Background: In addition to genetic factors, environmental and lifestyle factors are thought to play an important role in the onset of adolescent idiopathic scoliosis (AIS). This cross-sectional study was conducted to explore lifestyle factors related to AIS.
Methods: This study included 2,759 Japanese female junior high school students who planned a secondary screening after an initial moiré topography screening indicated possible scoliosis. The students and their mothers, or guardians, were asked to fill out a questionnaire consisting of 38 questions about demographic factors, lifestyle-related factors, social factors, household environment, participation in sports, health status, and factors related to the mother’s pregnancy and delivery. The questionnaire was completed by 2,747 students (a 99.6% response rate). After excluding students with heart disease, neurological disease, or a congenital vertebral anomaly, 2,600 students were eligible for assessment. After undergoing a secondary screening with standing radiographs of the spine, students were assigned to the normal (control) group if radiographs showed a curve of <15° or to the scoliosis group if they had a curve of ≥15°. The odds ratios (ORs) for AIS in relation to the possible risk or preventive factors were estimated by logistic regression analyses.
Results: No lifestyle-related factor was significantly associated with AIS. However, AIS was associated with classical ballet training (OR, 1.38; 95% confidence interval [CI], 1.09 to 1.75); the odds of AIS developing increased as the child’s frequency of training, number of years of experience, and duration of training in ballet increased. The OR for AIS was 1.5 times higher for participants whose mothers had scoliosis. AIS was also associated with a low body mass index (BMI). These associations remained even after mutual adjustment was performed.
Conclusions: No association was found between AIS and lifestyle-related factors. However, classical ballet training, a family history of scoliosis, and low BMI may be associated with AIS.
Level of Evidence: Prognostic Level III. See Instructions for Authors for a complete description of levels of evidence.
Adolescent idiopathic scoliosis (AIS) is a structural, lateral curvature with rotation of the spine that develops in otherwise healthy children around puberty1. Epidemiological studies have estimated that 1% to 3% of the at-risk population (females from 10 to 16 years old) will develop some degree of spinal curvature1,2.
Much research has gone into determining factors that might influence the onset of AIS, including genetic factors3-10, environmental factors11, hormonal imbalances12-15, nervous system abnormalities16-19, and biomechanical factors20. However, none of these factors, except genetics, identified thus far have been widely accepted. The AIS concordance rate in monozygotic twins is 73% to 92%3-5, which strongly supports the importance of genetic background in the onset of AIS. However, the Danish Twin Registry21 and Swedish Twin Registry22 have reported lower rates, of 13% and 10%, respectively. The Swedish Twin Registry study also found a unique environmental effect of 60%23, which suggested that factors related to different intrauterine environments are important in the etiology of AIS3.
McMaster et al. reported that introducing infants to a heated indoor swimming pool influences the onset of AIS11. Associations between the onset of scoliosis and participation in sports, such as ballet dancing24, rhythmic gymnastics25, swimming26,27, and other athletics28, have also been reported. However, previous studies evaluating the roles of lifestyle and environment in the onset of scoliosis have limitations. Most of them were retrospective case-control studies, which have a bias in retrieved information. Possible confounding factors, such as home environment, social factors, and menarcheal status, were not properly adjusted for. Moreover, in many studies, the presence or absence of scoliosis was not radiographically confirmed. Thus, to clarify the etiology of AIS and identify lifestyle adjustments that might reduce the risk of AIS, the present study explored lifestyle and environmental factors that might be associated with AIS. To reduce the limitations faced by previous studies, we diagnosed AIS using standing radiographs and evaluated confounding factors, such as age, status of menarche, and social factors, in a large cohort of subjects.
Materials and Methods
Study Design and Participants
In Japan, students in elementary and junior high school are screened for scoliosis at 10 to 14 years of age. The students are first screened with moiré topography or consultation29, conducted by nearby doctors or companies without making radiographs. If scoliosis is suspected, the student is referred for a second screening with whole-spine radiography.
