Spinal Deformity After Combined Thoracotomy and Sternotomy for Congenital Hea

weisskopf

[B]Spinal Deformity After Combined Thoracotomy and Sternotomy for Congenital Heart Disease[/SIZE][/B]
Herrera-Soto, Jose Antonio MD; Vander Have, Kelly L. MD; Barry-Lane, Patricia MD; Woo, Alice

Abstract
Abstract: Patients with congenital heart disease are at an increased risk to develop scoliosis. The purpose of this study was to determine the incidence of spinal deformity in patients after thoracotomy and sternotomy for congenital heart disease.

Methods: Sixty-eight patients underwent thoracotomy followed by a sternotomy and met inclusion criteria. The medical records were reviewed to gather demographic data and medical and surgical history. Serial radiographs were reviewed.

Results: Scoliosis developed in 26% of the patients (10 boys, 8 girls). The mean Cobb angle was 40 degrees (range, 15-78 degrees). The mean age at diagnosis of scoliosis was 10.7 years (range, 2.9-17 years). The mean follow-up was 14.9 years (range, 5-20 years). Twelve percent (8 patients) required posterior spinal fusion. A kyphotic deformity developed in 21% (14 patients). In patients with scoliosis, the mean kyphosis was 38 degrees (range, 2-88 degrees). Patients with a cyanotic cardiac condition had a 4-fold incidence of scoliosis. There was no correlation between the development of scoliosis or kyphosis and the age at time of procedures, number of surgeries, sex, heart size, or side of the aortic arch.

Conclusions: The risk of developing scoliosis in children with congenital heart disease is more than 10 times that of idiopathic scoliosis. Spinal deformities, including scoliosis and/or hyperkyphosis, were found in 38% of the patients. Curves develop at a younger age, which increases the risk of progression. The sagittal alignment in scoliosis patients tends toward hyperkyphosis. The thoracic spine receives a "double hit" when both procedures are combined.

Congenital heart disease occurs in about 1% of live-born infants.1 Its etiology is considered to be multifactorial and, depending on the diagnosis congenital heart disease, is surgically addressed through a thoracotomy, sternotomy, or both. Patients who undergo a thoracotomy are at an increased risk to develop scoliosis.2-14 The etiology of scoliosis in these patients remains controversial.13

Previous reports of scoliosis in congenital heart disease have included patients who underwent thoracotomy and, in a small percentage of the cases, combined thoracotomy and sternotomy. In fact, some reports present the rate of scoliosis after any type of open-heart surgery for congenital heart disease, thus mixing the patient population.5,6,11 The results may be inconsistent because the incidence of scoliosis after a specific approach is not clear. Typically, the focus has been on thoracotomy as the cause of scoliosis.2,6,9,10,14,15 However, there was an increased incidence of scoliosis (28%) and hyperkyphosis (22%) in patients who underwent sternotomy only for the treatment of congenital heart disease.16 To the best of our knowledge, there are no formal reports in the literature on the incidence of spinal deformities (scoliosis or hyperkyphosis) after combined thoracotomy and sternotomy or the treatment of congenital heart disease. We hypothesized that the thoracic cage takes a "double hit" after thoracotomy and sternotomy, contributing to both coronal and sagittal deformity. The purpose of this study was to determine the incidence of spinal deformity in patients with congenital heart disease treated with thoracotomy and sternotomy.

METHODS
A total of 910 patients underwent open-heart surgery for congenital heart disease at our institution. Seventy-eight patients underwent combined thoracotomy and sternotomy at different times to treat 2 cardiac conditions that required different approaches. Ten patients were lost to follow-up, died, or had insufficient documentation. After institutional review board approval was obtained, a retrospective review of 68 patients with sufficient data and radiographic follow-up was conducted.

The medical records were reviewed, including demographic, medical, and surgical data. Of note, our cardiothoracic surgeon does not resect ribs during the thoracotomy. Patients were subgrouped based on whether they had a cyanotic or acyanotic cardiac condition.1 If one of the cardiac diagnoses included a cyanotic lesion, they were classified as cyanotic. Cyanotic conditions included tricuspid and pulmonary atresia, tetralogy of Fallot, hypoplastic right or left ventricles, and transposition of the great vessels. Acyanotic conditions included atrial and ventricular septal defects, coarctation of the aorta, and patent ductus arteriosus.

