Adolescent idiopathic scoliosis (AIS) affects 0.47%–11.1% of the general population [1]. As the name suggests, the cause of the condition remains unclear. Several theories have been proposed, including genetic, neuro-developmental abnormalities, motor control and motor-sensory integration dysfunctions, vestibular and proprioceptive disorders, biomechanical growth modulation, uncoupled spinal neuro-osseous growth, thoracospinal concept, systemic and metabolic disorders [2], and decreased bone density [3-5] (Table 1).

Recently, Schlösser et al. [6] in 2014 performed a structured review of the literature regarding all the proposed and diverse etiologies proposed for the causes of AIS. They found that of all the theories, only those related to impaired gait control and decreased bone mineral density (BMD) had a moderate strength of evidence (Table 2). Other theories, including those of different volume of the cerebellar regions, asymmetric somatosensory evoked potentials, reduced trunk strength, decreased body weight, increased breast asymmetry, and impaired bone quality were only supported by weak evidence [6].

Considering that vitamin D may reportedly play a role in the cerebral process of postural balance [7] and that serum vitamin D level is positively correlated with the BMD of the hip [8,9] and negatively with the Cobb angle [10], vitamin D insufficiency or deficiency is speculated to influence AIS etiopathogenesis (Table 3).

In this study, we attempted to determine whether there is an association between vitamin D and the different current and relevant theories regarding AIS. We also assessed the impact of vitamin D deficiency or insufficiency on AIS development.

It has long been known that hereditary factors play a role in the etiology of idiopathic scoliosis (IS) [11-13]. In the previous 10–20 years, genetic association studies and genome-wide association studies (GWAS) have been used to identify single nucleotide polymorphism (SNP), namely, the variants of genes associated with AIS [14]. A genetic association study examines if one or more genes within a population co-occur with a disease more often than expected by chance occurrence, while GWAS involve the examination of the entire genome [15].

While several GWAS have identified different SNPs associated with AIS [16], they did not consider the specific mechanisms of how genes function within biological networks [14]. Pathway analysis has subsequently been developed to perform statistical tests on the lists of genes involved in the biological pathways for a significant relationship with a particular phenotype or observable trait [17].

Favaro [14] in 2017 reported the results of a pathway analysis of a list of genes affected in AIS. He found that the multiple pathways were significantly affected in AIS, including those of cancer; the endocrine system; the lung and heart stress response and repair; stem cell differentiation; as well as skeletal tissue growth, development, and repair.

It is noteworthy that vitamin D was involved in some pathways, including the spinal cord injury pathway, the lung fibrosis pathway, the endochondral ossification pathway, and the endocrine resistance pathway [14].

Abnormalities in the morphology and function of the central nervous system have long been implicated in the etiopathogenesis of AIS. Cervicothoracic syrinx [33,34], low-lying cerebellar tonsils [35-37], and morphological abnormalities in the midbrain [38], pons medulla [39], and vestibular system [40] have all been reported in subgroups of AIS subjects.

Functionally, AIS patients are reported to have a range of abnormalities in postural balance, proprioceptive disturbance [41,42], oculo-vestibular function [43], vestibular-evoked postural response [44], somatosensory functions [45-47], and motor control [48]. Beaulieu et al. [47] in 2009 and Haumont et al. [49] in 2011 both showed that postural balance dysfunction is related positively to the severity of curvatures. AIS patients with greater curvatures and poorer balance control use fewer anticipatory strategies to stabilize body oscillations [49].

Several studies have reported that the skeletal growth pattern in AIS subjects is disturbed [73-77]. This has been attributed to the age of onset of menarche [76,77], reduced serum vitamin D level, and/or polymorphism or variations in the VDRs [78].

AIS subjects are reportedly taller than healthy controls [73-77]. However, after the age of 13 years, the height difference between AIS and control subjects disappeared [77,79]. With respect to the sexual maturity, Siu King Cheung et al. [80] in 2003 reported that girls with AIS were shorter before menarche; however, they grew taller and had a longer arm span during the growth spurt compared to healthy controls. Studies have shown that there are insignificant differences in the height between patients and controls at the end of the growth period, and the growth pattern of AIS patients has thus far been regarded as essentially normal [76,77]. In fact, in a recent review, Schlösser et al. in 2014 stated that there was only weak evidence in support of corrected body height changes as an etiological factor of AIS [6].