Our cross-sectional study included Japanese female junior high school students in the Tokyo metropolitan area who visited the Tokyo Health Service Association (Tokyo, Japan), a public service corporation that provides a variety of early-detection medical examinations, for a second screening for scoliosis after an initial screening conducted on school premises by the Tokyo Health Service Association suggested scoliosis. Of the students suspected to have scoliosis after the initial screening, 76.9% visited the Tokyo Health Service Association for their second screening and were evaluated with standing posteroanterior whole-spine radiographs. Of 2,759 female students who underwent the second scoliosis screening between January 2013 and February 2015 at the Tokyo Health Service Association, 2,747 (99.6%) participated in this study. None of these students had previously been diagnosed with scoliosis. After excluding students with congenital vertebral anomalies (n = 36), congenital heart disease (n = 20), epilepsy (n = 27), or spina bifida (n = 1), and those whose menarcheal status was unknown (n = 63), the remaining 2,600 students were included in our analyses. This study was approved by the ethics committees of the Keio University School of Medicine in Tokyo, Japan, and the Tokyo Health Service Association. Students and their guardians were given a pamphlet describing the study, a brief oral explanation, and a questionnaire. Those who completed the questionnaire were deemed to consent to participation in this study.
Students and their guardians were given the descriptive pamphlet and questionnaire when they arrived at the Tokyo Health Service Association for the second scoliosis screening, and then all students underwent standing posteroanterior radiographs of the spine. The students and their guardians completed the questionnaire before they received the results from the radiographs or any recommendations for further follow-up or treatment. This stepwise process was designed to minimize bias arising from a preoccupation with the presence or absence of scoliosis. The participants submitted their questionnaires at the reception desk before leaving the institution.
The self-administered questionnaire, designed to investigate factors related to lifestyle, physical activity, home environment, health status, sports participation, and the pregnancy and delivery experience of the participants’ mothers, was created by a panel of orthopaedic surgeons specializing in the treatment of scoliosis, biostatisticians, and epidemiologists after extensive discussion. The questionnaire contains 38 questions related to demographics, lifestyle, social factors and household conditions, health status, pregnancy and delivery, and sports participation. The questionnaire is presented in the Appendix.
In the second screening, height and weight were measured with the students in stocking feet and light clothing. Body mass index (BMI) was calculated in kg/m2. Menarcheal status was determined in an interview conducted by the orthopaedic surgeons involved in the study.
Standing posteroanterior radiographs of the whole spine were evaluated by orthopaedic surgeons specializing in the treatment of scoliosis. The Cobb angles of the major curves were measured if spinal curvature was present, and AIS was defined as a Cobb angle of ≥15°. The Cobb angles were measured using the DICOM (Digital Imaging and Communications in Medicine) POP (Post Office Protocol)-net server system (ImageONE, Tokyo, Japan). The present study met established standards for the interobserver reliability of the measurements30. The Risser sign (grade 0 to 5) was recorded for each participant as an indicator of bone maturity31. Any congenital anomaly of the vertebrae was also recorded for later exclusion from analysis.
All statistical analyses were conducted using the Stata statistical package (version 13; StataCorp). All statistical tests were 2-sided, and confidence intervals (CIs) were calculated at the 95% level. Odds ratios (ORs) for AIS in relation to the possible risk or preventive factors were estimated by logistic regression analyses. When calculating ORs, we adjusted for age, menarcheal status, and social factors, including public or private school (as a surrogate marker of household income) and family composition (single-parent family, 2-parent family, or 3-generation family). Because prolonged hypoestrogenism has been suggested as a factor in the development of AIS24, we included menarcheal status in the model to examine the independent effects of other factors, such as playing sports, on AIS. We did not adjust for maternal and paternal educational background as a social factor because these data were not available for all participants, and the results were not essentially changed by further adjustment for these factors. We also analyzed our results by excluding participants with a Cobb angle of 10° to 19° and recalculating the ORs for AIS for students with a Cobb angle of <10°, who did not need treatment, and those with a Cobb angle of ≥20°, for whom treatment for AIS was indicated.
Table I summarizes the background information of the participants. Most questionnaires (81.5%) were filled out by the participants and their mothers, 91.6% of the participants had reached menarche, and 68.6% showed a Risser grade of 4 or 5. The median Cobb angle value was 14° (interquartile range, 10° to 19°); 47.2% of the participants had a Cobb angle of ≥15° and were diagnosed with AIS.
Lifestyle, Home Environment, and Health Status
Lifestyle factors, including playing a musical instrument, the type and weight of school bags, time spent studying, sleep posture, and hours spent sleeping, were not significantly associated with AIS (Table II). Participants whose mothers had scoliosis had a 1.5 times higher odds of AIS compared with those with no family history of scoliosis.