Most of the radiographs examined were standing posteroanterior and lateral chest films. These radiographs were standard for patient follow-up during cardiology visits. All posteroanterior chest films included at least L2 or lower. If no inclination was seen, the lumbar spine was considered straight. When available, standing anteroposterior and lateral 36-in spine films were used preferentially. Spinal deformity was measured on both radiographs using the Cobb method.17 Scoliosis was graded as mild (11-20 degrees), moderate (21-30 degrees), or severe (>30 degrees).4-7,9,11 The curves were classified as high thoracic, thoracic, and lumbar, and the convexity of the curve was recorded as either left or right. Kyphosis was measured on the lateral films (T2-T12) and graded as hypokyphotic (<=20 degrees), normal kyphosis (21-40 degrees), or hyperkyphotic (>40 degrees).18 A radiologist used the chest films to classify the heart size as normal or enlarged and the aortic arch as left or right.

RESULTS
Sixty-eight patients (38 boys and 30 girls) were studied. All but one patient underwent thoracotomy first, followed by a sternotomy. The mean follow-up was 14.3 years (5-30 years) after the first procedure.

Scoliosis
Eighteen patients, 10 boys and 8 girls, developed scoliosis (26%). The mean age at the time of the thoracotomy for those with scoliosis was 12 months (range, 0-63 months). No patient had scoliosis before the first surgery. In those patients with scoliosis, the sternotomy was performed at an average of 35.2 months (range, 1-142 months). Only one patient was found to have scoliosis before the second surgery. Scoliosis was detected at a mean age of 10.7 years (range, 2.9-17.5 years). On average, scoliosis was detected 7 years after the sternotomy. Patients with scoliosis were followed up at an average of 14.9 years (range, 1.5-25.7 years). These results are illustrated in detail in Table 1. There were a total of 31 curves identified in 18 patients. They were distributed as 11 right thoracic curves, 4 left thoracic, 7 lumbar, and 9 upper thoracic curves.

Scoliosis was classified as mild in 6 patients, moderate in 1 patient, and severe in 11 patients. The mean Cobb angle of the main coronal curve was 40 degrees (range, 15-78 degrees). Sixty-seven percent (12/18) of patients with scoliosis had moderate to severe deformity. Eight of the patients with severe scoliosis required posterior spinal fusion to prevent progression. For those patients with severe curves, the mean age at detection of scoliosis was 8.5 years (range, 2.9-15.6 years). There was a tendency for this group of patients to develop scoliosis at a younger age, although not statistically significant (P = 0.07; Fig. 1).

[Help with image viewing]
[Email Jumpstart To Image]  FIGURE 1. A 12-year-old girl with Ebstein anomaly and atrial septal defects presented with a 70-degree right thoracic curve.  

Sagittal Evaluation
Review of the thoracic sagittal alignment revealed a mean Cobb angle of 31 degrees (range, 2-88 degrees; Table 2). Fourteen patients (4 girls and 10 boys) developed hyperkyphosis (20.9%), 10 of which were more than 50 degrees. Normal and hypokyphotic curves were seen in 55.2% (n = 38) and 23.9% (n = 16) of patients, respectively. Patients with hyperkyphosis underwent thoracotomy at an average of 19 months and sternotomy at 48 months. Those with normal kyphosis or hypokyphosis underwent the first procedure at 8 months and the second at 37 months. These differences were not statistically significant.

Thirty-five percent of the patients with scoliosis were kyphotic in the sagittal plane, 47% were normal, and 18% were hypokyphotic. Eight patients with hyperkyphosis did not have an associated scoliosis (11.8%). Nine percent of patients (6/68) demonstrated kyphoscoliosis; that is, 12 patients with scoliosis (17.6%) did not demonstrate hyperkyphosis. The average sagittal Cobb angle in those with scoliosis was 38 degrees (range, 12-88 degrees; Fig. 2). Twenty-five percent of patients with moderate to severe scoliosis were hyperkyphotic, and all were boys.

Heart Size and Aortic Arch Position
Chest posteroanterior radiographs revealed that 36 patients had a normal heart size and 32 had cardiomegaly. Eleven (34%) of 32 patients with enlarged hearts developed scoliosis, whereas 7 (19%) of 36 subjects with a normal heart size developed scoliosis. The mean Cobb of the structural curve measured 33 and 47 degrees in the normal and enlarged heart groups, respectively. This difference was not statistically significant (P = 0.24). Of note, 6 of the 8 patients with severe curves had cyanotic cardiac disease as well as an enlarged heart. Position of the aortic arch side was adequately assessed in all patients. Ten of the arches were to the right, and 3 (30%) of these 10 patients developed scoliosis, 2 of which were left thoracic curves. Fifty-eight patients had left aortic arches, 15 (26%) of which developed scoliosis (2 left thoracic curves).

