 |
 |
 |
|
| home | CME | help | feedback | subscriptions | archive |
|
© 2000 The Journal of Bone and Joint Surgery, Inc.
Etiology of Idiopathic Scoliosis: Current Trends in Research* 
*No benefits in any form have been received or will be received
from a commercial party related directly or indirectly to the subject
of this article. No funds were received in support of this study.
 Current population studies characterize idiopathic scoliosis
as a single-gene disorder that follows the patterns of mendelian
genetics, including variable penetrance and heterogeneity.
Idiopathic scoliosis is a pathological entity of unknown etiology. Although the entity was first described by Hippocrates, the term idiopathic scoliosis was probably introduced in the middle of the nineteenth century by Bauer39; it was used by Nathan in 190967, defined by Whitman in 192293, included by Cobb in his classification19, and popularized by the Scoliosis Research Society40. Although most physicians who treat spinal deformities understand the term idiopathic scoliosis, the important questions concerning its etiology remain unanswered. The objectives of this paper are to provide an update on a number of aspects of the etiology of idiopathic scoliosis, to present an inventory of current investigational work, and to suggest directions for future research. The identification of etiological factors will depend on continued research in each of the areas discussed in this review. Further understanding of this disorder will enable the clinician to better predict prognosis and to aid in the development of more effective treatment modalities. This work represents an effort on the part of the Scoliosis Research Society Etiology Committee to promote an awareness of the research in this field.
The role of hereditary or genetic factors in the development of idiopathic scoliosis has been widely accepted9,20,26,45,97. Clinical observations as well as population studies have documented scoliosis within families, with the prevalence higher among relatives than within the general population24,34,37,77,79,97. Harrington45 studied women with a scoliotic curve that exceeded 15 degrees and found a 27 percent prevalence of scoliosis among their daughters. Population studies involving index patients and their families have indicated that 11 percent of first-degree relatives are affected, as are 2.4 and 1.4 percent of second and third-degree relatives, respectively77. Studies of twins have consistently shown that monozygous twins maintain a high concordance rate for the condition of approximately 73 percent, whereas dizygous twins have a concordance rate of 36 percent31,35,50,55,64. These values are higher than those reported for first-degree relatives from studies of populations of families identified through an affected individual in which the affection status of the first-degree relatives was determined primarily through an initial examination77,95. This may be related to the high rate of radiographic confirmation of the presence of this condition in studies of twins compared with that in large population studies of families. Radiographic confirmation of the disease potentially lowers the false-negative rate, as small scoliotic curves undetectable by clinical examination are identified. Despite documentation of the familial nature of this condition, the mode of inheritance has been debated. Studies based on a wide variety of populations have suggested an autosomal dominant, X-linked, or multifactorial inheritance pattern15,22,37,61 (Fig. 1). Wynne-Davies95 reported on a series of 2000 individuals in which all first-degree relatives of an identified affected individual (the proband) were clinically examined. Radiographs were made only if a positive diagnosis was suspected clinically. The results suggested a dominant mode of inheritance. In a series of 2869 individuals reported on by Riseborough and Wynne-Davies77, all first-degree relatives were examined clinically and radiographically and 81 percent of second and third-degree relatives were evaluated radiographically. That series showed a multifactorial pattern of inheritance.
In 1972, Cowell et al.20, noting the paucity of reports of male-to-male transmission, selected seventeen families (192 individuals) for physical and radiographic examination and reported a pattern consistent with an X-linked inheritance. With the advent of statistical analysis directed to the potential linkage of genes to known disorders (genetic linkage analysis), Miller et al.61 investigated X-linkage in fourteen families (136 individuals). This analysis calculates the odds (LOD score) that the genetic loci are linked to an observed trait compared with the odds that a given segregation pattern of genetic loci occurs by chance alone. Overall results did not support X-linkage within the entire population; however, when each family was considered independently, the LOD scores suggested that at least two loci (one autosomal and one X-linked) may have been involved in the expression of scoliosis within these families. An alternative statistical method directed at the clarification of a genetic model and the penetrance of a familial disease is complex segregation analysis. This methodology is applied to an unscreened population and can confirm clinical observations that a genetic determination exists for a specific disorder. Aksenovich et al.2 used complex segregation analysis to study a population of ninety families (283 individuals). While their results confirmed the clinical observation of a genetic determination of the disorder, the best-fitting genetic model proves to be equivocal. Collectively, these studies characterize idiopathic scoliosis as a single-gene disorder that follows the simple patterns of mendelian genetics. This concept defines the gene as the inherited unit transmitted from parent to offspring, which is responsible for the observable trait. Traits can be dominant in expression, meaning that the presence of one gene is sufficient to express the condition, or they can be recessive, meaning that the condition is expressed only in the absence of the dominant factor. Careful descriptions of a study population are essential to determine the expression of a specific gene and hence the transmission pattern of an observable trait. For example, Marfan syndrome is a dominant trait in which the gene is responsible for the condition transmitted from each affected parent to half of his or her offspring. However, simple gene disorders are susceptible to the genetic principles of variable penetrance and heterogeneity. Variable penetrance occurs when a certain percentage of individuals carrying the gene of interest do not express the observable trait. Genetic heterogeneity exists when two or more genes within a study population act independently and each leads to the observable trait. Thus, clinically identical subgroups are the result of distinct genetic etiologies. Common disorders among the population, such as scoliosis, that present clinically in different ways and are believed to have a genetic basis are potentially the result of these complex genetic interactions. Consequently, it is difficult to conduct isolated studies of a family or a small sample population in an effort to yield positive results for genetic linkage.
