery, providing a rationale as to why this skill set remains important to maintain for deformity surgeons treating AIS.

14.2 History and Evolution of Anterior Approaches

Since the introduction of instrumented fusions by Paul Harrington, ¹ significant advances in the surgical treatment of AIS have been made. Harrington instrumentation was a posterior-based system and yielded acceptable results at that time for most deformities. Dwyer et al, ² learning from the anatomy of scoliosis by Roaf, ³ noted scoliosis to be a “rotational lordosis,” highlighting that the deformity could be treated with the distraction of the concavity as with Harrington instrumentation or with compression of the convexity which they described through an anterior approach. ² This anterior approach was felt to better address the rotational deformity. The anterior approach to the spine was first reported to resect congenital vertebrae ^{4 , 5} and was then popularized for the treatment of tuberculosis and Pott disease by Hodgson and Stock ⁶ before being adopted by Dwyer for the treatment of idiopathic scoliosis. Floman et al ⁷ outlined the shortcomings of isolated posterior and/or anterior fusions in a group of patients with mixed pathology but including idiopathic scoliosis. The posterior approach limited the ability to address structural changes (congenital or acquired) in the vertebral bodies and limited rotational correction. The anterior approach allowed for ample correction of the deformity with release and osteotomies as needed but was not mechanically stable enough to allow for fusion mass to form. The Dwyer instrumentation provided some stability, but Floman et al noted a relatively high incidence of pseudarthrosis. Therefore, these authors resorted to a combined anterior approach for mobilization with or without Dwyer instrumentation and then Harrington posterior instrumentation with Hibbs spinal fusion. ⁷

The indications for this combined approach from their study were as follows:

• 
  

  Severe rigid kyphosis.

  
  
• 
  

  Major thoracolumbar or lumbar scoliosis with trunk imbalance or pelvic obliquity.

  
  
• 
  

  Congenital scoliosis curves associated with a hemivertebra or anterior unsegmented bar.

  
  
• 
  

  Absent posterior elements associated with severe scoliosis or kyphosis.

  
  
• 
  

  Failed anterior fusion.

There was later labeling of the “crankshaft phenomenon,” ⁸ and the authors advocated for the addition of an anterior approach to negate this occurrence. With a focus on idiopathic scoliosis patients only, the list can be refined to a more modern indications list:

• 
  

  Large, rigid scoliosis.

  
  
• 
  

  Restoration of sagittal balance.

  
  
• 
  

  Thoracolumbar and lumbar curves.

  
  
• 
  

  Prevention of crankshaft.

Posterior-based instrumentation evolved through Luque wires to Cotrel–Dubousset instrumentation, and the list of indications for combined procedures remained similar, ⁹ with the addition of the recognition of the crankshaft phenomenon by some authors. ¹⁰ A combined anterior and posterior surgery in the early descriptions ^{7 , 11 , 12} was a two-stage surgery with recovery interposed and therefore a prolonged hospitalization. Shufflebarger et al and Powell et al began to investigate the possibility of combined anterior and posterior surgeries as the surgical and anesthetic techniques were refined, allowing for same-day surgery to be performed. ^{9 , 10}

It was in this era with the first descriptions of same-day anterior and posterior spine surgery that better optics were developed and the ability to perform video-assisted thoracoscopic surgery (VATS) of the spine was first reported by Regan et al and Mack et al. ^{13 , 14} Benefits of thoracoscopy for lung procedures included reduced pain, improved shoulder girdle function, and shorter hospital stay ¹⁵ compared to an open thoracotomy approach. These benefits were applied to spine surgery, and the barrier to anterior spine access was reduced. ^{16 , 17 , 18 , 19 , 20} The perceived reduction in surgical comorbidity of the VATS procedure allowed increased adoption of a same-day anterior release–posterior spinal fusion or minimization of the comorbidities experienced by patients in recovering from anterior release while awaiting posterior fusion that was hypothesized to prolong hospital stay beyond the scheduled delay between procedures. ¹⁰

Newton et al in 1997 demonstrated the safety and effectiveness of the thoracoscopic approach compared to open thoracotomy. ²¹ This article focused on the first 14 cases performed, but despite being early in the thoracoscopic surgeon’s learning curve, the curve correction, estimated blood loss, and complications were similar to those of an open group. The chest tube output was higher, although different surgeons were performing the open approach and therefore differences in postoperative chest tube protocols may have accounted for this. The cost was found to be 29% less in the open group, and there was no difference in hospital length of stay. Again, these were the initial cases of thoracoscopic surgery, and the muscle sparing approach was felt to be worth the additional cost due to benefits in shoulder girdle function and recovery. The educational effort required by the surgeon was also highlighted. The same authors then evaluated the feasibility of thoracoscopic surgery in patients weighing less than 30 kg. ²² The small size of the chest cavity presents new challenges, and in very small children (<20 kg), the benefits in reduced exposure are not as significant. The technique was shown to be effective in the small children.

