Trends in the Prevalence of Adolescent Idiopathic Scoliosis in a Japanese Prefecture: A 25-Year Population-Based School Screening Study Using Moiré Topography

Abstract

Study Design

Retrospective analysis of a prospectively maintained school-based screening database.

Objective

To investigate long-term trends in the prevalence of adolescent idiopathic scoliosis (AIS) in a Japanese population over 25 years.

Methods

This study evaluated a screening program conducted from 1998 to 2022 in Ehime, Japan. All fifth- (ages 10-11) and seventh-grade (ages 12-13) students were included. Primary screening used moiré topography (Tsukidate criteria). Students with positive findings were referred for radiographic confirmation (AIS defined as Cobb angle ≥10°). Linear regression assessed temporal trends in positivity rate, positive predictive value (PPV), and estimated prevalence.

Results

Of 534 322 examinations performed, the overall positivity rate increased significantly from 2.24% in 1998 to 4.06% in 2022. In a sub-cohort with complete radiographic follow-up (n = 221 318), estimated AIS prevalence showed a significant linear increase from 0.85% (initial 5 years) to 2.14% (final year). Notably, PPV remained stable throughout the period (mean 47.3%) with no significant trend.

Conclusions

AIS prevalence increased substantially over 25 years in this cohort. The stability of the PPV suggests a true epidemiological increase rather than a diagnostic artifact. These findings highlight the importance of continued screening and suggest lifestyle or developmental factors may contribute to the rising incidence.

Keywords

adolescent idiopathic scoliosis school scoliosis screening moiré topography prevalence positive predictive value

Introduction

Adolescent idiopathic scoliosis (AIS) is the most common pediatric spinal deformity, characterized by a three-dimensional spinal curvature that develops between approximately 10 years of age and skeletal maturity. With a generally accepted prevalence of 2-4%,¹ AIS is common in girls and typically manifests during the peripubertal growth spurt. Early detection is paramount, as nonoperative interventions, such as bracing, can effectively halt curve progression,^2,3 thereby reducing the need for surgical correction. Consequently, school-based screening programs have been widely implemented as the primary means of identifying at-risk individuals in treatable stage.⁴

Evidence on temporal trends in AIS prevalence is limited. Internationally, reports are scarce, and their findings are inconsistent. Long-term studies based on large-scale school screening programs in Singapore and Hong Kong have reported that the prevalence of AIS has increased over time. Wong et al demonstrated a significant increase in prevalence among 11-12-year-old girls in Singapore from 1982-1997, positing that earlier physical maturation was a contributing factor.⁵ Fong et al reported a trend of increasing prevalence from 1995-2000 in a cohort study of approximately 400 000 individuals in Hong Kong.⁶ Contrastingly, a United States report stated that after school screening was discontinued in 2004, the incidence of newly diagnosed AIS decreased compared to the screening period.⁷ Thus, long-term trends in AIS prevalence remain unclear, with few long-term reports from Japan.

Although earlier menarche has been reported in other Asian countries and suggested as influencing prevalence,^5,6 national survey data from Japan show that the average age at menarche has remained remarkably stable at approximately 12 years and 2 months since 1997, spanning the study period.^8-10 This observation led to our initial hypothesis that the prevalence of AIS in Japan remains stable. We posit that the dataset from Ehime Prefecture used holds unique value. This dataset is rare globally, representing 25 years of consistent screening (since 1998) in a cumulative cohort of >530 000 participants. Critically, all screenings were performed using the same Moiré topography system and a single, unchanging diagnostic standard (Tsukidate criteria).¹¹ This methodological consistency minimizes confounding because of temporal changes in diagnostic criteria, allowing accurate assessment of true temporal trends in AIS prevalence. Therefore, this study aimed to leverage this longitudinal dataset to quantitatively determine temporal trends in the screening positivity rate, positive predictive value (PPV), and estimated prevalence of AIS in a regional Japanese population.

Materials and Methods

Study Design and Population

We conducted a retrospective review of a prospectively maintained database from a school-based scoliosis screening program in the Ehime Prefecture, Japan. This cohort study covered fifth (aged 10-11 years) and seventh grade (aged 12-13 years) students in public schools who were screened from 1998-2022.