In 1983, Dubousset et al.27 found that scoliosis routinely developed in pinealectomized chickens, and they attributed this effect to decreased melatonin production. This led Dubousset and Machida28 to measure the levels of melatonin in thirty adolescents with idiopathic scoliosis and in fifteen age-matched controls. The patients had severe scoliosis ranging from 57 to 75 degrees; the curves were divided into those that had progressed more than 10 degrees in the preceding year and those that had not. Patients with progressive scoliosis had a 35 percent decrease in melatonin levels throughout the night compared with those with stable scoliosis or the control subjects28. Bagnall et al.7 suggested that the action of melatonin may be mediated by growth hormone. Sporadic cases of a rapid increase in scoliotic curvature have been reported in patients undergoing growth-hormone therapy48. In a study of side effects of growth-hormone treatment, Allen3 noted scoliosis in fewer than 1 percent of patients, but progression sometimes occurred during treatment. However, the pharmacological relationship of growth hormone to melatonin is still unclear. The diurnal variation in melatonin levels seems to be important in determining the effect of this factor on the development of idiopathic scoliosis. However, this rhythm is obliterated in several diseases and does not have an obvious effect on the development of scoliosis. Moreover, patients with idiopathic scoliosis do not have documented difficulties with sleep or immune function, which might be expected with a substantial decrease in melatonin. There is no evidence that patients with idiopathic scoliosis have an inability to form melatonin. Thus, if a lack of melatonin is a factor in the development of scoliosis, it must be due to an alteration of its synthesis. On the basis of the available data, it is possible that melatonin plays a secondary role in the development of idiopathic scoliosis. However, it seems unlikely that scoliosis results from a simple absence of melatonin. Rather, it could result from alterations in the control of melatonin production, with either direct or indirect consequences on growth mechanisms.
Collagen and elastic fibers are principal elements in the supporting structures of the spinal column and have been the focus of many studies dealing with the pathophysiology of idiopathic scoliosis. Because scoliosis is a phenotypic characteristic of many connective-tissue disorders, such as Marfan syndrome, the hypothesis that a defect within the connective tissue is the causative factor of idiopathic scoliosis is plausible. Work focusing on the quality and quantity of the proteoglycan and collagen contents of the intervertebral discs has produced conflicting results. Pedrini et al.74 demonstrated an abnormal proportion of glycosaminoglycans and collagen content of the nucleus pulposus of intervertebral discs in patients who had idiopathic scoliosis. This finding was supported by Taylor et al.90 but not by Oegema et al.73. Using specimens obtained at autopsy as controls, Roberts et al.78 performed a histological and biochemical study of vertebrae and intervertebral discs in adolescents with idiopathic scoliosis. While changes in the distribution of collagen compared with that in normal subjects were evident, they were not consistent among the subjects who had scoliosis. Those authors suggested that the changes may be secondary to the abnormal mechanical forces applied to the discs rather than being the primary cause of the scoliotic deformity itself. The elastic fiber system, the second major component of the extracellular matrix, has been studied in individuals with idiopathic scoliosis by two groups of investigators. Echenne et al.29 examined the skin of patients who had idiopathic scoliosis and found substantial differences within the middle and deep dermis compared with those of normal subjects. However, type-I muscle abnormalities also were observed in seventeen (50 percent) of the thirty-four scoliotic patients. In 1994, Hadley-Miller et al.43 reported elastic fiber abnormalities in the spinal ligaments of a substantial number of patients with idiopathic scoliosis compared with those of normal individuals. Analysis of harvested fibroblasts cultured in vitro indicated a potential failure of matrix incorporation of elastic fiber components in several patients with scoliosis. Carr et al.15 and Miller et al.60 focused on the genes responsible for the structural components of the extracellular matrix system. Both groups of investigators selected families in which scoliosis was expressed in an autosomal dominant pattern and analyzed genetic linkage within the families through a candidate-gene approach. With this approach, genes are selected for study on the basis of scientific data, physiological rationale, and clinical understanding of the disease. Genetic linkage is established if the disease phenotype (that is, scoliosis) segregates with a particular allele of the gene. Conversely, if the gene is inherited independent of the disease, it cannot be responsible for the disorder. The structural genes of collagen types I and II, fibrillin 15 (FBN), and elastin were excluded as potential causative factors for idiopathic scoliosis within the study populations. While this information may be regarded as definitive, the studies were limited to a selected number of families with idiopathic scoliosis. The dilemma of whether the changes observed within the connective tissues of individuals with idiopathic scoliosis might be the consequence of scoliosis rather than the causative factor is still ongoing. However, most researchers concede that abnormalities reported within these elements of the majority of individuals affected with idiopathic scoliosis are probably secondary to the structural forces of the scoliotic deformity itself30,45.
