Secondary Logo

Journal Logo

The Feasibility of Anterior Thoracoscopic Spine Surgery in Children Under 30 Kilograms

Early, Sean D., MD*; Newton, Peter O., MD; White, Klane K., MD; Wenger, Dennis R., MD; Mubarak, Scott J., MD

Technique
Free

Study Design. A retrospective comparison of pediatric patients weighing less than 30 kg who underwent thoracoscopic anterior spinal release and fusion for deformity correction. This group was compared to two control groups: patients weighing over 30 kg (thoracoscopic) and patients under 30 kg (open).

Objective. To determine the efficacy and safety of thoracoscopic anterior spinal release and fusion in small pediatric patients weighing less than 30 kg.

Summary of Background Data. Recently, thoracoscopic methods have been utilized to perform anterior spinal release/fusion in the treatment of pediatric and adult spinal deformity. The safety, efficacy, and technical challenges of thoracoscopic spinal surgery in small children have not been established.

Methods. “Small thoracoscopic children,” defined as those under 30 kg who had thoracoscopic spinal surgery, are the main focus of this study. They were compared to “large thoracoscopic children” (>30 kg, thoracoscopic surgery) and “small open children” (<30 kg, open surgery). Preoperative, intraoperative, and postoperative parameters were analyzed.

Results. Small thoracoscopic children (n = 33) had greater estimated blood loss/kg body weight (13.6 mL/kg vs. 6.2 mL/kg;P = 0.003), greater chest tube output (27.5 mL/kg vs. 17.1 mL/kg;P = 0.003), and a longer intensive care unit stay (4.2 days vs. 1.5 days;P = 0.001) than did large thoracoscopic children (n = 48). Conversion to an open thoracotomy occurred in one patient from each of the thoracoscopic groups. Small thoracoscopic children required more anesthesia preparation time (79.2 minutes vs. 64.2 minutes;P = 0.002) than the small open children (n = 25). There was no significant difference in estimated blood loss, chest tube output, or intensive care unit stay between these two groups. Additionally, no significant difference was found between the three groups with regard to the number of discs excised, operative time, and total hospital stay.

Conclusion. Despite the decreased working space within the chest and difficulties of selective intubation, anterior thoracoscopic surgery for spinal release and fusion can be performed as safely in “small” children as in “large” children; however, additional intraoperative challenges should be anticipated. Although the outcomes were similar in the small thoracoscopic children compared to the small open children, the authors believe that very small patients (under 20 kg) should remain a relative contraindication to thoracoscopic surgery, especially during a surgeon’s learning curve.

From *Children’s Hospital, Los Angeles, California, and

†Children’s Hospital and Health Center, San Diego, California.

This project was supported by the Children’s Hospital–San Diego Orthopedic Research and Education Fund.

Acknowledgment date: October 2, 2001.

First revision date: April 17, 2002.

Acceptance date: June 7, 2002.

Address correspondence and reprint requests to

Peter O. Newton, MD

3030 Children’s Way, Suite 410

San Diego, CA 92123-4293

E-mail:pnewton@chsd.org

The manuscript submitted does not contain information about medical device(s)/drug(s).

No funds were received in support of this work. No benefits in any form have been or will be received from a commercial party related directly or indirectly to the subject of this manuscript.

There are a number of indications for anterior approach to the thoracic spine for correction of spinal deformity. In patients with severe, rigid scoliosis or kyphosis, anterior longitudinal ligament release and intervertebral discectomy improve mobility of the spine and allow greater surgical deformity correction. 6 Anterior bone grafting decreases the pseudarthrosis rate and prevents deformity recurrence after posterior instrumentation and fusion. 18 In immature patients, anterior fusion halts anterior spinal growth, preventing development of the crankshaft phenomenon that may be seen with isolated posterior spinal fusion. 4

Traditionally, an anterior approach to the thoracic spine has been performed by open thoracotomy. With recent technological advances, thoracoscopy has allowed video-assisted anterior spinal release and fusion with minimally invasive methods. Although the injury to the chest wall is less, thoracoscopy affords the same surgery on the spine as does open thoracotomy. In both caprine and porcine models, the thoracoscopic and open techniques were equally effective in increasing spinal flexibility after anterior release and discectomy. 12,21 In a sheep model, thoracoscopic interbody spinal fusion demonstrated histologic, biomechanical, and radiographic equivalence when compared with fusion by open thoracotomy. 3 Clinically, thoracoscopic and open techniques allowed similar correction of both scoliosis and kyphosis without an increase in complication rates. 14