Investigators also sought to evaluate the mechanical efficacy of the anterior release. Cheung et al published two articles with a small cohort and an expanded cohort demonstrating that the anterior release in idiopathic scoliosis increased the fulcrum bending flexibility from 39 to 54% after release prior to posterior instrumentation. ^{23 , 24} The Cobb angle on the fulcrum bend was decreased from 45 to 34 degrees, and this was correlated with the postoperative correction of 31 degrees. They concluded that the release improves spinal flexibility and increases the spinal deformity correction. Another study questioned the benefit of the anterior release. Hempfing et al in 2007 evaluated improved Cobb angle in patients with AIS after anterior release. The average preoperative curve for six patients was 89.7 degrees. Their protocol consisted of Cotrel dynamic traction for all patients, but they found an increase in coronal correction in traction of only 5.5 degrees after anterior release. They advocated for judicious use of anterior release for hyperkyphosis, coronal imbalance, or massive curves. ²⁵

As with all technology, there was no universal adoption of thoracoscopic surgery or even anterior release. Burton et al in 2005 ²⁶ compared their outcomes in large (>70 degrees) thoracic idiopathic scoliosis curves with posterior instrumentation and found comparable correction to the series published in the literature utilizing a circumferential approach highlighted previously. They advocated for isolated posterior instrumentation and arthrodesis for curves up to 90 degrees without the additional expense and morbidity of an anterior approach. The same group also found no need for anterior fusion to prevent crankshaft with the use of third-generation instrumentation involving hooks, wires, and screws for fixation. ²⁷ Arlet et al in 2004 also found comparable correction of 70- to 90-degree curves with limited flexibility (Cobb angle remains >45 degrees on bend) treated with third-generation instrumentation (hooks, wires, screws). ²⁸ Their correction with a posterior-only approach was similar to a cohort of combined approach patients from the literature.

Further development of posterior instrumentation included the increased application of pedicle screw fixation as anchors to the spine. Luhmann et al in 2005 found at their institution that the use of all pedicle screw constructs resulted in similar correction to an anterior release and thoracic hook instrumentation. ²⁹ They did find anterior release allowed for more coronal curve correction when comparing thoracic hook constructs with and without release. Luhmann et al also performed a financial analysis and found no difference in an all- pedicle screw construct and circumferential surgery. ³⁰ The increased pedicle screw cost was balanced by the increased surgeon charges, operating room (OR) charges, anesthesia charges, and inpatient room charges in the circumferential fusion group. Suk et al in 2007 demonstrated similar radiographic outcomes when comparing their results in patients with curves between 70 and 100 degrees treated with posterior pedicle-screw–based constructs also to a literature-based cohort. They highlighted that posterior-only correction obviated the need for anterior release and the complications related to anterior surgery. ³¹ The balance of improved correction with minimal complications is the overall goal.

14.3 Complications

A pendulum swings regarding complications. The anterior approach to the spine was added to avoid complications—crankshaft, pseudarthrosis, poor curve correction, and inability to address the congenital defect. As is appropriate in an ever- evolving and refining medical specialty, the analysis of the complications created is important to help further refine the disease treatment. A review of complications of anterior procedures of the spine reviewing 447 patients revealed a 31% complication rate when factoring in major and minor complications. ³² This study included children and adults and many diagnoses. In further breakdown, there was a 16% complication rate in 100 idiopathic scoliosis patients in the group—a much lower rate than the neuromuscular scoliosis group with a 52% rate of complications. Overall, patients 3 to 20 years old being treated for idiopathic scoliosis, neuromuscular scoliosis, kyphosis, and congenital scoliosis had a 26% complication rate with a 9% major complication rate. Also noted in this study is the rate of complications: 32% in the combined anterior and posterior approach group in one OR session, 36% in staged OR sessions, and 24% with an anterior-only approach. In conclusion, the group found the safety profile of anterior spine surgery to be on par with other large operations.