The idea that an abnormality of the paraspinous muscles might be the cause of idiopathic scoliosis has been entertained for many years. Spencer and Eccles87 were apparently the first to describe the two types of muscle fibers in paravertebral muscles in patients with adolescent idiopathic scoliosis. They differentiated between type-I (slow-twitch) and type-II (fast-twitch) fibers and noted that the number of type-II fibers was decreased in their patients, suggesting a myopathic process. Sahgal et al.79 noted similar findings in the gluteus medius muscle. Bylund et al.13 described a normal distribution of type-I and type-II fibers on the convexity of the curve but a lower frequency of type-I fibers on the concavity. Slager and Hsu85 biopsied paravertebral muscle in thirty-one patients with adolescent idiopathic scoliosis and noted a decrease in the number and size of type-II fibers in twenty-one and seventeen patients, respectively, with no preference for either the convex or the concave side. Yarom et al.99,103 noted similar findings in muscles from distant sites (the deltoid, trapezius, gluteus, and quadriceps) and concluded that this represented a myopathic process. Yarom and Robin100 studied paraspinous muscles as well as other muscle sites with use of light microscopy; they found fiber-splitting, tubular bodies, and contraction bands and confirmed the presence of central core formation in many fibers. Myofilament disarray and Z-line streaming also were found both in paraspinous muscle (to a greater extent on the concave side) and at distant sites (the gluteus maximus). The sarcomere was shortened primarily on the concave side; the A-band also was shortened, but equally on both sides. In a similar study, Low et al.53 noted that lipid, glycogen, and membranous bodies in the muscle fibers were increased and the sarcoplasmic reticulum was slightly dilated. Both groups of investigators postulated a myopathic process. Ford et al.36 noted a marked decrease in muscle spindles in all paraspinous muscles that were tested in patients with adolescent idiopathic scoliosis. Using biochemical analysis of the paraspinous muscles with x-ray fluorescence spectrometry in patients with adolescent idiopathic scoliosis, Yarom and Robin100 demonstrated a markedly increased calcium content. The authors thought that patients with adolescent idiopathic scoliosis might have a generalized membrane defect - namely, an impaired calcium pump. Using the stable isotope-labeled l-leucine, Gibson et al.38 analyzed muscle protein synthesis in paravertebral muscle biopsy specimens obtained bilaterally from the top, bottom, and apex of the curve in nine children with idiopathic scoliosis. No differences were noted between the two sides of the spine; however, at the apex of the curve, synthesis was higher on the convexity than on the concavity in all patients. Muscle ribonucleic acid activity at the curve apex was lower on the concave side than on the convex side. Gibson et al. believed that these results were consistent with effects on muscle protein turnover secondary to increased muscle contractile activity and functional immobilization of the muscle on the curve concavity. In summary, no definite conclusions can be reached with regard to the etiological involvement of skeletal muscle abnormalities. Most of the abnormalities that have been noted are likely secondary to the deformity itself; however, the histochemical changes that have been described might indicate a defect of the cell membrane.