These different findings may be related to the age of onset of menarche reported in different studies [76,77,80]. Ylikoski [76] in 2003 and Goldberg et al. [77] in 1993 reported that AIS girls experience earlier menarche with an earlier growth spurt. Other authors have reported contradictory findings [81-84]. Warren et al. [81] in 1986 reported a higher prevalence of scoliosis in ballet dancers, all of whom had delayed menarche. Mao et al. [82] in 2011 found that the average age of onset for menarche in Chinese girls with AIS girls was significantly later than in normal controls. Yim et al. [83] in 2012 also reported that Chinese AIS girls had an average delay of 0.49 years in the age of onset of menarche compared to healthy controls.

Many factors influence the age of onset of menarche. These include genetics, estrogen level, vitamin D status, polymorphism of VDR, and latitude [78,85] (Table 4). Twin studies have shown that as much as 68% of the variance in age at menarche may be genetically determined [85]. A decrease in the estrogen level is related to a delay in menarche [82]. Likewise, vitamin D deficiency and VDR polymorphism affect the timing of menarche [78,86]. Villamor et al. [86 ]in 2011 showed that vitamin D3 deficiency is associated with an early onset of menarche by as much as 9 months. Latitude reportedly affects the time of onset of menarche, possibly through its influence on melatonin secretion and vitamin D synthesis [87-89]. Girls who live at higher latitudes appear to experience menarche later than girls who live closer to the equator [87-89]. Grivas et al. [87] in 2006 attributed the differences in the age of onset of menarche in different latitudes to the possible differences in melatonin secretion [87]. However, it is also possible that the differences in menarche onset are related to the serum vitamin D level. At higher latitudes where the ultraviolet radiation is less intense, the cutaneous synthesis of vitamin D is decreased [88,90]. This may also be related to the delay in menarche onset. Dossus et al. [88] in 2013 showed that on a continuous scale, an increase of 1° latitude was associated with 0.04 years older age of onset of menarche, whiles a 1-kJ/m² increase in the annual ultraviolet radiation dose resulted in a 0.42 years younger age at menarche. Hagenau et al. [90] in 2009 showed that there was a 25(OH)D level decrease with latitude in European Caucasians; however, this was not observed in non-Caucasians. Studies have shown that an increase in the serum vitamin D level is associated with earlier menarche onset and vice versa [88,90].

Overall, the results are controversial. Studies have shown that both low and increased serum vitamin D level are associated with an earlier age of onset of menarche [86,88]. The contradictory findings may be attributable to the confounding effects of childhood obesity and higher leptin levels [91]. In obese girls, more serum 25(OD)D is bound to the adipose tissues. Moreover, the higher amount of adipose tissues increases the tissue dilution of the vitamin, resulting in lower serum 25(OH)D levels, influencing the outcome.

Furthermore, the mechanism underlying the association of vitamin D with the timing of menarche has not been clearly elucidated. It is possible that vitamin D acts through estrogen to affect the timing of menarche. The deficiency of vitamin D may reduce estrogen synthesis [30] and the expression and activity of several enzymes related to the production of sex steroids [92]. In addition, the interactions between the signaling pathways of estrogen and 1,25-vitamin D3 have also been identified [93].

Osteopenia is a condition characterized a lower than normal BMD. In particular, it is defined as BMD with a T-score between −1.0 and −2.5. T-score is the number of standard deviations above or below the mean value for a healthy 30-year-old adult of the same sex and ethnicity as the patient. Recent studies have shown that 29.5% [94] to 38% of AIS patients have osteopenia [3,95,96]. Burner et al. [97] in 1982 were the first to report relative osteoporosis in children with IS. Similarly, Velis et al. [98] in 1988 used the Singh femoral trabecular index, a classification system for the bone density of the femoral neck based on the visibility of the trabecular types that can be seen in the femoral neck to show that scoliosis patients presented with early trabecular bone loss in the hips. Thereafter, osteopenia in AIS patients has been confirmed in many studies [95,99-101] using dual photon-absorptiometry and dual energy X-ray absorptiometry. Osteopenia has been found not only in the spine [3,99], but also in the radius [102], hips [99-101], and tibia; therefore, the decrease in bone density is a generalized rather than a local phenomenon.

Follow-up studies have also shown that the bone quality in AIS patients is also affected [100,102]. Tam et al. [102] in 2014 found that the cortical density of the radius of the non-dominant hand of AIS patients was significantly reduced as compared to that in healthy controls. The trabecular density, trabecular number, and thickness also reportedly decreased [102], making the patients prone to fracture.