De la Garza Ramos et al reviewed the Nationwide Inpatient Sample and found a 7.6% rate of inpatient complications following surgery for AIS. This rate was 6.7% for a posterior-only approach, 10.0% for isolated anterior approach, and 19.8% for combined approaches. ³³ In the review of the 3,582 patients in the AIS arm of the Harms Study Group database, the complication rate was 3.0% for anterior spinal fusion, 2.4% for isolated posterior spinal fusion, and 5.6% for posterior spinal fusion with anterior release. ³⁴

These complication data are not surprising as the patients indicated for anterior release generally have the largest, stiffest curves. Some of the older data also include neuromuscular patients who have a higher rate of complications postoperatively. There are also some specific risks regarding the positioning and anesthesia for thoracoscopic surgery that are not encountered with posterior-only surgery. The details of the technique will be covered later in this chapter, but the typical position is lateral decubitus with the convexity of the curve up. This allows the convexity of the curve to be addressed and the great vessels typically lie within the concavity somewhat protected. The positioning does necessitate single lung ventilation with either a double-lumen endotracheal tube or deployment of a bronchial blocker to ensure lung deflation. Mucous plugging is a concern after surgery for the lung that is dependent during the procedure, and the double-lumen endotracheal tube has a learning curve for the anesthesia team. Sucato and Girgis reported on bilateral pneumothoraces, pneumomediastinum, pneumoperitoneum, pneumoretroperitoneum, and subcutaneous emphysema from a slightly distal position of the double-lumen tube which led to single left lobe ventilation causing higher than necessary airway pressures and barotrauma. ³⁵ As a result, they modified their technique for thoracoscopy to involve prone positioning as reported by King et al in ³⁶ and Sucato and Elerson in 2003. ³⁷ Roush et al reported in 2001 on a tension pneumothorax in the inflated hemithorax due to guidewire migration during placement of instrumentation. ³⁸ While this is not routine for anterior release alone, care needs to be maintained with any work across the midline.

14.4 Pulmonary Function

Impaired pulmonary function was once thought to occur only in idiopathic scoliosis with very large curves (90–100 degrees). ³⁹ Avoiding curve magnitudes of that size was the primary reason for surgical management beginning around a 50-degree thoracic curve, as curves of this magnitude were shown to progress in adulthood at an average of 1 degree per year. ³⁹ Burrows et al in 1983 demonstrated that during teenage years, the predictive equations available did not relate to observed pulmonary function in children, as the actual forced vital capacity (FVC) and forced expiratory volume in one second (FEV-1) were found to be better than predicted from a height alone. ⁴⁰ This gain was thought to be due to improved chest wall mechanics and muscle strength of the diaphragm, intercostals, and chest wall musculature. Burrows et al also demonstrated a plateau of pulmonary function relative to height in early adulthood with a decline occurring beginning in the mid-30s. Improved population normal values for children were provided when Wang et al compiled the excellent longitudinal prospective collection of pulmonary function in six cities across America for patients aged between 6 and 18 years. ⁴¹ In total, they followed up 11,630 white children and 989 black children with a total of 76,138 annual examinations in white children and 6,324 annual examinations in black children. With this study, percent predicted values by race, height, and age of the patient can more accurately be determined. Based on percent predicted value, mild lung impairment is defined as 65 to 80% of the predicted value, moderate impairment is defined as 50 to 64%, and severe impairment is defined as less than 50% of the predicted value. Further investigation into preoperative pulmonary function revealed moderate or severe impairment in curves as small as 40 to 50 degrees and the majority of curves greater than 80 degrees. ⁴² Correlating this preoperative pulmonary function with curve characteristics (magnitude of the thoracic curve, number of vertebrae involved in the thoracic curve, thoracic hypokyphosis, and coronal imbalance) explained only 19.7% of the variability of FVC, 18% of FEV-1, and 8.8% of the variability between patients in total lung capacity (TLC). ⁴²