Due to its less invasive nature, thoracoscopic surgery affords a number of advantages over open thoracotomy. By avoiding extensive muscle dissection and rib resection, thoracoscopy allows decreased postoperative pain and narcotic requirements when compared to open thoracotomy. 5 This may improve postoperative pulmonary function. With less skin, serratus anterior, and latissimus dorsi division, thoracoscopy is expected to allow a more rapid return of normal shoulder motion and strength. In a study comparing thoracoscopic to limited open pulmonary resection, Landreneau et al demonstrated less pain, less shoulder dysfunction, and reduced early pulmonary impairment in the thoracoscopic group. 9 It is believed that such decreased morbidity is also present in patients undergoing thoracoscopic spine surgery. 14,17,19,20

Despite the increasing acceptance of thoracoscopic techniques, challenges exist in its use for the treatment of small children. Video-assisted thoracoscopic surgery takes advantage of the space afforded by the thoracic cage. With single lung ventilation of the nonoperated hemithorax, a large working space can often be created. However, patient size is an important factor in determining the difficulty of the procedure. The diminutive bronchial tree in young children necessitates special methods of single lung ventilation. 13 In addition, the small chest cavity, narrow rib spacing, and reduced working distance to the spine create additional technical challenges by decreasing the volume available for the thoracoscope and necessary surgical instruments. 14 It was hypothesized that this makes performance of discectomy and fusion more difficult in small children than in adolescents and adults.

The purpose of this study was to evaluate the safety and efficacy of thoracoscopic spine surgery in small children while also identifying the technical and anesthetic challenges that should be anticipated.

Back to Top | Article Outline

Methods

This study was an Institutional Review Board-approved, retrospective analysis of pediatric patients who underwent anterior thoracic spinal release and fusion for deformity correction at our institution. Patients who underwent anterior thoracic instrumentation were excluded. The focus group, small thoracoscopic children, comprised all patients weighing less than 30 kg who had thoracoscopic anterior spinal release and fusion between 1994, when this technique was first introduced at our institution, and 2000. There were two comparison groups. The large thoracoscopic children were defined as those patients weighing over 30 kg who underwent anterior thoracoscopic spinal release and fusion between 1994 and 1998. The data from this group were included in a previous publication. 13 The small open children group comprised all children weighing less than 30 kg who had open thoracotomy, without diaphragm division, for anterior thoracic spine surgery at our institution from 1991 to 2000.

The choice of 30 kg as the division between “small” and “large” children was somewhat arbitrary but based on the frequency distribution of patients’ weight in children treated thoracoscopically at our institution over time. The <30 kg cutoff for small children represents approximately the lower 25% of cases by patient weight. It is also roughly the weight at which our anesthesia service believes single lung ventilation becomes more challenging.

Retrospective data collection was identical for each group. Age, weight, gender, and diagnosis were recorded. Intraoperative parameters analyzed for the anterior thoracic spine surgery included anesthesia preparation time, quality of lung collapse, number of discs excised, normalized estimated blood loss (mL/kg body weight), and total operative time. Quality of lung collapse was obtained from the dictated operative report. It was graded as satisfactory or poor. Any difficulty with spinal visualization secondary to the lung, which was not easily remedied, was considered poor lung collapse. Anesthesia preparation time was defined as the interval from entrance into the operating suite until the start of the surgery. Total operative time was defined as the interval from the start of the anterior thoracic surgery until its completion. Both time parameters were directly extracted from the intraoperative anesthesia record. All patients had either staged or same-day posterior surgery. The posterior procedures included arthrodesis with or without a variety of posterior implant systems, depending on the nature of the deformity. The data from the posterior procedure were not included in this analysis. Three patients also had endoscopic anterior retroperitoneal lumbar surgery following the thoracoscopic procedure. Data for this lumbar procedure were not included.

Postoperative parameters that were analyzed included the duration of mechanical ventilation, normalized chest tube output (mL/kg of body weight), chest tube duration, intensive care unit (ICU) stay duration, and total length of the hospital stay. All complications were recorded.