Screening Protocol (Primary Screening)

Primary screening was conducted using a moiré topography system (Fujinon FM-40; Fujifilm Corp., Tokyo, Japan) to capture standardized topographic images of each student’s back in a relaxed standing position. Moiré topography is a non-invasive optical method that generates a contour map of the back surface, allowing the detection of truncal asymmetry associated with scoliosis. All captured images were anonymized and transferred to the Department of Orthopedic Surgery at the university hospital for centralized review by a team of specialist spine surgeons.

A positive or negative screen was determined according to the Tsukidate classification,¹¹ a standardized criteria that was applied consistently throughout the study period. This classification system divides the back into 4 quadrants (right and left thoracic and right and left lumbar regions) and quantitatively scores the asymmetry of the Moiré fringe patterns and truncal shift. A screen was deemed positive if the composite score exceeded sex- and grade-specific cutoff values (Figure 1).

Figure 1.

The Tsukidate classification for moiré topography.

Diagnostic Confirmation (Secondary Screening)

Students with a positive primary screen were formally recommended to undergo a secondary evaluation by a community-based orthopedic specialist. This secondary workup included a standing full-length posteroanterior spinal radiograph. A definitive diagnosis of AIS was made as the Cobb angle measured ≥10° on a standing PA radiograph, which is the standard diagnostic threshold for scoliosis.

Data Collection and Outcomes

A systematic data feedback mechanism was in place for a subset of schools primarily located within Matsuyama City, representing approximately 41.4% of the cohort. For students from these schools, the secondary screening results, including the measured Cobb angle, were reported to the university hospital via the local public health association. This created a traceable cohort that enabled the calculation of performance metrics of the screening program.

The primary outcome measures for this study were defined as follows:

1. Positivity Rate: The proportion of all screened students who had a positive moiré topogram, calculated as number of positive screens/total number of students screened.

2. PPV: The proportion of students with a positive screen result who were subsequently confirmed to have AIS. This was calculated within the traceable cohort as number of confirmed AIS cases with Cobb angle ≥10°/number of students with a positive screen who attended secondary screening.

3. Estimated Prevalence: The prevalence of radiographically confirmed AIS within the traceable cohort, calculated as number of confirmed AIS cases with Cobb angle ≥10°/total number of students screened in the traceable cohort.

Statistical Analysis

Linear regression analysis was used to assess the temporal trends in the annual positivity rate, PPV, and estimated prevalence over a 25-year study period. The coefficient of determination (R²) was calculated to quantify the proportion of variance in each outcome variable predicted from the year, thereby assessing the strength of the linear relationship. Statistical significance was set at P < 0.05.

Ethical Considerations

This study was approved by the Institutional Review Board of the HITO Hospital (IRB number: 20240529001). In accordance with local ethical guidelines for large-scale public health screening analyses, an opt-out method of consent was used. The study plan was publicly announced on the hospital’s clinical research website, allowing individuals or their guardians to opt out of the study.

Results

Overall Cohort and Positivity Rate

A total of 534 322 moiré topography screenings were performed from 1998-2022. Of these, 15 792 tested positive, with an overall mean positivity rate of 2.96% (Table 1). A clear and consistent temporal trend was observed in the annual positivity rate, demonstrating a strong and statistically significant linear increase over time. The rate increases from 2.24% (1998) to 4.06% (2022) (R² = 0.84, P < 0.001) (Figure 2).

Table 1.

Annual Results of Primary Moiré Topography Screening From 1998-2022

Fiscal year	Number of examinees	Number of positive	Positivity rate (%)
1998	21 863	490	2.24
1999	24 116	397	1.65
2000	20 354	365	1.79
2001	22 388	402	1.80
2002	19 383	373	1.92
2003	20 499	433	2.11
2004	19 299	390	2.02
2005	22 362	470	2.10
2006	22 031	384	1.74
2007	21 839	400	1.83
2008	22 450	529	2.36
2009	22 382	669	2.99
2010	22 263	739	3.32
2011	22 130	800	3.62
2012	21 767	746	3.43
2013	21 738	750	3.45
2014	22 041	871	3.95
2015	20 971	891	4.25
2016	20 720	797	3.85
2017	20 866	806	3.86
2018	20 862	818	3.92
2019	20 849	809	3.88
2020	20 653	835	4.04
2021	20 386	812	3.98
2022	20 110	816	4.06
Total	534 322	15 792	2.96

Figure 2.