We observed associations between AIS and a history of dental braces, a family member in the household who smoked, a decrease in weight of >1 kg since the previous year, and a low current BMI. The adjusted ORs were 1.24 (95% CI, 1.03 to 1.48) for dental braces, 0.82 (95% CI, 0.70 to 0.97) for a family member who smoked, 0.72 (95% CI, 0.55 to 0.94) for weight reduction, and 1.38 (95% CI, 1.17 to 1.63) for low BMI. In the mutually adjusted model, including all factors associated with AIS, only the association between dental braces and AIS was attenuated.
Classical ballet training was associated with increased odds of AIS (adjusted OR, 1.38; 95% CI, 1.09 to 1.75). In contrast, the odds of AIS decreased with playing basketball (OR, 0.69; 95% CI, 0.51 to 0.92) or badminton (OR, 0.61; 95% CI, 0.45 to 0.81) (Table II). These associations remained even after mutual adjustment for these 3 sports activities, a family history of scoliosis, a household member who smoked, a history of dental braces, weight loss, and current BMI was performed. To evaluate the associations between AIS and sports in detail, we assessed the age when the child began a sport, how many years she played, how often she played (frequency), and the duration (defined as a product of years played and frequency) (Table III). With respect to classical ballet, AIS was significantly associated with the start of ballet training at <7 years of age (preschool years) (OR, 1.38; 95% CI, 1.07 to 1.78). As the child’s years of ballet training, frequency of dancing, and duration of participation increased, there was a clear trend toward increased ORs for AIS (p value for trend, <0.01). For basketball and badminton, the frequency was inversely associated with AIS (p value for trend, <0.05); however, we could not evaluate associations between AIS and long-term involvement in basketball or badminton because students rarely started these sports before the age of 7 years (the median starting age for either sport was 12 years). The results from other sports are summarized in the Appendix.
Factors Related to Pregnancy and Delivery and Other Factors
We did not find measurable associations between AIS and factors related to pregnancy or delivery (Table IV). For example, compared with the group with the lowest maternal age at delivery (<25 years), the adjusted ORs for AIS were 1.02 (95% CI, 0.75 to 1.37) for those 25 to 29 years old, 1.15 (95% CI, 0.86 to 1.55) for those 30 to 34 years old, and 0.96 (95% CI, 0.69 to 1.34) for those ≥35 years old. The remaining factors not associated with AIS are shown in the Appendix.
Cobb Angles of <10° Compared with Angles of ≥20°
To further clarify the results of the present study, we compared factors for participants with a Cobb angle of <10° (who did not require AIS treatment) and those with a Cobb angle of ≥20°, for whom AIS treatment was recommended; the 1,222 participants with a Cobb angle of 10° to 19° were excluded (see Appendix). In this analysis, we saw tendencies similar to those in our main analysis.
In the present study, classical ballet training was associated with increased odds of AIS (OR = 1.38). This result was replicated in our evaluation of participants with Cobb angles of <10° or ≥20°, in which the odds increased to 1.58. We also observed clear increases in the odds of AIS as the years, frequency, and duration of classical ballet training increased. An association between classical ballet and AIS has been reported; Warren et al. reported that 24% of professional ballet dancers (mean age, 24.3 years) had scoliosis24. Since menarche was delayed in 83% of the dancers with scoliosis, those authors speculated that delayed menarche and prolonged intervals of amenorrhea might predispose ballet dancers to scoliosis24. The delay in menarche or amenorrhea in ballet dancers may have been caused by diet restrictions to maintain a leptosomatic figure. A meta-analysis by Arcelus et al. concluded that dancers had a 3-times higher risk of eating disorders, particularly anorexia nervosa and eating disorders not otherwise specified32. Since the present study adjusted for menarcheal status, our results indicated that ballet dancing itself might be a risk for the development of AIS.
The present study showed that playing basketball, badminton, or volleyball was associated with decreased odds of AIS. McMaster et al.11 suggested that children who develop AIS have a long-standing proprioception defect that makes them less likely to participate in sports. Another possibility is that girls with scoliosis were not drawn to the ball sports that require more dynamic, explosive athletic bursts, such as basketball or volleyball, since the curvatures may limit their ability to compete effectively in the ball sports.
Most parents or guardians are anxious about whether lifestyle-related factors, especially the weight and type of school bags, affect the onset or progression of AIS. Our results indicated that lifestyle-related factors were not associated with higher odds of developing AIS. We could not actually weigh the bags, since children usually came for the second screening on the weekend and were not carrying their school bags. Interestingly, 76% (1,978) of the students described their school bags as heavy.