Preoperative and initial erect postoperative radiographs were analyzed for each of the three groups. Standard Cobb angle measurements were used for patients with scoliosis. Mean values for scoliosis and percent correction of scoliosis were calculated using only the data from patients whose scoliosis necessitated surgical treatment. Patients with predominantly kyphotic deformity and those that underwent an in situ arthrodesis for scoliosis were not included in the analysis of percent scoliosis correction. Sagittal plane alignment was measured from T2 to T12. 2 Mean values for kyphosis and percent correction of kyphosis were calculated using only the data from patients whose sagittal plane alignment necessitated surgical treatment.

Back to Top | Article Outline

Statistical Analysis.

Multiple analyses of variance (ANOVA) were performed. Statistical comparisons of large thoracoscopic children compared to small thoracoscopic children and small thoracoscopic children compared to small open children were analyzed using simple contrasts with the small thoracoscopic children as the reference category. The simple contrast reports an F value (such as that used in the ANOVA), Bonferroni-type P values, and simultaneous confidence intervals based on Student t distribution. Levene’s test for homogeneity of variances and normal Q-Q plots were run to ensure that variances were equal and the data were normally distributed. For all variables except chest tube days, the data were either nonnormal or the variances were unequal. The nonparametric median test was run for all of these variables, and the results did not deviate from the F statistic reported with the simple contrasts. Therefore, the values of the contrasts were reported. Two Pearson correlation coefficients were calculated to identify potential relationships among: anterior surgical time and blood loss; and ICU stay and quality of lung collapse. The nonparametric alternative to the χ2 statistic, Fisher’s exact test, was run with quality of lung collapse (satisfactory vs. poor) and both thoracoscopic surgical groups (small thoracoscopic children and large thoracoscopic children) as the variables. The overall alpha level was set at P = 0.05 and adjusted by Bonferroni correction to P = 0.004.

Back to Top | Article Outline

Results

Inclusion criteria were met by 106 patients. The small thoracoscopic children included 33 patients (31%) who underwent anterior thoracoscopic spinal surgery and weighed less than 30 kg. Thirteen were male and 20 female. Mean age was 10.7 years, and the mean weight was 22.8 kg. Two patients had idiopathic scoliosis (6%), 25 neuromuscular scoliosis (76%), and 6 congenital scoliosis (18%). The small open children included 25 patients (24%) who weighed less than 30 kg and underwent open anterior spinal surgery. Ten were male and 15 female. Mean age was 7.8 years, and the mean weight was 19.4 kg. Fourteen patients had neuromuscular scoliosis (56%), 10 congenital scoliosis (40%), and 1 postlaminectomy kyphosis (4%). The large thoracoscopic children included 48 patients (45%) who weighed over 30 kg and underwent anterior thoracoscopic spinal surgery. There were 23 males and 25 females. Mean age was 14.7 years, and the mean weight was 51.1 kg. Thirteen patients had idiopathic scoliosis (27%), 9 idiopathic kyphosis (19%), 20 neuromuscular scoliosis (42%), 2 congenital scoliosis (4%), and 4 a tumor or syrinx (8%). The number of discs excised, total operative time, duration of mechanical ventilation, chest tube duration, and average hospital stay did not differ significantly between the three groups (Table 1). The number of discs excised was significantly greater in the neuromuscular group compared to the nonneuromuscular group (P = 0.001;Table 2). Hospital stay was also significantly greater in the neuromuscular group compared to the nonneuromuscular group (P = 0.001) (Table 2). However, the total operative time, duration of mechanical ventilation, and chest tube duration did not differ between these two groups (Table 2). There were no significant interactions between weight/surgical procedure and diagnosis for any of the following variables: number of discs excised, total operative time, duration of mechanical ventilation, chest tube duration, and average hospital stay. Quality of lung collapse (satisfactory or poor) did not differ between the two thoracoscopic surgery groups, although there was a trend toward more difficulty in the small thoracoscopic children (P = 0.06).