Any surgical approach to scoliosis needs to acknowledge that the preoperative pulmonary function of the patient is not a direct linear correlation with curve magnitude, and the surgeon must be mindful of the impact of any surgical technique on lung function. The prior study by Wang et al in 1993 also showed that there is continued maturation and increase in lung function until on average age 16 in girls and age 18 in boys ⁴¹ ; so, any surgical procedure prior to these ages should allow for this continued pulmonary maturation. That expected increase in pulmonary function with maturation is incorporated into the percent predicted pulmonary function, making it a useful metric for comparing pulmonary function test (PFT) values over time. That same percent predicted function is predicated on height of the patient, which is complicated in scoliosis patients due to their deformity and the sudden acquisition of height in the immediate postoperative period. Using raw height data would likely oversell their preoperative percent predicted value or hide any pulmonary limitation. Postoperatively, more correction would lead to more height change and the expectant higher pulmonary function with this increased height. If a patient’s raw numbers remain exactly the same, their postoperative predicted pulmonary function would decline significantly just based on the increase in height. Controlling for this with arm span substituting for height is a consideration, although the standard values were based on height.

In 2000, literature regarding pulmonary function with anterior approaches began to be published. Graham et al detailed the 2-year results of pulmonary function following anterior spinal fusion through an open thoracotomy. ⁴³ This was the first discussion that leaving instrumentation in the thoracic cavity may be an additional hindrance to pulmonary function independent from the chest wall violation. Thoracoplasty had been noted to lead to pulmonary function decline. In the study by Graham et al, absolute FVC, FEV-1, and TLC demonstrated decline at 3 months postoperative but returned to preoperative values at 2 years. Reflecting the idea of improvement in pulmonary function through adolescence, there was a statistically significant decline in percent predicted value, although the values were within 94 to 96% of their preoperative value. ⁴³ Also in 2000, Vedantam et al published a report that would serve as the first entry in a 10-year minimum follow-up of pulmonary function relative to surgical approach out of Washington University. ⁴⁴ Prior to this study, there were often mixed operative approaches combined in single studies without clear delineation or nonhomogenous populations regarding age, gender, curve patterns, severity, and curve etiology. In this Washington University cohort, patients were placed in one of four groups: (1) instrumented posterior spinal fusion; (2) posterior spinal fusion with thoracoplasty based on rib rotation greater than 15 degrees on scoliometer and patient preference; (3) open anterior instrumented spinal fusion; or (4) combined anterior fusion with or without instrumentation and posterior instrumented spinal fusion. They substituted arm span for height to allow a more realistic comparison of pre- to postoperative percent predicted value. At the 2-year mark, patients in group 1 who underwent instrumented posterior fusion demonstrated significant improvement in raw pulmonary function with 14% improvement from preoperative to 2-year postoperative measures in FVC and FEV-1 and 5% improvement in TLC. As a result, they demonstrated relative stability or gain in percent predicted pulmonary function over 2 years. The other three groups demonstrated no statistical change in raw lung function or percent predicted values of lung function except in the group who underwent thoracoplasty—these patients had a statistically significant decrease in percent predicted FVC, FEV-1, and TLC relative to preoperative. ⁴⁴

Kim et al published an update at minimum 5 years postoperative of the Washington University cohort tracking surgical approach and lung function. They found there was an increase in raw FVC and FEV-1 in group 1 (posterior spinal fusion only). All other groups demonstrated only maintenance in absolute FVC and FEV-1 over 5 years which resulted in statistically significant declines in percent predicted values due to lacking expected improvement with maturation. There was no change in percent predicted value of FVC and FEV-1 in the patients with posterior fusion only, but any chest wall violation resulted in loss of percent predicted pulmonary function at 5 years. However, the combined anterior/posterior group maintained FEV-1 percent predicted function but lost FVC. ⁴⁵ A follow-up study of the same group in 2011, with some patients lost to follow-up, reported 10-year results and again found any chest cage disruption to result in mere maintenance of pulmonary function numbers and relative decline in percent predicted values, while posterior spinal fusion alone led to increased absolute pulmonary function measures and maintenance of the percent predicted function. ⁴⁶