Abnormalities in the structure and the function of thrombocytes have been noted in patients with adolescent idiopathic scoliosis by many investigators51,63,101,102. Because the protein contractile systems (actin and myosin) of platelets and skeletal muscle resemble each other, it follows that both of these elements would be affected if there is an underlying systemic disorder involving either the structure or the function of the protein contractile system. The platelet is an attractive model to study because its contractile system, which controls its shape, is independent of the axial skeleton, making it independent of the secondary effects in skeletal muscle that potentially are produced by the scoliotic deformity. Yarom et al.102 first noted that calcium and phosphorus levels in skeletal muscle were elevated in several diseases; they subsequently observed that intracellular calcium and phosphorus levels also were elevated in the platelets of the same patients. Increased calcium levels also were found within the intracellular dense bodies. Some of the platelets were larger than normal. A calcium transport defect in connection with the cell membrane or contractile protein metabolism was suggested as an explanation. Muhlrad and Yarom63 noted decreased activity of the intracellular contractile proteins as well as myosin adenosine triphosphatase activity within the platelets. They also noted decreased platelet aggregation with adenosine diphosphate and epinephrine in patients with idiopathic scoliosis. Using electron microscopy and a metal impregnation technique, they demonstrated three distinct platelet types: reticular, metallophilic, and pale cells. The metallophilic platelet was seen most commonly in patients with idiopathic scoliosis, especially those with larger curves, whereas the reticular type was found in controls. This difference was thought to be related to differences in membrane permeability. In a different study102, the authors analyzed platelets from forty-nine patients with mild scoliosis (less than 15 degrees) and twenty patients with curves greater than 20 degrees, and they compared these groups with twenty-five normal adults and thirty normal adolescents. The number of dense bodies was increased in patients with idiopathic scoliosis, especially in those with curves of greater than 20 degrees. Those authors noted a distinct difference in the surface electrical charge, with the patients with idiopathic scoliosis, especially those who had larger curves, having more negatively charged metallophilic platelets than the controls. Calmodulin, a calcium-binding receptor protein, is a critical mediator of eukaryotic cellular calcium function and a regulator of many important enzymatic systems. As mentioned, because the contractile protein systems (actin and myosin) of platelets and skeletal muscle resemble each other, it follows that an underlying systemic disorder affecting the contractile system of skeletal muscle would also affect the contractile system of platelets. Calmodulin regulates the contractile properties of muscles and platelets through its interaction with actin and myosin and its regulation of calcium fluxes from the sarcoplasmic reticulum. Increased calmodulin levels in platelets have been shown to be associated with the progression of adolescent idiopathic scoliosis. Kindsfater et al.51 compared seventeen patients who had adolescent idiopathic scoliosis of varying severity with ten age and gender-matched controls and found that platelet calmodulin levels in skeletally immature patients with curves progressing more than 10 degrees per year were considerably higher than the levels in patients with stable curves (3.8 compared with 0.7 nanogram per microgram of protein); the levels in the patients with the stable curves and those in the control group were similar. This data is particularly important when compared with the recent findings of Dubousset and Machida28, who noted decreased levels of melatonin in patients with adolescent idiopathic scoliosis that had progressed more than 10 degrees and normal melatonin levels in patients with stable curves. Recent evidence suggests that melatonin may act by modulating calcium-activated calmodulin28. Melatonin binds to calmodulin with high affinity and has been shown to act as a calmodulin antagonist; as such, it may modulate diurnally many cellular functions involving calcium transport. Multiple pathological biochemical and histological changes have been noted in the platelets of patients with adolescent idiopathic scoliosis. As previously discussed, these changes are similar to those in paraspinous muscle and suggest a primary defect in cells with a contractile system. Some of these abnormalities appear to be related to a defect in the cell membrane and include elevation of intracellular calcium and phosphorus levels, decreased activity of intracellular contractile proteins, decreased platelet aggregation, increased numbers of intracellular dense bodies, a greater number of metallophilic cells, a higher negative surface charge of platelets, increased calmodulin activity, abnormal peptide structure of the myosin chains, and a decreased number of alpha-2 adrenergic receptor sites on platelets. A small percentage of thrombocytes in patients with idiopathic scoliosis are larger than normal. This has not been described in controls. These changes in platelet morphology and physiology suggest a cell-membrane defect in patients with idiopathic scoliosis.