Table 1

Table 1

Table 2

Table 2

Back to Top | Article Outline

Blood Loss

The mean estimated blood loss per kg of body weight was significantly greater for the small thoracoscopic children compared to the large thoracoscopic children (P = 0.003) (Table 1). There was no significant difference in normalized blood loss between the small thoracoscopic children (13.6 mL/kg) and the small open children (16.3 mL/kg)(P = 0.4). There was no significant difference in estimated blood loss between children with a neuromuscular and nonneuromuscular diagnosis (P = 0.5;Table 2). The interaction between weight/surgical procedure and diagnosis was also not significant (P = 0.08). Examining this interaction further, there was a nonsignificant trend for neuromuscular patients to lose more blood than nonneuromuscular patients when the surgery was thoracoscopic; however, when the surgery was an open thoracotomy, neuromuscular patients were found to lose less blood than nonneuromuscular patients. A significant positive correlation was found between estimated blood loss and operative time for the small thoracoscopic children (P = 0.003), but this relationship was not significant for the other two groups (P > 0.10).

Back to Top | Article Outline

Chest Tube Output

The small thoracoscopic children had greater normalized chest tube output compared to the large thoracoscopic children (P = 0.003;Table 1). Chest tube output per kg of body weight did not differ significantly between small open children and small thoracoscopic children (P = 0.5).

Back to Top | Article Outline

ICU Days

Mean ICU stay was significantly greater in the small thoracoscopic children compared to the large thoracoscopic children (P < 0.001;Table 1). The average stay for small open children was not significantly different from the small thoracoscopic children (P = 0.1). Mean ICU stay was significantly greater for neuromuscular patients compared to nonneuromuscular patients (P < 0.001;Table 2). The interaction was not significant (P = 0.05); however, this nonsignificant trend indicated that neuromuscular patients had an increased ICU stay when the procedure was a thoracoscopic procedure compared to nonneuromuscular patients who underwent a thoracoscopic procedure. Intensive care unit stay for all diagnoses were similar when the procedure was an open thoracotomy.

Back to Top | Article Outline

Anesthesia Preparation Time

The mean time from induction of anesthesia until the start of surgery was significantly greater for the small thoracoscopic children compared to the small open children (P = 0.002;Table 1). There were no differences between the small thoracoscopic children and the large thoracoscopic children (P = 0.2).

Back to Top | Article Outline

Radiographic Measures

The average preoperative scoliosis for the three groups was similar (small thoracoscopic children: 62°± 12°; large thoracoscopic children: 56°± 20°; small open children: 66°± 29°). The average postoperative percent scoliosis correction (for those patients in which correction was attempted) was 53%, 62%, and 37%, respectively (P = 0.06). The average preoperative kyphosis of these patients with surgically indicated hyperkyphosis was 72°± 29°, 71°± 22°, and 76°± 22° for the small thoracoscopic, large thoracoscopic, and small open groups, respectively. For the two thoracoscopic groups, the percent correction, calculated as: [preop − (postop − 40)/preop] × 100, was 95% ± 8% for the small group and 94% ± 17% for the large group. The sagittal correction of the small open group was 68% ± 9% (P = 0.04).

Back to Top | Article Outline

Complications

There were three complications in the small thoracoscopic children. Surgery was rescheduled in one 24-kg girl with neuromuscular kyphoscoliosis due to inability to selectively intubate her lungs. She was only the third patient to undergo thoracoscopic spine surgery at our institution. The procedure was successfully completed at the later date utilizing a balloon bronchial blocker. There was one conversion to open thoracotomy in a 24-kg girl with neuromuscular kyphosis. Despite repeated repositioning of a balloon blocker, there was intermittent reexpansion of the lung that prevented proper visualization. One 20-kg girl with neuromuscular scoliosis developed a pleural effusion after chest tube removal. The effusion resolved with reinsertion of the chest tube on postoperative day 6.

There were three complications in the large thoracoscopic children. One 35-kg girl with myotonic dystrophy and kyphoscoliosis had conversion to open thoracotomy due to excessive bone bleeding, which prevented adequate visualization of the spine. There was no vessel injury. One 50-kg girl with idiopathic scoliosis developed a pleural effusion after chest tube removal. Despite thoracentesis, the effusion reaccumulated, necessitating chest tube reinsertion. One 36-kg girl with juvenile idiopathic scoliosis developed chylothorax after surgery. This resolved with a low-fat diet. No further invasive measures were needed.

There was only one complication in the small open group. One 18-kg girl with neuromuscular scoliosis had excessive bleeding causing postponement of her posterior procedure. There were no neurologic complications in any of the three study groups.

Seven large thoracoscopic children, 8 small thoracoscopic children, and 3 small open children remained in the ICU for prolonged postoperative mechanical ventilation of greater than 72 hours. One of these children had scoliosis associated with a large syrinx, and the remaining children had neuromuscular scoliosis.