Annual positivity rates in the entire cohort.

Traceable Follow-Up Cohort

The subcohort for which secondary screening results were systematically collected comprised 221 318 students, representing 41.4% of the total screened population (Figure 3). Within this traceable cohort, 6893 students screened positive, corresponding to a rate of 3.11%. Of these, 5832 students (84.6%) underwent the recommended secondary screening for radiographic confirmation. The trend of an increasing positivity rate was present and highly significant within this subcohort, mirroring the findings from the overall population (R² = 0.61, P < 0.001) (Figure 4).

Figure 3.

Flowchart of the study cohort.

Figure 4.

Annual positivity rates in the traceable cohort.

PPV

Among the 5832 students in the traceable cohort who had a positive screen and underwent secondary evaluation, 2757 were diagnosed with AIS (Cobb angle ≥10°). This resulted in an overall mean PPV of 47.3% over the study period (Table 2). Although there were annual fluctuations in PPV, linear regression analysis revealed no statistically significant temporal trends (Figure 5). PPV stability is a critical finding of this study.

Table 2.

Outcomes in the Traceable Follow-Up Cohort From 1998-2022

Fiscal year	Number of examinees	Number of positive	Positivity rate (%)	Number of examinees who underwent secondary screening	Positive predictive value (%)	Number of examinees with a Cobb angle	Estimated prevalence (%)
1998	10 274	398	1.57	397	35.26	140	1.36
1999	9784	192	1.80	161	49.07	79	0.81
2000	9786	235	2.20	201	39.80	80	0.82
2001	9323	169	1.73	136	43.38	59	0.63
2002	9076	208	2.32	169	46.15	78	0.86
2003	8838	205	2.39	179	42.46	76	0.86
2004	9113	192	2.11	143	55.94	80	0.88
2005	9411	173	1.84	122	46.72	57	0.61
2006	8951	136	1.52	108	62.96	68	0.76
2007	9299	201	2.16	143	39.86	57	0.61
2008	8887	197	2.22	172	50.00	86	0.97
2009	9068	267	2.94	233	51.50	120	1.32
2010	8855	324	3.66	276	41.67	115	1.30
2011	8825	316	3.58	275	48.36	133	1.51
2012	8685	303	3.49	266	51.88	138	1.59
2013	8678	300	3.46	259	49.03	127	1.46
2014	8661	349	4.03	295	47.46	140	1.62
2015	8387	360	4.29	298	41.61	124	1.48
2016	8256	315	3.82	276	42.75	118	1.43
2017	8309	349	2.39	302	53.31	161	1.94
2018	8131	313	3.85	262	48.09	126	1.55
2019	8225	345	4.19	278	48.56	135	1.64
2020	8270	334	4.04	285	47.02	134	1.62
2021	8033	311	3.87	269	56.13	151	1.88
2022	8193	401	4.89	327	53.52	175	2.14
Total	221 318	6893	3.11	5832	47.27	2757	1.25

Figure 5.

PPV over time in the traceable cohort.

Estimated Prevalence of AIS

The overall mean estimated prevalence of AIS (Cobb angle ≥10°) in the traceable cohort was 1.25% (2757 cases out of 221 318 students). The primary finding of this study was a significantly strong positive linear trend in estimated prevalence over time. The annual prevalence, which was frequently <1.0% in the early 2000s, increased steadily to 2.14% in 2022 (Table 2),which was statistically significant (R² = 0.74, P < 0.001) (Figure 6).

Figure 6.

Estimated prevalence of AIS over time in the traceable cohort.

