Infantile Scoliosis

Michael T. Rohmiller and Behrooz A. Akbarnia

Initially, James divided idiopathic scoliosis into three categories based on the age at presentation: infantile (0 to 3 years), juvenile (4 to 9 years), and adolescent (10 years to skeletal maturity).1 At the time of the categorization, it was believed that the three groups corresponded with the periods of increased spinal growth velocity and thus increased risk of scoliosis progression. However, later published data demonstrated that spinal growth velocity is relatively constant or even declines between 5 and 10 years of age.^2–4 Due to the rarity of juvenile onset scoliosis and the data demonstrating decreased spinal growth velocity during this age frame, Dickson recommended that idiopathic scoliosis be divided into two categories: early onset (0 to 5 years) and late onset (after the age of 5).^5,6

The age of onset remains important as a prognostic tool because curves that demonstrate a significant thoracic deformity before age 5 will have a higher likelihood of cardiopulmonary abnormalities such as restrictive lung disease, pulmonary hypertension, cor pulmonale, and thoracic insufficiency syndrome.^7–10 In addition, the treatment of children with early onset scoliosis has proved to be difficult. Standard modalities used in older patients with scoliosis such as orthotic treatment or spinal fusion have limited roles due to the potential deleterious effects on the growth of the spine, lungs, and thoracic cage. This chapter will focus on the evaluation, treatment, and outcomes of patients diagnosed with infantile (early onset) idiopathic scoliosis.

♦ Natural History

Infantile idiopathic scoliosis is a curvature of the spine in the coronal plane of a child in the first 3 years of life for which no identifiable cause is known. Unlike idiopathic scoliosis, there are multiple described etiologies of curvature of the spine in infants. Common causes include congenital vertebral or chest wall anomalies, neuromuscular (cerebral palsy, myelomeningocele, and muscle disease), syndromic scoliosis such as neurofibromatosis, and scoliosis due to intraspinal pathology such as tumor, diastomatomyelia, syrinx, or tethered cord. After eliminating all possible causes of early onset scoliosis, a child is diagnosed with infantile idiopathic scoliosis.

Infantile idiopathic scoliosis is relatively rare and accounts for less than 1% of idiopathic scoliosis seen in the United States. The incidence of infantile idiopathic scoliosis is slightly higher in Europe.^11,12 Infantile scoliosis presents as a left-sided mid to low thoracic curve in 75 to 90% of cases.^2,11,13 Unlike late onset idiopathic scoliosis, boys are affected more frequently than girls with a 3:2 ratio.^11,13 Fortunately, the prognosis for infantile scoliosis also differs from late onset. In the initial description of infantile scoliosis by James, he described a curve that “developed rapidly and relentlessly, causing the severest form of orthopaedic cripple with dreadful deformity, marked dwarfing and shortening of life.”¹⁴ The original series reported that 4 out of 33 cases resolved with the remaining developing major deformities.¹⁴ In the following years, he added more patients to his study and ultimately reported on a total of 212 patients with infantile scoliosis. In the larger group, 36% of the cases resolved spontaneously.¹¹ Similarly, Scott and Morgan reported a significant likelihood of progression in infantile scoliosis curves.¹³ Within 10 years, several authors reported an increase in the frequency of spontaneous resolution.^15,16 Later, Ceballos et al. reported a 74% rate of resolution.¹⁷ Today, as in the study by Lloyd-Roberts and Pilcher, more than 90% of cases will spontaneously resolve without the need for treatment.¹⁸ However, one must be aware that girls who present with a right-sided thoracic infantile curve have a worse prognosis and do not follow the typical rate of spontaneous resolution.¹⁹

Since the initial description almost 75 years ago, the number of cases of infantile scoliosis has declined.20 Further investigation reveals possible factors responsible for the difference in incidence rates. Infantile scoliosis was initially attributed to intrauterine molding. Browne hypothesized that intrauterine molding caused not only infantile scoliosis but was also responsible for the associated crowding deformities. In his series, 83% of patients exhibited some form of intrauterine crowding deformity such as plagiocephaly, plagiopelvy, decreased hip abduction, or rib molding.^21,22 Mehta subsequently agreed that intrauterine molding was responsible for infantile scoliosis.²³ The view has subsequently been refuted due to the absence of scoliosis at birth. Although intrauterine molding is likely responsible for the other deformities that are occasionally present at birth, the association with scoliosis has not been shown. The second hypothesis to explain infantile scoliosis was proposed by Mau.¹⁵ Mau believed that prolonged oblique supine positioning in the crib was responsible for infantile scoliosis. The focused pressure from positioning a child that is unable to roll may induce scoliosis, plagiocephaly, hip adduction, and pelvic molding. Wynne-Davies’ data also supports the postnatal pressure theory as she observed that 97 of 134 cases developed scoliosis and plagiocephaly within the first 6 months of life (time during which children are unable to reposition).²⁴ The rate of plagiocephaly in children without scoliosis is 28%.²⁵ Given that European children were sleeping supine whereas American children were positioned prone, this may also account for the difference in incidence of infantile scoliosis between the two continents. A new study evaluating the incidence of infantile scoliosis in children in Europe and North America is needed as the current recommendations from the American Academy of Pediatrics advocate supine sleeping to decrease the risk of sudden infant death syndrome.²⁶