The benefits of reduced dissection and chest wall violation that led to the finding of reduced pain of thoracoscopy compared to thoracotomy were also theorized to minimize the effect on pulmonary function. Faro et al in 2005 compared postoperative pulmonary function changes after thoracoscopic anterior fusion and open anterior fusion. Both groups had a significant decline in pulmonary function at 3 months postoperative as measured by absolute values and percent predicted values. In the thoracoscopic group, preoperative FVC of 2.81 L declined to 2.42 L at 3 months but had recovered to 2.82 L on average at 1 year. The thoracotomy group started with an FVC of 3.09 L which declined to 2.50 L at 3 months postoperative and had improved to 2.83 L at 1 year, still below preoperative levels. Similar decline and recovery was noted in the FEV-1 with the thoracoscopy group slightly exceeding preoperative levels at 1 year postoperative FEV-1. The percent predicted values returned to nearly normal for the thoracoscopic group at 1 year but had not returned to normal in the open group, indicating a protective effect of the thoracoscopic approach despite instrumentation. ⁴⁷ This was theorized to be due to the preservation of the latissimus dorsi, serratus anterior, and intercostal muscles. There is still a stall of improvement in pulmonary function which does not fit the expected trend of improvement throughout this adolescent period.

When comparing open versus thoracoscopic anterior release without instrumentation, Lenke et al in 2004 examined 37 patients (21 thoracoscopic, 16 open) who underwent anterior release followed by posterior spinal fusion. All thoracoscopic releases were performed in San Diego and all anterior open releases in St. Louis; so, the pulmonary function testing within group would be standardized. Both groups experienced an improvement in FVC and FEV-1 at 2 years postoperative. Percent predicted values were not reported. The thoracoscopic group improved FVC from 2.48 to 2.85, while the open group improved FVC from 1.97 to 2.43. The FEV-1 improved from 2.06 to 2.37 for the thoracoscopic group and 1.65 to 2.08 for the thoracotomy group. There was no statistical difference in any of these findings. ⁴⁸ Sucato et al in 2009 investigated prone thoracoscopic release with posterior fusion versus posterior spinal fusion alone and found greater improvement in FVC, FEV-1, percent predicted FVC, and percent predicted FEV-1 in the circumferential group but also found improvement in the posterior fusion group. ⁴⁹ When a thoracoplasty was added to either procedure, these benefits were blunted and at 1 year there was a return to baseline but no improvement, unlike in the groups with no thoracoplasty.

A meta-analysis was performed by Lee et al in 2016 to help guide through the multitude of studies regarding the effect of the surgical approach on pulmonary function. In order to be included, the studies had to be clear about the approach used in each case and had to include absolute PFT values and standard deviations to calculate effect sizes. These effect sizes were then compared over time. Twenty-two studies met their criteria. Patients undergoing VATS anterior release with posterior spinal fusion experienced moderate to large improvement at 2 years postoperative. There was no further follow-up. They found that anterior spinal fusion with instrumentation and video-assisted anterior spinal fusion with instrumentation had similar absolute PFT values at 2-year follow-up compared to their baseline. In comparison, posterior spinal fusion with instrumentation with or without a release demonstrated small-to-moderate increases in absolute postoperative PFT values. The small improvement with open anterior spinal release was lost at 5-year follow-up. Any addition of thoracoplasty led to the loss of the improvement noted for the posterior spinal fusions. ⁵⁰

14.5 Modern Approach

Anterior release has evolved since the introduction of thoracoscopy to spine surgery in the mid-1990s. Arunakul et al noted their institution’s trends in 2015. Initially, hyperkyphosis was the most frequent indication, but that was only intermittently the indication for release by 2006. Prevention of crankshaft remains an indication in very young patients with open triradiate cartilage. Severe and rigid scoliosis is the most common indication for anterior release at this point. ⁵¹ A similar decline in combined approaches has been seen nationwide. In a study by Theologis et al published in 2017, a review of the nationwide inpatient sample noted a decline in anterior and combined approaches with an increase in posterior-only surgeries from 1998 to 2011. In 2011 only 2% of idiopathic scoliosis cases reviewed underwent a combined approach down from 9% in 1998. ⁵² In 2014, an international consensus was developed regarding optimal surgical care for AIS. In these guidelines, the posterior approach was among more than 60 aspects considered optimal care, and there was consensus that routine use of anterior release was not considered optimal care. ⁵³