Subgroup Analyses

Stratification by sex and grade revealed marked differences in PPV and prevalence. Both metrics were consistently higher in girls than in boys, and in seventh grade students than in fifth grade students (Figures 7 –9). This disparity was striking for the estimated prevalence among seventh grade students, which reached 3.36% in girls, >9 times higher than the 0.36% observed in boys of the same age. The prevalence of more significant curves requiring clinical monitoring or intervention (Cobb angle ≥25°) was 0.29% overall, with the highest rate found in seventh grade girls (0.81%).

Figure 7.

Distribution of Cobb angles in patients undergoing secondary screening.

Figure 8.

Secondary screening results for fifth grade students.

Figure 9.

Secondary screening results for seventh grade students.

Discussion

This study, based on a 25-year methodologically consistent school screening program, provides strong evidence that the estimated prevalence of adolescent idiopathic scoliosis has significantly increased in the regional Japanese population. The primary finding of the near-doubling of prevalence was substantiated by the critical observation that the positive predictive value of the screening test remained stable over time.

PPV stability serves as the methodological keystone of this study. A common criticism of longitudinal screening studies is the potential for diagnostic drift in which interpretation standards may change over time, leading to an artificial increase in case detection. If spine surgeons interpreting the moiré topograms had become more liberal in their criteria over the past 25 years, the positivity rate would increase. However, this would have been accompanied by an increase in false-positive screens, which would have decreased PPV. The fact that the PPV remained stable while the positivity rate and estimated prevalence increased significantly provides powerful evidence that the observed trend reflects a true increase in the underlying prevalence of the condition in the screened population. Although the absolute accuracy of moiré topography has been debated in the literature, its consistent application in this program, using the same hardware and unchanging human-interpreted criteria, created a stable longitudinal yardstick. Any systematic error inherent in the tool remains constant over time, indicating that the observed change in prevalence is a robust finding.

The trend of increasing prevalence observed in our cohort aligns with reports from other Asian settings, such as Singapore⁵ and Hong Kong,⁶ in which similar increases were noted in studies conducted in the 20th century. Our findings contrast with those of a United States study that reported a decrease in the incidence of newly diagnosed AIS after the discontinuation of school screening. This divergence reflects a difference in methodology; the US findings probably represent a detection artifact in which fewer mild cases were identified in the absence of a systematic screening program. Contrastingly, our study, with its continuous and consistent screening, captured a genuine underlying epidemiological shift.

To explain this trend, we propose a multifactorial hypothesis that implicates a synergistic interaction between secular trends in adolescent development and profound changes in modern lifestyles over the past quarter-century. First, although the age of menarche has been stable in Japan,^8-10 national data indicate that children are, on average, taller and heavier for their age than in previous decades.¹² A more rapid pubertal growth spurt is a well-established biomechanical trigger for the progression of spinal deformities.¹³ This accelerated growth velocity may unmask or amplify latent spinal asymmetries. Second, the study period of 1998-2022 coincides perfectly with the digital revolution,¹⁴ leading to a dramatic increase in sedentary behavior and screen time among adolescents, with a corresponding decrease in overall physical activity.^13,15,16 Chronic poor posture and deconditioning of the core and paraspinal musculature may create an environment conducive to the initiation or progression of scoliosis. Our hypothesis posits that this modern sedentary lifestyle may initiate a cascade of biomechanical deficits, whereas the concurrent phenomenon of accelerated growth provides “fertile ground” for these deficits to rapidly progress into clinically significant scoliosis, thereby amplifying the “vicious cycle” of deformity progression.

The PPV of 47.3% was comparable to that of other screening techniques, indicating the robust performance of our program.^3,17-20 Our secondary findings have important clinical implications. The consistently lower PPV observed in boys than in girls (31.1% vs 56.2% in seventh grade students) suggests that the Tsukidate criteria may be overly sensitive for males, leading to a higher rate of unnecessary referrals. The development and implementation of sex-specific screening thresholds can significantly improve the efficiency of screening programs by reducing the burden of false-positives in boys.