Back to Top | Article Outline

Discussion

The authors have demonstrated the safety and efficacy of spinal thoracoscopy in a group of children (n = 33) weighing less than 30 kg. A comparison to larger children treated thoracoscopically as well as similarly sized patients treated by an open thoracotomy allowed differentiation of the effects of patient size, as well as the method of exposure–thoracoscopic versus open.

There were a number of similarities between the study groups that validate the comparisons made. The study design aimed to compare differing patient groups who had the same extent of surgery. Only patients who underwent anterior release and fusion were included in the study. None of the patients had anterior spinal instrumentation. The small open children group was limited to patients who had thoracic surgery only. Any patient with division of the diaphragm was excluded in order to maintain maximal uniformity of the surgical procedures between the open and thoracoscopic groups. The extent of the surgery did prove similar, as there was no difference in the mean operative time or mean number of discs excised between the three groups. Preoperative spinal deformity was similar in the three groups. With patients undergoing in situ fusion (congenital scoliosis) omitted from analysis, the percent deformity correction was also similar in the three groups. All thoracoscopic procedures were performed by a single surgeon. This eliminates intersurgeon variability when comparing small thoracoscopic children and large thoracoscopic children, in a technically demanding procedure with a well-documented learning curve. 7,11,13,15 The small thoracoscopic children and small open children had similar mean body weights of 22.8 kg and 19.4 kg, respectively.

Despite efforts to standardize our study groups, there were notable differences. The mean age varied between the two small children (7.8 years, open vs. 10.7 years, thoracoscopic) groups. This age difference may reflect some selection bias, as very young children may have been chosen for open thoracotomy rather than thoracoscopic surgery. However, many of the small open children were operated before our institution’s use of thoracoscopic techniques and account for the differences in the length of follow-up. The number of patients varied among the three groups. Only 25 patients in the computer database, which began in 1991, met the under-30 kg inclusion criteria for the small open children. Since the thoracoscopic technique has been used at our institution, the vast majority of patients who need anterior thoracic release and fusion are treated by the less invasive thoracoscopic method. The group of large thoracoscopic children represent 48 patients from a previously published study who weighed over 30 kg. 13 The remaining 17 patients from that study are included in the group of 33 small thoracoscopic children. The small thoracoscopic children represent approximately one-fourth of all patients who have undergone thoracoscopic anterior release and fusion at our institution since 1994. The predominant diagnoses in the three groups differed slightly. In the large thoracoscopic children, there was a nearly equal division between patients with neuromuscular disease (42%) and idiopathic scoliosis or kyphosis (46%). In the small open children, there were a similar number of patients with neuromuscular (56%) and congenital (40%) scoliosis. The vast majority of the small thoracoscopic children had neuromuscular scoliosis (76%). This lack of congenital scoliosis patients reflects the authors’ impression that, when two to three levels require treatment (as is often the case in congenital deformities), a limited open approach is simpler and preferred.

Four of the small open children and one patient from the group of small thoracoscopic children had staged posterior procedures. The results of analysis of the postoperative variables (time intubated, chest tube output, chest tube duration, ICU stay, and total hospital stay) did not differ with exclusion of the data from these five patients. Thus, their data were included in the study. This is also true for the three patients who had endoscopic anterior retroperitoneal lumbar surgery following their thoracoscopic procedure.

The effectiveness of curve correction varied among the three groups. Mean correction of scoliosis was 54% for the small thoracoscopic children and 62% for the large thoracoscopic children. This is similar to previous reports, where mean correction of scoliosis ranged from 42% to 77%. 8,10,11,15,20 Arlet 1 reported a 60% mean correction of scoliosis in a meta-analysis and review of 10 studies on thoracoscopic spine surgery. Mean correction of kyphosis was 95% for the small thoracoscopic children and 94% for the large thoracoscopic children in the present study. Niemeyer et al15 reported a 77% correction of kyphosis whereas King et al8 reported an 87% correction. The mean scoliosis correction for the small open children was 38%, increasing to 44% with the omission of patients with congenital scoliosis in whom in situ fusion was performed. Kyphosis correction averaged 68% in this group. In many of the small open children, postoperative casting, rather than posterior instrumentation, was used and is likely the reason for the lower percentage correction.