Age at First Procedure
Thoracotomy was the first procedure for all but one patient. The mean age at the time of the thoracotomy was 10 months (range, 0-83 months). The sternotomy was done at about 40 months of age (range, 1-142 months). The thoracotomy was performed before 1 year of age in 52 patients. Thirteen (25%) of 52 were diagnosed with scoliosis, 9 severe and 1 moderate. Ten of 12 moderate to severe curves underwent surgery for congenital heart disease before 1 year of age. One patient developed mild scoliosis after the thoracotomy and before the sternotomy. This patient developed progressive scoliosis that required posterior spinal fusion. In the remaining patients, scoliosis was detected after the second procedure.

Cyanosis Versus Acyanotic Conditions
Forty-four of 68 patients were found to have cyanotic cardiac conditions, and 24 were acyanotic. Sixteen patients (36%) with cyanotic cardiac condition developed scoliosis, whereas only 2 patients (8.3%) with acyanotic conditions developed scoliosis. In the cyanotic group, scoliosis was detected at a mean age of 10.5 years (range, 2.9-17.5 years). The average coronal Cobb angle for this group was 41 degrees (range, 15-78 degrees). For the acyanotic group, scoliosis was detected at a mean age of 10 years (range, 9.3-10.7 years), and the mean coronal Cobb was 32.5 degrees (range, 13-52 degrees). The small sample size prohibited a statistical comparison.

Side of Thoracotomy
Of those patients with scoliosis, there were 10 left thoracotomies, 6 right thoracotomies, and 2 bilateral thoracotomies. There was no correlation between the side of the thoracotomy and the convexity of the structural curve.

Multiple Procedures
Eleven patients (16%) underwent a second sternotomy, a contralateral thoracotomy, or both. Patients with multiple procedures were not at an increased risk for developing spinal deformity.

DISCUSSION
There is an increased incidence of scoliosis in children with congenital heart disease.2-14 Previous authors report the incidence of scoliosis to be between 2% and 31%.8,10,11,14,16,19 Our study yielded a 26% incidence of scoliosis. This is 10 times the rate of adolescent idiopathic scoliosis in the general population.7,20 Of the 26%, 9% had an associated hyperkyphosis, and 17% presented with either hypokyphosis or normal kyphosis. Herrera-Soto et al 16 previously reported a 28% incidence of scoliosis after a median sternotomy for patients with congenital heart disease. This is similar to the incidence of scoliosis in previous thoracotomy studies.2-14 This raises the question of a vascular etiology as opposed to surgical approach. The mean Cobb angle of the structural coronal curve was 40 degrees for patients undergoing combined thoracotomy and sternotomy for the treatment of congenital heart disease. This is higher than that reported by Herrera-Soto et al 16 on patients with sternotomies only where the mean Cobb angle was 24.5 degrees. Scoliosis was mild in 8.8% of all patients studied and was moderate to severe in 17.6%. Most curves (67% of all curves) were moderate to severe curves. This is in contrast to other reports of open-heart surgery where severe curves are rare with an approximate incidence less than 5%.6,9-11,21 It is important to consider following up these patients to maturity to monitor curve progression.

In this series, the incidence of scoliosis was almost equal in boys and girls. This is in contrast to idiopathic scoliosis, which tends to occur with greater frequency in girls.22 Girls with congenital heart disease were more likely to develop severe deformity similar to adolescent idiopathic scoliosis.22

Van Biezen et al 11 reported a 22% incidence of scoliosis in 160 patients with coarctation of the aorta after undergoing thoracotomy. None of their patients had scoliosis before the index thoracotomy. They concluded that the increased risk of scoliosis was caused by the surgical approach and not the congenital anomaly. Twenty-five of these underwent a second procedure (19 median sternotomies and 6 thoracotomies). They did not evaluate the risk of developing scoliosis before or after the second procedure or progression of scoliosis after a second procedure. None of our patients presented with scoliosis before the index procedure. There was no correlation between the age at time of thoracotomy or sternotomy surgery and the probability of developing scoliosis. The mean age at the detection of scoliosis in this study was 10.7 years, almost 10 years after the first procedure. Early onset likely increases the risk of curve progression.