Over the last twenty years, sophisticated neurological investigations have been used to compare patients who have idiopathic scoliosis with controls and to compare patients who have progressive curves with those who have nonprogressive scoliosis41,49,54,80; however, many of the results have been inconsistent103. No clear-cut neurological tests either for diagnosing idiopathic scoliosis or for predicting its progression have so far been established. Evaluation of vibration sense by means of a biothesiometer has been used to investigate dorsal column dysfunction, but the results have varied from study to study and the biothesiometer has now been deemed unreliable54. Similarly, impaired peripheral proprioception is not a constant finding49. Interestingly, patients with idiopathic scoliosis have responded poorly compared with controls when tests for visual and proprioceptive function have been combined103 or when spatial orientation has been evaluated46. It may be difficult to distinguish cause from effect when a neurological abnormality is found in association with idiopathic scoliosis. For example, an abnormal sway pattern (measured with stabilometry) is one area of proprioception that was originally thought to be a primary abnormality (that is, an abnormality that was causal to the scoliosis)65. Increased sway is now considered to be secondary to scoliosis of any cause (that is, it is an effect of the spinal deformity)98. It tends to be more marked when the central nervous system is still immature80. When growth ceases, the enhanced-sway phenomenon apparently reverts to normal despite the scoliosis98. A number of studies have shown an abnormal nystagmus response to caloric testing in patients with idiopathic scoliosis, suggesting an oculovestibular abnormality72,80,98. Herman et al.46 proposed that a dysfunction of the motor cortex that controls axial posture results from a sensory input deficiency concerning spatial orientation and that this effect probably results from central proprioceptive sources involving visual and vestibular function. Other reports have supported this concept12,49,72. The clinical syndrome of symmetrical horizontal or lateral gaze palsy is associated with a high prevalence of scoliosis of the idiopathic type44. The site of neurological abnormality is thought to be the paramedian pontine reticular formation, which links the preocular motor nuclei and the vestibular nuclei. It is reasonable to speculate that the site of neuropathy in idiopathic scoliosis could also be the paramedian pontine reticular formation. Other authors have deemed the cerebral cortex to be the likely source of postural misinformation. Recent studies involving electromyography and corticosensory evoked potentials in patients undergoing surgery for idiopathic scoliosis have demonstrated abnormal and asymmetrical latencies correlating with the side and indeed the progression of the scoliosis18,58. The advent of magnetic resonance imaging has led to renewed interest in abnormal neuroanatomy linked to scoliosis. A cervicothoracic syrinx associated with a Chiari type-I malformation at the foramen magnum has a substantially increased prevalence in patients with idiopathic scoliosis, particularly those who exhibit a juvenile onset of this disorder5,42,47. A review of the literature demonstrated that the prevalence of a syrinx in comparable series of patients with scoliosis ranged from 17 to 47 percent42. The site and extent of the curve are no different from those found in idiopathic scoliosis without a syrinx, but there is a greater prevalence of left-sided thoracic deformity in patients with a syrinx5,32. Neurological function may be normal, although in some patients the abdominal reflexes are asymmetrical104. It is not yet known whether this is secondary to the syrinx formation as the first sign of syringomyelia or whether this asymmetry might reflect a more proximal hindbrain or midbrain lesion. Such a lesion could be linked to or even causative of the syrinx, the tonsillar prolapse, and the scoliosis. Alternatively, the Chiari malformation and the syrinx could be the result of traction on the medulla distally through the foramen magnum. It is also possible that growth of the spinal cord is slower in patients with idiopathic scoliosis, resulting in a cord that is shorter than a more rapidly growing vertebral canal, or that reduced growth of the cord is the result of pineal dysfunction, perhaps involving a melatonin deficiency, as discussed elsewhere57. Experimental scoliosis has been produced in mammals by both anterior and posterior spinal rhizotomies over several levels89. It also occurs when the posterior gray matter and adjacent Clarke columns are damaged in the thoracic spine75. These curves have provided useful animal models but are probably unrelated to idiopathic scoliosis per se, being more akin to a paralytic or neuropathic deformity. Recently, experimental scoliosis has been produced in rats and rabbits by the creation of microscopic electrocoagulation lesions of the brain with use of stereotactic devices8,98. These small lesions lie in the area of the pons and the periaqueductal gray matter, close to the vestibular and preocular nuclei. This zone, the paramedian pontine reticular formation, is the one that Herman et al.46 suggested was abnormal. It is also consistent with the site of neuropathology in symmetrical horizontal gaze palsy44. However, it is important to note that scoliosis developed in only a small proportion of the animals in these experiments, and although the deformity was thought to be idiopathic in at least one study the animals were ataxic98. Pinealectomized chickens with induced scoliosis have increased latency of corticosensory evoked potentials56. This might suggest damage affecting sensory conduction in the roof of the third ventricle adjacent to the stem of the excised pineal gland. Alternatively, it might reflect a reduction in melatonin levels causing interference with nerve conduction or with the maturation of the nervous system7,57. Any hypothesis that proposes a neurological defect must account for the impression that many patients with idiopathic scoliosis have above-average ability in sports. These observations have been largely anecdotal, but a study of girls attending ballet school showed that the prevalence of scoliosis may be as high as 20 percent93. It is difficult to account for a neurological defect that allows the patient to excel in activities demanding high proprioception and coordination.