Normalized estimated blood loss correlated more with patient age and size than with the surgical technique employed. Both the small thoracoscopic children and small open children had over twice the volume of blood loss, per unit weight, than did the large thoracoscopic children. There was no difference between the small thoracoscopic children and the small open children, thus the method of spinal access does not explain the differential blood loss. Although the small thoracoscopic children had a high percentage of neuromuscular patients, it is unlikely that variability in diagnosis was a major factor in the blood loss differences between the thoracoscopic groups; both the small open children and the large thoracoscopic children had similar numbers of neuromuscular patients yet very different normalized blood loss values. Instead, age and vertebral body development are more important. Whalen et al22 showed that although the disc is avascular, young children have numerous arterioles and venules in their cartilaginous endplates. Ratcliffe 16 demonstrated extensive venous and arteriole anastomoses throughout the vertebral bodies of younger children. These vessels involute near age 15 years, the mean age of the large thoracoscopic children in our study. The bony endplates are thinner in young children, compared to older adolescents. This may allow more frequent entrance into the cancellous bone of the vertebra during disc and cartilaginous endplate excision, in these immature patients. Thus, in younger children, more blood loss should be expected following discectomy and endplate excision given this increased vascularity of the vertebral bodies.

The anesthesia preparation time was greater in the two thoracoscopic groups when compared to the small open children. This statistically significant difference was expected given the specialized methods of single lung ventilation required for thoracoscopic surgery. The small thoracoscopic children and large thoracoscopic children had statistically similar anesthesia preparation times and quality of lung collapse. More difficulty was anticipated with the small thoracoscopic children, and although trends supporting these ideas did exist, statistical significance was not reached. In the vast majority of large thoracoscopic children, a Univent or double lumen endotracheal tube readily allowed selective lung ventilation. These tubes are often too large for smaller children, necessitating specialized methods. This usually involves a small balloon catheter, such as a Foley, Fogarty, or balloon removed from the Univent tube that is placed adjacent to a standard endotracheal tube. The balloon is inflated in the mainstem bronchus on the operative side, allowing selective ventilation of the opposite lung. Given the number of thoracoscopic procedures performed at the authors’ institution, the anesthesiologists have become quite facile with this technique. Although one patient from each group required conversion to open thoracotomy, there was no statistically significant difference in the quality of lung collapse between the two thoracoscopic groups. At the authors’ institution, little extra time is required when standard double lumen endotracheal tubes cannot be used.

The benefits of thoracoscopy relate largely to the limited chest wall dissection that is associated with this approach. The lengths of the portal incisions are typically 15 mm. Thus, the total incision length varies between 4.5 and 7.5 cm, depending on the number of portals required. Although difficult to quantitate, it is the authors’ impression that as the size of the patient decreases, the relative benefit of the thoracoscopic approach also decreases. This is because the summed length of the portal incisions approaches that of a thoracotomy for a small child. This is especially true if only two or three levels of the spine require treatment (e.g., hemiarthrodesis for congenital hemivertebra). The smallest child in the thoracoscopic group (age: 1 year and 8 months, weight: 10 kg) had congenital scoliosis in which a hemiepiphysiodesis was performed over three levels. The benefits of thoracoscopy, compared to thoracotomy, for such an exposure are limited and likely not worth the effort. Although there were similar outcomes in the small and large children thoracoscopic groups, the authors believe that patient size under 20 kg should remain a relative contraindication to thoracoscopic surgery, especially during a surgeon’s learning curve.

Back to Top | Article Outline

Conclusion

Despite the decreased working space within the chest and difficulties of selective intubation, anterior thoracoscopic surgery for spinal release and fusion can be performed safely in small children. Intraoperative challenges associated with increased blood loss (mL/kg), as well as increased chest tube output (mL/kg), longer ICU stay, and the need for specialized intubation methods should be anticipated.