Even in the heyday of anterior release, it was not routine, as noted earlier, with just 9% of cases undergoing anterior release in 1998. In idiopathic scoliosis, the definition of a severe and rigid curve is shifting larger and stiffer after the advent of pedicle screw instrumentation. There is continued focus on the three-dimensional deformity of scoliosis leading some to focus on the sagittal plane as a reason for the release. Ruf et al in 2013 nicely outlined the theory: in severe scoliosis, there is a relative overgrowth of the anterior column. With maximal axial rotation correction that can be achieved with pedicle screws, there is no place to put the anterior spine unless a shortening is performed or if the posterior column is allowed to lengthen significantly. Even with posterior release via the Ponte osteotomy or wide posterior release without anterior shortening through the discs and vertebrae, Ruf et al hypothesized that ultimate correction and restoration of kyphosis will suffer. ⁵⁴ This is a common theme in modern scoliosis treatment. Kyphosis must be restored in order for lordosis to be restored in the lumbar spine. Anterior release is theorized to help. Ferrero et al in 2014 investigated anterior release for patients with hypokyphosis. A posterior fusion with a pedicle screw and sublaminar band construct focusing on posteromedial translation rather than correction of rotation was performed with or without anterior release. All patients had T4–T12 kyphosis less than 20 degrees preoperatively. The group with anterior release improved postoperative kyphosis from 11.7 to 18.3 degrees. The group without anterior release had kyphosis change from 12.1 to 15.2 degrees. The difference was not statistically significant. The authors noted that less focus on rotational correction and more focus on restoration of coronal and sagittal balance may have allowed for the lack of difference between groups. ⁵⁵

In 2004, Davis et al noted that traction radiography under general anesthesia could better evaluate curve flexibility and help eliminate the need for anterior release, as often patients’ curves were more responsive when the patient was anesthetized than the bending film exhibited. The authors noted that anterior release surgery was scheduled on 13 patients in the cohort with curves that did not bend to less than 40 degrees. ⁵⁶ These are more inclusive indications than most surgeons use today, but the traction radiograph remains in the arsenal to help evaluate rigidity. Larger curves are designated as severe in current literature. Curves of 70 degrees were an indication for anterior release in early studies. Now, severe curves are typically greater than 90 degrees and in some studies greater than 110 degrees. The options for treatment of these severe curves include traction as noted in Chapter 13, anterior release with or without intervening traction, or posterior osteotomies—either posterior column only or three-column osteotomies as noted in Chapter 15 Fig. 14‑1

The current literature for severe curves is frequently a mixed bag of techniques employed as well as diagnoses being treated. Many studies are from a single center, and accumulating enough patients to make a study meaningful is a challenge; certainly, finding enough patients to exhibit a comparison is difficult. Watanabe et al in 2010 reported on 21 patients with curves greater than 100 degrees. Ten were neuromuscular scoliosis, 9 idiopathic, and 2 congenital. Fifteen of the 21 patients underwent anterior release. All patients had halo-gravity traction. They found a 51% correction in the major Cobb angle and a 20% improvement in space for the lungs. The traction resulted in 27% improvement without release and 37% improvement with the release. ⁵⁷ Yamin et al in 2008 reported 21 cases with Cobb angle greater than 80 degrees treated with anterior release and halo-pelvic traction followed by posterior instrumented fusion often with a pedicle subtraction osteotomy. Mean preoperative Cobb angle was 110.5 degrees and the mean correction was 65.2%. There were two postoperative neurologic complications in patients with congenital scoliosis. The patients after release and traction demonstrated an average of 44% correction of their curve. ⁵⁸ Another mixed diagnosis population was reported by Qiu et al in 2007 with 30 out of 60 patients diagnosed with idiopathic scoliosis. In these patients, the average Cobb angle was 91.6 degrees preoperative with bending correction 24% on average. Correction after release and halo-femoral traction averaged 39.3%, and correction after posterior fusion averaged 57.5%. These patients were treated with hybrid posterior instrumentation. ⁵⁹ In 2016, Kandwal et al reported on 21 patients, 13 with idiopathic scoliosis who underwent anterior release and staged posterior fusion without traction between stages for Cobb angles greater than 100 degrees. After release, the Cobb angle improved from 116 to 74 degrees on average. Final postoperative Cobb angle was 26.5 degrees on average. An all pedicle screw construct was employed. ⁶⁰