Idiopathic scoliosis is known to be associated with hypokyphosis; a relative imbalance of growth of anterior and posterior structures has been postulated as a cause of hypokyphosis. According to this hypothesis, the anterior structures grow more rapidly than the posterior ones, and with bending forward the vertebral bodies at the apex tend to move out of the way by rotating to the side. Although this hypothesis is gaining in popularity, few studies have addressed the growth of the spine or spinal growth relative to other body segments. In a prospective study, Nissinen et al.68 found that children who had scoliosis were taller and had less kyphosis compared with normal children. Those authors also found that children with scoliosis had increased sitting height. The most comprehensive study of vertebral growth to date was reported by Skogland and Miller84, who found no substantial differences in the radiographic length of the thoracolumbar spine between patients with idiopathic scoliosis and controls. However, when the lengths of the scoliotic spines were corrected with use of the method of Bjure et al.11, they were found to be considerably larger than the spines in the control group. Those authors noted that the period of accelerated spinal growth in puberty started about one year earlier in the girls with scoliosis. There were no differences in the maximum growth rate between the two groups. The height of the sixth thoracic vertebra was notably greater in the patients with idiopathic scoliosis. The height-to-width ratio at the sixth thoracic level also was greater in the patients with scoliosis. They noted that girls have been found to have a greater height of the vertebral bodies than boys, and this difference increases with age91. Archer and Dickson6 and Shohat et al.83 reported that children with larger curves are taller than those with smaller curves. Archer and Dickson attributed this to flattening of the thoracic kyphosis but provided no data to support this conclusion. Carr et al.16 concluded that growth is important to the development and progression of scoliosis but is not an etiological factor. Several authors have studied hormonal control of growth in idiopathic scoliosis; however, the results have been somewhat contradictory and the implications of this research are presently unclear. Misol et al.62 found no differences in growth-hormone levels between patients with idiopathic scoliosis and controls following glucose-tolerance tests and insulin-induced hypoglycemia. Willner et al.94 compared somatomedin-A and growth-hormone levels in girls who did and did not have idiopathic scoliosis. They found that the mean fasting growth-hormone levels were higher in the group with idiopathic scoliosis and that the growth-hormone levels after exercise increased more rapidly in this group. Serum somatomedin values were also greater in the group with idiopathic scoliosis. Willner et al. interpreted these findings as showing an altered sensitivity of the growth-hormone-release mechanism in girls with idiopathic scoliosis. They noted that their findings differed from those of Misol et al.62. Spencer and Zorab88 bioassayed somatomedin activity in normal children and children with scoliosis and found no substantial differences. It is well established that girls with idiopathic scoliosis have a tendency to be taller and more slender than their peers. There are indications that the spine in patients with scoliosis is more slender and longer than that in nonscoliotic children. This spinal pattern has been implicated in a tendency toward column-buckling. As there is a tendency toward rapid growth in early adolescence, just when the scoliosis is most prone to increase, it is presumed that the spine buckles with growth and the posterior ligaments consequently fail to grow in response to anterior growth, acting instead as a tether and forcing the spine into lordosis. Forward bending forces the apical vertebrae of this lordotic segment to translate to the side, resulting in scoliosis. Interpretation of these studies must take into account two factors. First, the age of the patients studied may be an important variable. If, as several studies suggest, patients with idiopathic scoliosis have rapid growth in early adolescence followed by normal growth later, only younger patients would be expected to show abnormalities. Second, our knowledge of the growth hormone-somatomedin axis has increased considerably in the last decade. Somatomedin A is no longer considered to be a distinct growth factor; its supposed actions are now thought to be the result of other growth factors acting singly or in combination. Therefore, studies that have measured somatomedin-A levels are difficult to interpret in light of the new data. The control of growth is extremely complex and involves the interaction of many hormones and growth factors. These include hormones such as thyroxine, sexual hormones, and growth hormone and its releasing factor; various growth factors; and modulators such as calmodulin. While the relationship between calmodulin and scoliosis has been well studied, other growth factors, such as fibroblast growth factor, have not been studied in patients with idiopathic scoliosis, to our knowledge. In addition, we know little about the growth-factor receptors in these patients. Previously described effects of melatonin might not be completely separate from the growth-hormone axis. Furthermore, melatonin has recently been shown3 to independently induce the production of insulin-like growth factor-1; therefore, it may have the capacity to affect growth in a manner independent of growth hormone.