In addition to her contributions in other aspects of pediatric orthopaedics, Wynne-Davies also described several associations with infantile scoliosis. She confirmed the predominance of left-sided thoracic curves and noted that most were diagnosed between 1 and 6 months of age. Because very few curves were present at birth, she also refuted the claim that these curves were congenital in nature. However, she has hypothesized that intrauterine molding plays a role similar to congenital muscular torticollis. Although torticollis is rarely present at birth, it is believed to be due to intrauterine molding and presents after the body resumes the intrauterine shape in the first months of life. In her study population, 13% of male infants with progressive curves were mentally retarded, and 7% of males had concomitant inguinal hernias. Progressive curves were commonly associated with advanced maternal age. The genetics of infantile scoliosis were similar to that of late onset scoliosis in which parents or siblings of affected children are 30 times more likely than controls to have scoliosis.²⁴ Mehta’s data concurred with that of Wynne-Davies. She noted that infants with hypotonia were unable to resist deformation compared with children with normal tone.²³ Connor reported that children with congenital malformations, including hiatal hernia, were at increased risk of progression.²⁷

♦ Anatomic and Physiologic Variants of Infantile Scoliosis

One cannot attempt to evaluate and treat patients with infantile scoliosis without an understanding of the anatomic involvement of the condition. Infantile scoliosis affects the growth of the chest wall, spine, and lung parenchyma. Normal development of the lungs requires adequate space in the thoracic cavity. When inadequate space is available, abnormalities occur based on the time at which the restriction occurs. If the compression is prenatal, major airway development is abnormal. This is most commonly seen with congenital diaphragmatic hernia and resultant pulmonary hypoplasia. This entity does not occur even in severe cases of congenital scoliosis. When the restriction develops after birth and before age 8, the alveoli are affected. Davies and Reid have shown that the alveoli are normal in size but reduced in number.²⁸ This scenario occurs in infantile scoliosis and the reduction in alveoli is greater with earlier presentation.

Major airway development is complete at birth, but limited bronchioles and alveoli exist. It is estimated that 20 million alveoli are present at birth. Normal alveoli develop rapidly during the first year of life, increasing in both size and volume. By age 4, 250 million alveoli have developed, and the full adult alveoli development is complete by age 8. Similarly, by age 8, the respiratory tree has matured to the final adult composition with 23 generations.

Spine and thoracic cage growth occur simultaneously with the pulmonary system. Dimeglio demonstrated that the spine from T1 to S1 grows an average of 2 cm per year from birth to age 5. Two thirds of the final sitting height is achieved by 5 years of age. Between age 5 and 10, deceleration in the rate of growth occurs followed by a second increase after age 10 during the pubertal growth spurt. However, the growth velocity is fastest during the initial phase between birth and age 5.⁴ The volume of the thoracic cage also increases during early life although more slowly than the spine or lungs. The thoracic volume increases long after the pulmonary development has ceased. Full adult thoracic volume is attained by age 15 in males and females. At birth, the thoracic volume is 6% of final adult volume. At age 5, the volume has increased to 30% and reaches 50% of the adult volume by age 10.⁴

Given that the growth of the spine, chest wall, and pulmonary system are rapid during the first 5 years of life, conditions that interfere with normal development of one system have serious ramifications on the development of related structures. Scoliosis that presents during this time has a higher chance of cardiopulmonary compromise. Early onset scoliosis impairs the normal development of the alveoli and pulmonary vessels. The abnormal lung parenchyma and vasculature result in ventilation defects.²⁹ The severity of pulmonary compromise is related to the age of onset of scoliosis. Patients who develop scoliosis at a younger age will have more severely involved lungs.⁷ Restrictive lung disease with decreased vital capacity (VC) and total lung capacity (TLC) and increased residual volume (RV) is seen in patients with early onset scoliosis. Compliance of both the lungs and thoracic cage are decreased and account for the restrictive pattern. The actual percent decrease in VC is related to the severity of deformity and the age of onset. Early onset curves will have more deleterious effects than similarly severe curves that present after maturation of the pulmonary system. When severe, restrictive pulmonary disease results in pulmonary arterial hypertension and cor pulmonale. The gas exchange is normal in these patients so hypoxemia is related to diminished tidal volumes.30 Respiratory compromise or failure is common in patients with severe early onset scoliosis, although it frequently presents later in life due to significant pulmonary reserve.^7,8 This situation differs from patients with scoliosis and thoracic insufficiency syndrome that present with pulmonary failure at an early age.¹⁰