There are some articles comparing techniques for severe curves. Shi et al published an article in 2015 comparing patients who underwent anterior release and posterior fusion versus patients who underwent posterior fusion alone with starting curve greater than 80 degrees. ⁶¹ They achieved similar results between groups in terms of radiographic correction with the anterior release group improving from 90 degrees on average to 35 degrees postoperatively, and the posterior fusion only patients improved from 88 to 31 degrees. They noted a higher implant density in the posterior fusion-only group and noted more instrumentation complications in that group. Despite ultimate correction being similar, they implied that more force distributed through the screws resulted in dangers with instrumentation and advocated for higher implant density to minimize the risks. In patients with a high risk of implant complication, they advocated for release to minimize the spinal anchors required. ⁶¹ Zhang et al in 2010 compared anterior release and 2 weeks of halo-femoral traction with posterior surgery alone with intraoperative halo-femoral traction. They found similar curve correction but prolonged operative time, blood loss, and hospital stay in the release/traction group compared to the single procedure. Hybrid instrumentation was employed with a minimal implant density. ⁶¹ In 2012, Koptan and ElMiligui compared anterior release and traction to an earlier series of patients who underwent staged anterior release and posterior fusion without intervening time in traction. Using a hybrid screw/hook/wire construct they achieved 59% correction in the group with traction as opposed to 47% in the group with release and fusion only. ⁶² Some of the same concepts discussed earlier, just with larger curves, are holding true—release and/or traction can provide a big boost in groups with large curves, especially with hybrid posterior instrumentation.

All pedicle screw instrumentation was utilized in Ren et al’s comparison of the anterior and posterior vertebral column resection versus anterior release with posterior internal distraction in 2014. The anterior/posterior vertebral column resection group was the earlier group and then conversion to anterior release and posterior fusion led to improved radiographic outcomes with 76.8% main thoracic Cobb angle correction compared to 68.3% in the vertebral column resection group. The anterior release patients, due to internal distraction which lasted 1 to 4 weeks with possible repeat operative trips to increase distraction, had longer hospital stays and patient costs. Operation time and blood loss were the same between groups. ⁶³

Dobbs et al in 2006 provided a near modern comparison: anterior release through open thoracotomy and posterior instrumentation versus posterior instrumentation alone for treatment of scoliosis greater than 90 degrees with all pedicle screw constructs. They reported no differences in postoperative Cobb values or correction with the combined surgery group improving from 92.3 to 36.9 degrees and the posterior-only group improving from 94.3 to 34.9 degrees. There was more flattening of the T5–T12 kyphosis in the posterior only group. They lost 11.6 degrees of kyphosis on average, while the combined group lost 5.4 degrees. The lung function was also monitored in this group who underwent open anterior release. The patients lost 3.1% of percent predicted FVC and 2.1% of percent predicted FEV-1 compared to the posterior only group that gained 5% of predicted FVC and 3.1% of predicted FEV-1. There were no reported complications or reoperations in either group. ⁶⁴ As the severity of the curve shifts higher in reports, the management involves a larger range of tools including release, traction, and osteotomies. Surgeons need to consider the possible pulmonary implications as well as cost and neurologic risk for the correction of these very large curves. Release still plays a role in these large and rigid curves, as the force required through the pedicle screws is diminished with increased flexibility and the sagittal plane correction noted earlier for Ruf et al and seen in the study of Dobbs et al. With the anterior release, the anterior column is shortened, especially if no structural graft is placed in the disc space. This allows for more shortening of the anterior column, as the powerful axial rotation correction of the pedicle screws brings the spine back to the appropriate plane.