Biomechanical factors can affect spinal alignment in ways that are often evident in the pathogenesis of idiopathic scoliosis and obvious in some forms of nonidiopathic scoliosis. Mechanical properties of the spinal tissues, alignment of the spine, abnormal loading (either through forces or displacement), and the way that the spine is supported may lead to the development of scoliosis. The manner in which a structure is supported (the boundary conditions) is an important determinant of its mechanical behavior. This has been recognized in patients with neuromuscular scoliosis, in whom pelvic obliquity is reportedly an important contributor to spinal stability52. Scoliosis may develop due to weak or insufficient abdominal musculature that is unable to adequately support the spine52, but this biomechanical aspect has not been studied extensively in idiopathic scoliosis. Altered material properties of the tissues of the spine may affect its response to mechanical loading, and this is a possible mechanism for the development of scoliosis. One such material property that is commonly mentioned is axial stiffness, and much of the early research on spinal biomechanics focused on establishing the mechanical characteristics of spinal tissues. However, those were mostly observational studies of tissue from normal spines, and they provided limited insight into tissue from scoliotic spines4. Inferior bone quality has been suggested as a possible cause of spinal curvature on the basis of measurements of bone density in children with idiopathic scoliosis and age-matched controls17,19,76,81,93. Although subjects with idiopathic scoliosis have different bone material properties, authors have reached different conclusions regarding whether this is etiologically important. At this time, we are not aware of any evidence in the literature supporting inferior bone quality as an important factor in the etiology of idiopathic scoliosis. The results of the bone-density studies are intriguing in that changes were found at skeletal sites remote from the spine that were presumably unaffected by the presence of the scoliosis. Unless these changes are caused by some subtle aspect of the biomechanics of scoliosis, bone quality may need to be investigated as a possible etiological factor. Studies of soft-tissue extensibility and joint laxity suggestive of reduced muscular or ligamentous stiffness have yielded contradictory results10,33,59. Joint laxity may be an important risk factor for the progression of scoliosis, but there is little evidence that it is an important etiological factor. One possible etiology of idiopathic scoliosis is asymmetrical loading of the spine, forcing collapse of the overloaded side. Trontelj et al.91 studied the stretch reflex in normal children, children with neuromuscular scoliosis, and children with idiopathic scoliosis. Those authors concluded that a localized neurogenic defect that causes asymmetrical muscle weakness is a primary factor in the etiology of idiopathic scoliosis. Carpintero et al.14 demonstrated that lordoscoliosis could be induced in a rabbit by tethering of the spine. Whether these findings are the cause of idiopathic scoliosis or the result of spinal deformity was not demonstrated. Sevastik et al.81 proposed a theory referred to as the thoracospinal concept, in which the etiology of thoracic curves is based on an assumption that the thoracic spine is predisposed to rotate to the right; scoliosis develops when this tendency is combined with overgrowth of the left-side ribs. They supported their contention with a series of experiments showing that scoliosis can be induced in rabbits by stimulating unilateral rib growth1 and that this scoliosis was similar to idiopathic scoliosis in humans83. Normelli et al.69,70, in a study of human ribs from normal subjects and subjects with right thoracic scoliosis, found that the ribs on the left are consistently longer than those on the right in people who have scoliosis. These findings were supported by parallel studies that suggested that the left side is more vascular than the right side and hence is likely to grow more71. This work culminated in a recent paper96 describing a unilateral rib-resection procedure used to treat a child who had a scoliosis of 46 degrees. As described earlier, the hypokyphotic spine may be sensitive to axial loads and at risk for buckling. The direction of the collapse depends on local conditions but is often determined by the lateral pressure exerted by the aorta. Somerville86 described this as rotational lordosis initiated by a growth difference in the thoracic spine, whereby growth of the posterior elements lags behind that of the anterior aspect of the vertebrae (Fig. 2). The amount of growth difference need not be great or symmetrical, but the greater the difference the higher the risk of collapse. The extent of the coronal plane asymmetry differentiates children with true scoliosis from normal children with coronal and transverse-plane asymmetry due to aortic loading. There is considerable circumstantial evidence to support this thesis21,23,25,26,86,98.
There is no strong scientific evidence implicating any particular biomechanical factor in the etiology of idiopathic scoliosis. Biomechanical properties of the structural elements of the spine, alignment of the spine, abnormal loading, and spine-support conditions are static mechanisms important to spinal alignment, but spinal stability as a mechanical process involves continuous realignment of the spine based on position-sensing on a local scale (at the vertebral level) and involving the head and trunk as well as the spine. This dynamic process might also lead to the development of scoliosis in the presence of a normal biomechanical spinal structure12. Research efforts to validate this concept have only recently been initiated.