Back to Top | Article Outline

Key Points

  • Thoracoscopic anterior spinal surgery can be safely and effectively performed in children weighing less than 30 kg.
  • In these small children, greater blood loss per kg body weight should be anticipated.
  • Specialized methods of intubation for single lung ventilation may be required.
Back to Top | Article Outline

References

1. Arlet V. Anterior thoracoscopic spine release in deformity surgery: a meta-analysis and review. Eur Spine J 2000; 9: S17–23.
2. Boseker EH, Moe JH, Winter RB, et al. Determination of “normal” thoracic kyphosis: a roentgenographic study of 121 “normal” children. J Pediatr Orthop 2000; 20: 796–8.
3. Cunningham BW, Kotani Y, McNulty PS, et al. Video-assisted thoracoscopic surgery versus open thoracotomy for anterior thoracic spinal fusion. A comparative radiographic, biomechanical, and histologic analysis in a sheep model. Spine 1998; 23: 1333–40.
4. Dubousset J, Herring JA, Shufflebarger H. The crankshaft phenomenon. J Pediatr Orthop 1989; 9: 541–50.
5. Hazelrigg SR, Landreneau RJ, Boley TM, et al. The effect of muscle-sparing versus standard posterolateral thoracotomy on pulmonary function, muscle strength, and postoperative pain. J Thorac Cardiovasc Surg 1991;101:394–400; discussion 400–1.
6. Herring A. Anterior Spine Surgery. In: Weinstein S, ed. The Pediatric Spine. Practice and Principles. New York, NY: Raven Press; 1994; 1409–25.
7. Kaiser LR, Bavaria JE. Complications of thoracoscopy. Ann Thorac Surg 1993; 56: 796–8.
8. King AG, Mills TE, Loe WA Jr, et al. Video-assisted thoracoscopic surgery in the prone position. Spine 2000; 25: 2403–6.
9. Landreneau RJ, Hazelrigg SR, Mack MJ, et al. Postoperative pain-related morbidity: video-assisted thoracic surgery versus thoracotomy. Ann Thorac Surg 1993; 56: 1285–9.
10. Lapinksy AS, Richards BS. Preventing the crankshaft phenomenon by combining anterior fusion with posterior instrumentation. Does it work? Spine 1995; 20: 1392–8.
11. Mack MJ, Regan JJ, McAfee PC, et al. Video-assisted thoracic surgery for the anterior approach to the thoracic spine. Ann Thorac Surg 1995; 59: 1100–6.
12. Newton PO, Cardelia JM, Farnsworth CL, et al. A biomechanical comparison of open and thoracoscopic anterior spinal release in a goat model. Spine 1998;23:530–5; discussion 536.
13. Newton PO, Shea KG, Granlund KF. Defining the pediatric spinal thoracoscopy learning curve: sixty-five consecutive cases. Spine 2000; 25: 1028–35.
14. Newton PO, Wenger DR, Mubarak SJ, et al. Anterior release and fusion in pediatric spinal deformity. A comparison of early outcome and cost of thoracoscopic and open thoracotomy approaches. Spine 1997; 22: 1398–406.
15. Niemeyer T, Freeman BJ, Grevitt MP, et al. Anterior thoracoscopic surgery followed by posterior instrumentation and fusion in spinal deformity. Eur Spine J 2000; 9: 499–504.
16. Ratcliffe JF. Anatomic basis for the pathogenesis and radiologic features of vertebral osteomyelitis and its differentiation from childhood discitis. A microarteriographic investigation. Acta Radiol Diagn (Stockh) 1985; 26: 137–43.
17. Regan JJ, Mack MJ, Picetti GD3rd. A technical report on video-assisted thoracoscopy in thoracic spinal surgery. Preliminary description. Spine 1995; 20: 831–7.
18. Rinsky LA. Surgery of spinal deformity in cerebral palsy. Twelve years in the evolution of scoliosis management. Clin Orthop 1990; 100–9.
19. Rothenberg S, Erickson M, Eilert R, et al. Thoracoscopic anterior spinal procedures in children. J Pediatr Surg 1998;33:1168–70; discussion 1170–1.
20. Waisman M, Saute M. Thoracoscopic spine release before posterior instrumentation in scoliosis. Clin Orthop 1997; 130–6.
21. Wall EJ, Bylski-Austrow DI, Shelton FS, et al. Endoscopic discectomy increases thoracic spine flexibility as effectively as open discectomy. A mechanical study in a porcine model. Spine 1998;23:9–15; discussion 15–16.
22. Whalen JL, Parke WW, Mazur JM, et al. The intrinsic vasculature of developing vertebral end plates and its nutritive significance to the intervertebral discs. J Pediatr Orthop 1985; 5: 403–10.
Keywords:

thoracoscopy; scoliosis; spine deformity; endoscopy; pediatric]Spine 2002;27:2368–2373

© 2002 Lippincott Williams & Wilkins, Inc.