14.6 Crankshaft

The continued growth of the anterior spine after posterior spinal fusion was first reported in 1973 by Dubousset et al. ⁸ The concept was confirmed after review in 1989 in Texas and Miami of patients who were Risser 0 at the time of posterior spinal fusion for idiopathic or paralytic scoliosis. Their review demonstrated progression of the mean Cobb angle from 37 degrees 1 year after fusion to 52 degrees at final follow-up. There was increase in the Perdriolle angle from 29.5 to 45 degrees. The patients fused at younger ages demonstrated more progression, and the only patient who had undergone an anterior and posterior fusion demonstrated no progression. In 1995, a total of 51 patients from Texas Scottish Rite with idiopathic scoliosis alone were evaluated by Sanders et al. They found evidence of crankshaft phenomenon in one patient with closed triradiate cartilage at the time of operation, while they found 10 out of 23 patients with open triradiate cartilage at the time of operation which showed evidence of the crankshaft phenomenon most appreciable with an increasing rib vertebral angle difference. ⁶⁵ Lapinksy and Richards also in 1995 from Texas Scottish Rite compared 10 patients with open triradiate cartilage who underwent anterior and posterior spinal fusions to 12 patients with open triradiate cartilage who underwent posterior fusion alone. There was a greater than 10-degree increase in Cobb angle in two patients and greater than 10-degree increase in rib vertebral angle difference in three patients who underwent anterior and posterior spine fusion. None of these changes happened in the same patient; so, it did not meet their definition of the crankshaft, which required both changes. In the group that underwent posterior fusion alone, there were five patients with a greater than 10-degree increase in Cobb and rib vertebral angle and therefore with the crankshaft. There were another four patients who had an increase in Cobb or rib vertebral angle difference. ⁶⁶ Burton et al questioned the need for anterior release in 2000 on a review of 18 Risser 0 patients. Seven had open triradiate cartilage, but only one patient exhibited an increase in Cobb angle or rib vertebral angle difference of greater than 10 degrees. ²⁷ They attributed their success to rigid instrumentation with mostly screw and hook constructs, although they noted only two patients had surgery prior to their peak height velocity. This was likely the most significant finding, but the authors favored the rigidity of the implants as the only case of crankshaft they experienced was treated with sublaminar wires which would be inadequate to their current standard. They noted pedicle screw instrumentation to be more rigid for the anterior column. A study with pedicle screws in immature dogs by Kioschos et al in 1996 demonstrated that pedicle screws could overpower the lordosis created with posterior fusion. The pedicle screw fusion group maintained similar spinal alignment to the control group, while the posterior fusion without instrumentation group generated nearly 25 degrees of lordosis. ⁶⁷

Tao et al noted no crankshaft occurrence in a group of 46 patients with open triradiate cartilage who underwent posterior only fusion with pedicle screw instrumentation compared to 7 cases (33%) in 21 patients with hybrid instrumentation. ⁶⁸ They advocated that pedicle screws limit the occurrence of crankshaft complications but caution to their small sample size. While the anterior release is not routinely performed merely for the mitigation of crankshaft in our patient populations when all pedicle screw instrumentation is used, the youngest patients especially with larger curves will not only have more flexible curves but an additional anterior fusion with a release.

14.7 Lack of Posterior Elements/Neuromuscular Scoliosis

One of the original reasons for the anterior approach to the spine was the lack of posterior elements. This is seen most commonly in myelomeningocele as well as conditions associated with dural ectasia such as Marfan syndrome and neurofibromatosis. These paralytic and syndromic causes of scoliosis are not the focus of this book, but in a chapter regarding anterior release they may still play an important role. Especially when coupled with a discussion about the crankshaft phenomenon, younger patients with limited ability to gain pedicle screw fixation may benefit from anterior fusion. Neuromuscular scoliosis, most frequently due to cerebral palsy, is also not a topic of this book, but early studies regarding the anterior approach to the spine featured a mixed group of diagnoses. Any discussion of large scoliosis curves will likely include neuromuscular, congenital, and syndromic scoliosis potentially more than idiopathic scoliosis. In the neuromuscular population, it is important to balance the morbidity of any additional surgery as the relative patient health is often poor. Patients with cerebral palsy in particular frequently have pulmonary involvement, as do patients with spinal muscular atrophy and Duchenne muscular dystrophy. To this end, Keeler et al in 2010 examined anterior release in patients with spastic neuromuscular scoliosis due to cerebral palsy and found similar correction of pelvic obliquity with intraoperative halo-femoral traction compared to anterior release and posterior instrumented fusion. The patients also had shortened OR time, lower estimated blood loss lower frequency of postoperative intubation, and fewer cases of pneumonia than the anterior and posterior combined group. ⁶⁹ Similarly the Harms Study Group published in 2018 from the prospective cerebral palsy cohort that intraoperative traction may be a viable alternative to anterior surgery even in patients with curves greater than 100 degrees. The intraoperative traction group had improved correction of pelvic obliquity (33.5 vs. 14 degrees). In this study, patients with anterior/posterior combined fusion had a similar hospital length of stay, similar intensive care unit and days intubated, similar estimated blood loss, cell savers, and red blood cell volumes transfused. ⁷⁰ Patients with very large scoliosis often have an underlying disorder responsible for their scoliosis that will also carry additional comorbidities for surgery. It is important to thoroughly investigate these possibilities and understand the surgical consequences of the approach taken to deformity correction prior to proceeding with combined surgical approaches.