Correlating a local pathological process and the associated spatial deformation can aid in understanding the etiology of idiopathic scoliosis. Several authors have tried mathematical modeling of the spine with use of different mathematical tools and kinematic assumptions4,23,25,26,33,66. The only solid data that can be acquired with few approximations or assumptions pertains to spatial arrangements of bones, expressed as coordinates in a Cartesian system66. These data, both static (spatial arrangement) and dynamic (range of motion), can now be acquired through imaging techniques and can be measured with precision.
Research into the etiology of idiopathic scoliosis has focused on multiple areas and has demonstrated the complex pathophysiology of this disorder (Fig. 3). Although idiopathic scoliosis may develop from infancy to adolescence, most of the work has focused on adolescent idiopathic scoliosis. It is clear that a complex and probably multifactorial process is involved.
Naturally occurring scoliosis in vertebrates is seen almost exclusively in humans, although a number of animal models for the condition exist. Many authors have observed an array of differences between convexity and concavity of the curve, but it has been difficult to distinguish causative factors from those that may result from the condition13,36,38,87,100. The current thinking is that there is a defect of central control or processing by the central nervous system that affects a growing spine and that the spine's susceptibility to deformation varies from one individual to another. Girls may be more vulnerable to this process because of the short and rapid adolescent growth of the spine compared with that in boys. It is possible that melatonin plays a secondary role in the development of idiopathic scoliosis. However, it seems unlikely that scoliosis results from a simple absence of melatonin. Rather, it could result from alterations in the control of melatonin production, with either direct or indirect consequences upon growth mechanisms. The control of growth is extremely complex and involves the interaction of many hormones and growth factors. A possible relationship between calmodulin and scoliosis has been proposed and should be studied further. In addition, studies of other growth factors in patients with idiopathic scoliosis are needed. Growth hormone and melatonin tend to have a reciprocal relationship, and the presence of one can influence the production of the other. Most researchers have doubted that a collagen abnormality is a primary etiological factor in the appearance or evolution of idiopathic scoliosis10,15,33. These abnormalities are more likely to be associated with the presence of secondary degenerative changes. The same may be true with regard to muscle changes. Many of the morphological changes found in paravertebral muscle have been found in other muscles but to a lesser degree99,103. These changes suggest the possibility of a generalized muscle defect, but they are most likely related to asymmetrical muscle-loading resulting from the deformity itself. The importance of the research of platelet abnormalities in patients with adolescent idiopathic scoliosis is based on the fact that a systemic defect may be present in platelets and skeletal muscle. A small percentage of thrombocytes in patients with idiopathic scoliosis are larger than normal. This finding has not been reported in controls. These changes in platelet morphology and physiology, which are similar to those found in skeletal muscle, suggest a cell-membrane defect that may be genetic in patients with idiopathic scoliosis. In attempting to develop a logistical model for causality on the basis of information in the literature, one is impressed by the volume of data that do not appear to be interrelated. The familial aspect of this disorder, however, allows the application of current molecular genetic techniques for the identification of the gene (or genes) involved in the development of scoliosis. Although idiopathic scoliosis is known to aggregate within families, the pattern of inherited susceptibility is unclear. Multiple contradictory reports emphasize the importance of consistent diagnostic criteria and rigid methods of disease ascertainment in order to provide substantial evidence in support of or against a proposed model of inherited susceptibility2,20,22,24,34,95. Once these criteria have been identified, it may be possible to integrate these multiple observations and to develop an understanding of the relationship between the genome, the maturing skeleton, and spinal deformity. Abnormal magnetic resonance imaging findings, microlesions in the brains of experimental animals, and the most consistent clinical neurological studies point to the pontine and hindbrain regions as the likely sites of primary pathology that could lead to idiopathic scoliosis. High-resolution magnetic resonance images of the midbrain and hindbrain, neurohistological analysis of brain specimens obtained coincidentally from patients with idiopathic scoliosis, and more sophisticated neurological studies should be regarded as high research priorities. The etiology of idiopathic scoliosis continues to elude investigators, although research has provided much useful information, which has been presented in this review. The consensus is that the etiology is multifactorial. With time, continued research will lead to the identification of the various factors involved in the causation of this disorder, which affects so many children and adolescents. Early identification will lead to earlier treatment and perhaps eventually to eradication of the disease itself.
|
Top
Introduction
Genetic Factors