The OR for AIS was 0.82 (95% CI, 0.70 to 0.97) for smoking by other family members in the household. However, smoking generally has a negative impact on the spine. Maternal exposure to cigarette smoking is a risk factor for congenital scoliosis33. Exposure to smoke may promote vertebral disc degeneration34,35. The present results indicate that passive smoking (secondhand smoke) may not be associated with AIS.
Participants whose mothers had scoliosis had 1.5-times higher odds of AIS than those whose mothers did not have scoliosis; these results agree with previous reports that AIS has a genetic component3-5.
Participants who had worn dental braces tended to have an elevated OR for AIS. Associations between malocclusion and scoliosis36,37 or kyphotic posture38 have been reported, but it has not been demonstrated whether masticatory disorders, such as malocclusions, can influence scoliosis or other aspects of whole-body posture39,40. We speculate that spinal curvature and disorders of the masticatory system may share a cause in common.
Low BMI had a significant OR of 1.38 (95% CI, 1.17 to 1.63) for AIS. Patients with AIS have been reported to have a lower BMI than the general population14,41. Smith et al. reported that, besides having a lower BMI in general, 25% of patients with AIS have BMI scores in the anorexic range14. A low BMI may be indicative of an eating disorder, excessive exercise, decreased bone mass, and/or hormonal imbalances, all of which might be associated with AIS. However, interpretation requires caution because the BMI distribution in our population (47% had a BMI of <18.5 kg/m2) was different from that in Western populations.
The intrauterine environment is crucial in programming the fetus for various health and disease outcomes throughout life42. However, the present study found no measurable associations between AIS and factors related to pregnancy and delivery (Table IV), including maternal age at delivery. A higher maternal age or paternal age at birth is reported to have a negative impact on the onset of AIS6,43-46. These findings may differ from ours because of the wide differences in study and control populations: previous studies included patients diagnosed with scoliosis on the basis of a Cobb angle of >10°43, subjects recruited from a single hospital43, control groups not recruited from the same situation as the subjects43, study populations that included boys6,43, comparisons made with the general population6, subjects who only had thoracic scoliosis44, and studies in which 67% of the patients had severe deformities44. Since nondifferential misclassification by memory recall might lead the associations toward the null in this study, further investigations are required.
Our present study has several limitations. First, all of the participants were suspected of having scoliosis on the basis of an initial screening with moiré topography. However, all participants were assessed radiographically for scoliosis and were assigned to the control group or the AIS group on the basis of the Cobb angle. Since guardians and students relied on memory to fill out the questionnaire, the information is less accurate than prospectively collected information. In addition, because of the cross-sectional nature of this study, temporal associations between AIS and the factors explored remain to be resolved. Finally, we cannot rule out the possibility of chance findings due to multiple comparisons performed in this study.
Our study’s strengths include a high response rate (99.6%), which minimizes selection bias. All of the participants were evaluated with standing radiography, which is the most accurate method for diagnosing scoliosis, ensuring that the students were accurately assigned to the control or the AIS group. The radiographs were read and measured by orthopaedic surgeons specializing in scoliosis, whose judgment of the presence or absence of AIS is highly reliable. Confounding factors including age, menarcheal status, and social factors were adjusted in the evaluation. Compared with previous studies, our study had the largest number of participants. All of the data were collected within 2 years, which minimizes bias arising from lifestyle changes of children of a particular age over time. Finally, observations based on differences between students with a Cobb angle of <15° and those with a Cobb angle of ≥15° were verified in analyses comparing students with a Cobb angle of <10° and those with a Cobb angle of ≥20°.
In conclusion, no association was found between AIS and lifestyle factors. However, classical ballet training, a family history of scoliosis, and low BMI may be associated with AIS. Thus, students who have these factors may require selective screening for AIS.
The questionnaire entitled “Survey of Children’s Lifestyle and Environment” and tables showing the association between participation in other sports and AIS, the associations between other factors and AIS, and the associations between potential risk factors and AIS after excluding participants with a Cobb angle of 10° to 19° are available with the online version of this article as a data supplement at jbjs.org.
Investigation performed at the Department of Orthopedic Surgery, Keio University School of Medicine, Tokyo, and the Tokyo Health Service Association, Tokyo, Japan
Disclosure: No external funding was received for this study. On the Disclosure of Potential Conflicts of Interest forms, which are provided with the online version of the article, one or more of the authors checked “yes” to indicate that the author had a relevant financial relationship in the biomedical arena outside the submitted work.
- Copyright © 2017 by The Journal of Bone and Joint Surgery, Incorporated