Further, the presence of a generalized reduction in BMD prompted Cheng et al. [3] in 1999 to suggest that osteopenia may have an etiopathogenetic role in AIS. Lee et al. [4] in 2005 demonstrated that BMD was inversely related to curve severity in 919 girls with AIS during the peri-pubertal period. AIS patients with lower BMD generally had a higher Cobb angle, while those with normal BMD had smaller curves [4]. This concurred with the findings of Wang et al. [101] in 2017 In fact, Hung et al. [5] in 2005 who followed up a cohort of 324 girls with AIS until skeletal maturity or progression of curves ≥6° found that osteopenia was a significant prognostic factor for curve progression.

Lee et al. [4] in 2005 compared the calcium intake and time spent in physical activities in 596 girls with AIS girls and 302 age-matched healthy controls. They found that the median calcium intake of AIS patients did not differ significantly from that of healthy controls. However, the median time spent by the former (AIS patients) in physical activities was slightly less than that spent by the controls [5]. Other studies have shown that osteopenia observed in AIS patients is related to vitamin D insufficiency and deficiency [10,103,104]. Longitudinal and cross-sectional studies have also demonstrated that the 25(OH)D level positively correlated with BMD in healthy children and adolescents [105-108]. The serum 25(OH)D level of pre-pubertal children at ages 7.6, 9.9, and 11.8 positively correlated with cortical bone mineral content (BMC), cortical thickness, and predicted bone strength at the age of 15.5 years [108]. Pubertal girls with lower vitamin D levels were at risk of not reaching the maximum peak bone mass, particularly at the lumbar spine [105]. Moreover, Balioglu et al. [10] recently demonstrated a negative correlation between serum vitamin D levels and the Cobb angle. Thus, these results further suggest that vitamin D may play an etiopathologic role in AIS development.

The age of onset of menarche is a determinant factor of peak BMD in young girls. Fox et al. [109] in 1993 and Tuppurainen et al. [110] in 1995 reported that BMD is related to the age of onset of menarche. Early menarche is associated with higher circulating estrogen levels during and after menarche [111,112] and a high peak bone mass [113]. Late menarche is associated with a low BMD [109,110,113]. A Chinese cohort study found patients aged <40 years who had experienced late menarche (at age >14 years) had an approximately two-fold higher risk of low areal BMD at the hips [114]. Similarly, a Korea National Health and Nutrition Examination Survey (2008–2011) showed late menarche is associated with a lower BMD in the lumbar spine [114]. The age at menarche is correlated not only with BMD, but also with bone microstructure [115]. Young healthy girls with a higher age of onset of menarche displayed lower total volumetric BMD, cortical volumetric BMD, and cortical thickness at the forearm than those with a lower age of onset of menarche [115].

Many hormones and metabolites have been implicated in the etiopathogenesis of IS. Among these, estrogen, melatonin, and leptin are important [116-121]. Platelet calmodulin and osteopontin were possibly of secondary importance in AIS pathogenesis. It is possible that all these hormones and proteins interact to influence BMD and vertebral bone quality, thereby affecting AIS development.

Currently, AIS is considered a multifactorial heterogeneous disease [160], with genetic and environmental factors influencing its etiopathogenesis [161,162]. Vitamin D deficiency and osteopenia were prevalent among AIS patients [10] and were positively correlated with the Cobb angle. Therefore, we speculated that vitamin D deficiency may play a role in AIS development, possibly by influencing fibrosis regulation [163], postural control, and regulation of bone metabolism [10] (Table 5).

In children with vitamin D deficiency, spinal cord or spinal column injury in children may be repaired by fibrosis, setting the background for the tethering of the nerve roots. Abe et al. [24] in 2017 injured the right side of the paravertebral muscles of 30, 8-week-old male rats. Thereafter, they performed histological examination and comparison of the muscle fibers and dorsal root ganglia on both the sides 1–3 weeks post-intervention. They found the cicatrix formed after the injury may sensitize the dominant nerve in the dorsal root ganglia [24]. Li et al. [25] in 2016 studied the effects of compression of the sciatic nerve in rats to investigate the pathological alterations in the neurons of the dorsal root ganglia [25]. They reported at 3 weeks after the compression, collagen fibers accumulated around the compressed sciatic nerve, and at 8 weeks, excessive collagen formation with muscle atrophy was observed [25]. The volume of collagen gradually and significantly increased following sciatic nerve compression. The amount of collagen I fibers surrounding the ipsilateral dorsal root ganglion was higher than that on the contralateral side [25], suggesting that injuries to the paravertebral muscles and nerve roots are healed with fibrosis. It is possible that in a subset of AIS patients with vitamin D deficiency, spinal injury in childhood is healed by fibrosis, with resultant nerve root tethering. As per clinical reports, fibrosis and scar tissue formation after surgery in childhood may be associated with AIS [164,165]. Lebel and Lebel [164] in 2016 reported a case of a female patient with AIS who underwent surgical excision of a 4-cm benign ganglioblastoma from the left posterior rib cage at the age of 3 years. Post-surgical radiographs at the age of 8 years showed fusion of left 4th and 5th ribs. The patient was diagnosed with AIS at 11 years of age [164]. Similarly, Brooks et al. [165] in 2009 reported a case of AIS, possibly secondary to right inguinal herniorrhaphy performed before 6.25 years of age. By the age of 6.5 years, an asymmetric posture was apparent in a casual photograph, and after puberty, a marked deformity was evident [165].