In 1934, Bisgard 15 presented the concept of thoracogenic scoliosis. The theory is based on imbalance of opposing forces on the spinal column. Those procedures that destabilized the spine, including extensive rib resections or lobectomies, created clinically significant deformities. The convexity was toward the side of the thoracotomy. Although many authors have also reported scoliosis after thoracotomy to be convex toward the side of the procedure, this is not consistently found.4,8,10,11,23 Our cardiothoracic surgeon does not resect any portion of the rib, as this increases the risk of spinal deformity. Several reports on the incidence of scoliosis after thoracotomy for the treatment of esophageal atresia reveal the apex of the curve to be opposite to the side of the procedure.23 This phenomenon has been termed pleural scoliosis where pleural scarring is thought to contribute to the development of scoliosis. Most of the thoracotomies are closed by suturing ribs together, which may contribute to rib fusion. One report has implicated rib fusions after the repair of atresia as the cause of scoliosis.23 All of the patients in that series with scoliosis underwent more than one thoracotomy and had 4 or more fused ribs. The authors concluded that patients who undergo multiple procedures are at an increased risk. There was no correlation in our series between the side of the thoracotomy and the convexity of the structural curve. In addition, multiple procedures did not seem to influence the development of spinal deformities.

The process of sternal development begins by the fifth fetal month.19 A peripheral physeal plate with distinct zones of cellularity, similar to the epiphyseal plates in children, surrounds each ossification center.24 These centers expand and eventually coalesce. The rate of ossification is dependent on the amount of load applied. It has been proposed that shear stresses accelerate this process, whereas compressive stresses inhibit growth. Intuitively, if the sternum is stressed sufficiently after a sternotomy, this may cause premature closure of the sternal growth centers, leading to tethering forces that may create increased kyphosis.

In this series, the combination of sternotomy and thoracotomy affected the sagittal alignment as well. The mean sagittal thoracic curve in adolescent idiopathic scoliosis ranges from 20 to 23 degrees.25 Our series yielded 38 degrees of kyphosis. In a recently presented study on sternotomy alone for congenital heart disease, the incidence of kyphosis was 22% versus 21% in the current series.16 Of that 21%, 9% had kyphosis and associated scoliosis, and 12% did not. That report yielded a mean sagittal Cobb angle of 31 degrees for patients with adolescent idiopathic scoliosis. There are no reports, to the best of our knowledge, on the sagittal profile of patients who undergo a thoracotomy. Thirty-three percent of the scoliosis patients in our series presented with hyperkyphosis. This may have implications for bracing and surgical correction of these kyphoscoliotic curves.5 Because this is the first article to address sagittal deformity after combined thoracotomy and sternotomy, we cannot compare the incidence of hyperkyphosis to other reports. However, based on our findings, both the sagittal and coronal curves tend to be worse than those after sternotomy alone.16 It seems that the thoracic spine receives a "double hit" when both procedures are combined.

The presence of a cyanotic condition has been theorized as a possible explanation for the development of scoliosis.6-8,10 It has been proposed that decreased bony perfusion may impair symmetric bone growth. This has not been supported in other studies.11,14,16 In our study, we found that patients with cyanotic cardiac were 4.5 times more likely to develop scoliosis, and the curves were more likely to be severe.

Increased heart size has also been implicated as a possible risk factor.14,20 In our series, heart size did not increase the risk of scoliosis or hyperkyphosis. The side of the aortic arch did not influence the convexity of the curve.

CONCLUSION
The incidence of scoliosis in children with congenital heart disease is high. Spinal deformities, including scoliosis and/or hyperkyphosis, were found in 38% of the patients. The risk of developing scoliosis after a combined thoracotomy and sternotomy for the treatment of congenital heart disease is more than 10-fold than that of idiopathic scoliosis. The risk of developing severe curves after a combined procedure triples that of thoracotomy alone. The thoracic spine seems to receive a "double hit" when both procedures are performed. Spinal deformity develops at a younger age, increasing the risk of progression. There does not appear to be a sex predilection, although girls tend to develop more severe scoliosis. Patients with cyanotic conditions have a 4-fold increased risk of developing scoliosis. Unlike adolescent idiopathic scoliosis, the sagittal alignment tends toward hyperkyphosis. The authors recommend that patients with congenital heart disease be evaluated throughout skeletal growth and observed closely for coronal and sagittal spinal deformities (Fig. 3).


[Help with image viewing]
[Email Jumpstart To Image]  FIGURE 3. A 12-year-old girl who underwent a right thoracotomy at 4 months of age and a median sternotomy at 4 years for tetralogy of Fallot. These are her initial films. A, Anteroposterior spine film revealed a T5-T10 curve of 30 degrees and a T10-L4 left lumbar curve of 42 degrees. B, Lateral view demonstrates a normal sagittal contour with lack of normal lordosis at the apex.