Previous authors have reported that untreated idiopathic scoliosis resulted in cardiopulmonary compromise. Whereas these reports typically described the patients as having adolescent idiopathic scoliosis (AIS), subsequent data indicate that patients with adolescent idiopathic scoliosis are unlikely to develop curves of a magnitude large enough to create life-threatening cardiopulmonary compromise. Close examination of historic papers reveals that many of the patients previously categorized as AIS most likely had infantile scoliosis. Cardiopulmonary compromise relies on incomplete development of the pulmonary system with decreased space available for the lungs. Patients who develop AIS (or late onset scoliosis) have near normal lung and chest wall development. Therefore, the only aspect of the multifactorial formula of cardiopulmonary compromise present is scoliosis, and this is not powerful enough to create cardiopulmonary distress.

♦ Clinical Evaluation

The clinical evaluation of a child with suspected scoliosis or spinal deformity must proceed in a thorough, systematic fashion. A complete history must be obtained prior to physical examination. Prenatal history of the mother including health problems, previous pregnancies, and medications is performed. Birth history of the child should include details such as length of gestation, type of delivery (vaginal or caesarean), birth weight, and complications. Similar to developmental dysplasia of the hip, infantile scoliosis has an association with breech presentation.²⁴ Unlike developmental dysplasia of hip (DDH), infantile scoliosis is also seen more commonly in males with low birth weight. The developmental history should include both motor and cognitive milestones. This must be carefully performed because cognitive delay has been associated with progressive curves.^24,27

The physical examination attempts to analyze the spinal deformity and eliminate associated conditions from the potential diagnosis. Initial inspection should include the skin, entire spine, head, pelvis, and extremities. The skin must be examined for cutaneous stigmata such as café au lait spots or axillary freckles seen in neurofibromatosis, midline patches of hair seen in spina bifida occulta, or bruising seen in trauma. The spine exam should include inspection and palpation of the spine. In young children, the Adams forward bend test (looking for prominence of the ribs in the thoracic spine or transverse processes in the lumbar spine) is not possible, but the test can be simulated by lying the child prone over the examiner’s knee. Curve flexibility can be assessed by placing the child in the lateral position over the knee or suspending the infant under the arms. Notation of chest or flank asymmetry, chest excursion, and abdominal reflexes must be made. Limitation in chest excursion may indicate syndromic scoliosis and thoracic insufficiency syndrome.¹⁰ Abdominal reflex abnormalities should initiate a thorough neurologic evaluation. Muhonen et al. described the absence of an abdominal reflex as the only objective finding seen in some patients with a Chiari malformation.³¹ When abnormal, the absent reflex is usually on the convex side of the curve.³² Other studies suggest that even in the presence of a normal physical exam, total spine magnetic resonance imaging is indicated in patients with infantile scoliosis due to the high incidence of neural axis abnormalities.³³ In the report by Gupta et al. on six patients with infantile scoliosis and normal neurologic exam, three patients were found to have neural axis abnormalities on MRI.³⁴ Dobbs et al. subsequently reported on 46 neurologically normal patients with Cobb angles greater than 20 degrees who underwent MRI. Of those patients, 10 demonstrated a neural axis abnormality of MRI and 8 patients ultimately required neurosurgical intervention for the abnormality. The current recommendation is to perform an MRI on patients with infantile scoliosis and a Cobb angle greater than 20 degrees.³⁵

The head must be thoroughly examined in children referred for scoliosis. Plagiocephaly is common in patients with infantile scoliosis and responds well to therapy. Other conditions affecting the head that are associated with infantile scoliosis include bat-ear deformity and congenital muscular torticollis. Although these conditions frequently occur without scoliosis, one must be aware of the association. The pelvic exam should rule out plagiopelvy and developmental hip dysplasia, both of which are associated with infantile scoliosis.^21,22 The lower extremity exam must exclude limb length inequality as the etiology of scoliosis. When scoliosis is secondary to a limb-length inequality, the lumbar prominence is found on the side of the longer limb. Functional scoliosis due to limb length inequality may be negated by performing a sitting forward bend or placing a lift under the short limb to equalize the limb lengths.