Artaza and Norris [163] in 2009 exposed mesenchymal multipotent cells to the active form of vitamin D. They found that vitamin D decreased the expression of transforming growth factor β 1 and the plasminogen activator inhibitor, both of which are profibrotic factors, leading to reduced collagen deposition.

Vitamin D deficiency is also found to reduce postural control; supplementation of vitamin D improves the postural control in elderly individuals with vitamin D deficiency [69-71]. Whether this has any implications in AIS subjects warrants further investigation.

Insufficiency and deficiency of vitamin D may also play a role in AIS pathogenesis through its effects on bone metabolism. Vitamin D possibly interacts with estrogen, melatonin, and leptin to affect the BMD and BMC of the bones, including the vertebrae [3,9,10,30,131,146,158]. Recently, vitamin K2 was also found to be important for regulating bone metabolism. Vitamin K2 facilitates the deposition of calcium in the bones by the carboxylation of osteocalcin, the synthesis of which is induced by 1,25-(OH)2D3 [166 ]. Vitamin D deficiency reduces the carboxylation of osteocalcin and reduces the BMC of the total body and the lumbar spine [167]. It is noteworthy that many peri-pubertal healthy adolescents have subclinical deficiency of vitamin D [105,168-170] and vitamin K2 [167,171,172]. This may be attributable to the high bone turnover in children [105,172] and the fat-restricted diet that is commonly enforced by the parents, given the fact that both vitamin D and K2 are fat-soluble vitamins. Deficiency of vitamin D, vitamin K2, estrogen, melatonin, and leptin levels is associated with a reduction in the BMD [8,9,129,150,173]. It is possible that the resultant osteopenia from the deficiencies make the vertebrae susceptible to wedge deformity in the early stages of AIS [174-181], contributing to the initiation and/or progression of the curvature.

Recent studies have shown that asymmetric loading of the osteopenic vertebrae may cause vertebral wedging through creeping, a continuous deformation under constant load [182,183]. Lowered vertebral BMD is associated with a higher Cobb angle and increased risk for curve progression [4,5,184]. Studies have shown that sustained compressive loading can cause progressive and measurable creep deformity in osteopenic elderly human vertebrae [182,183], with a mean BMD <0.15 g/cm³ that results in vertebral wedging [182]. Wren et al. [185] in 2017 reported that lateral thoracic vertebral body wedging is negatively correlated with the vertebral cross-sectional area (CSA) in girls with AIS. Thus, vertebral body wedging is more likely to occur in the thoracic spine that has a small vertebral CSA, particularly in the presence of osteopenia [186]. In skeletally immature patients, continuous asymmetric growth on the concave and convex sides of a scoliosis curve would increase vertebral wedging, further contributing to an increase in the curvature [185,187].

Our literature review demonstrated an association between vitamin D status and BMD, menarche, and other hormones. Although lower vitamin D levels are correlated with the Cobb angle, the relationship between vitamin D and AIS may not be casual. Further studies are required to determine whether vitamin D definitely plays a casual role in the etiopathogenesis of AIS and whether vitamin D deficiency is related to the tethering of the nerve roots and postural control dysfunctions observed in some AIS subjects. Longitudinal studies may also be required to examine the effects of fat-restricted diets as well as vitamin D and K2 supplementation on the progression and management of the scoliosis curves.

Further, it may be interesting to determine if the widespread deficiencies of vitamin D and K2 in adolescents are related to the increased need for vitamin D and K2 during pubertal growth and/or to the adherence to the now challenged fat- and cholesterol-restricted diets [188] that have been promoted since the early 1980s to reduce the risk of cardiovascular diseases.