This study has some limitations. First, as only screen-positive children were referred for radiography, we could not assess the false-negative rate; thus, the true sensitivity, specificity, and negative predictive value cannot be calculated. Second, the dataset was cross-sectional during diagnosis and excluded longitudinal data on curve progression or treatment outcomes. Third, the prevalence estimates were derived from a 41.4% subcohort for which follow-up data were available, introducing selection bias if this group is not fully representative of the entire population. Fourth, our proposed etiological hypothesis is subject to ecological fallacy as we did not have individual-level data on lifestyle factors or growth parameters to directly test these associations. Finally, although we interpreted stable PPV as evidence of a consistent screening standard, rising disease prevalence would theoretically be expected to increase the PPV. The absence of a significant upward trend in PPV may suggest that the magnitude of the prevalence increase was not large enough to be statistically detected in the PPV analysis amidst annual variations or that other unmeasured factors were at play. However, this complex relationship requires further investigation.

Conclusion

This 25-year study, distinguished by its scale and methodological consistency, provides robust evidence of a true and significant increase in the prevalence of AIS in a regional Japanese population. PPV stability over this long period strongly supports the conclusion that this trend is a genuine epidemiological phenomenon and not an artifact of changing diagnostic standards. We propose the multifactorial hypothesis that this trend may be driven by the synergy between modern sedentary lifestyles and secular changes in adolescent growth patterns. These findings reaffirm the ongoing value of school-based screening for the early detection of AIS and highlight the urgent need for further research on the environmental and developmental drivers of this common spinal deformity.

Footnotes

ORCID iDs

Tadao Morino

Yusuke Murakami

Ethical Considerations

This study was approved by the Institutional Review Board of HITO Hospital (IRB number: 20240529001). An opt-out method for consent was utilized via public notification on the institution’s website, in accordance with local ethical guidelines.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

References

Choudhry

Ahmad

Verma

. Adolescent idiopathic scoliosis. Open Orthop J. 2016;10:143-154. doi:10.2174/1874325001610010143

Weinstein

Dolan

Wright

Dobbs

. Effects of bracing in adolescents with idiopathic scoliosis. N Engl J Med. 2013;369:1512-1521. doi:10.1056/NEJMoa1307337

Bettany-Saltikov

Weiss

Chockalingam

, et al. Surgical versus non-surgical interventions in people with adolescent idiopathic scoliosis. Cochrane Database Syst Rev. 2015;2015:CD010663. doi:10.1002/14651858.CD010663.pub2

Oetgen

Heyer

Kelly

. Scoliosis screening. J Am Acad Orthop Surg. 2021;29:370-379. doi:10.5435/JAAOS-D-20-00356

Wong

Hui

Rajan

Chia

. Idiopathic scoliosis in Singapore schoolchildren: a prevalence study 15 years into the screening program. Spine. 2005;30:1188-1196. doi:10.1097/01.brs.0000162280.95076.bb

Fong

Cheung

Wong

, et al. A population-based cohort study of 394,401 children followed for 10 years exhibits sustained effectiveness of scoliosis screening. Spine J. 2015;15:825-833. doi:10.1016/j.spinee.2015.01.019

Thomas

Stans

Milbrandt

Kremers

Shaughnessy

Larson

. Trends in incidence of adolescent idiopathic scoliosis: a modern US population-based study. J Pediatr Orthop. 2021;41:327-332. doi:10.1097/BPO.0000000000001808

Iwase

Taniyama

Koyanagi

, et al. A century of change: unraveling the impact of socioeconomic/historical milestones on age at menarche and other female reproductive factors in Japan. J Epidemiol. 2024;34:387-392. doi:10.2188/jea.JE20230155

Yoshimi

Matsumura

Takeda

. When and how do adolescent girls in Japan become aware of premenstrual symptoms from menarche? A cross-sectional study among senior high school students. BMJ Open. 2021;11:e045215. doi:10.1136/bmjopen-2020-045215

10.

Hinobayashi

. The evolutionary developmental psychology about the secular trend. Kakenhi-Project 2013;22330189.

11.

Mineko Tsukidate

Mashita

, et al. Observation of cutting point on Moire photography. Yobou Igaku Journal. 1998;299:26-31.

12.

Fujiwara

Harada

Hagiya

, et al. Trends in childhood obesity in Japan: a nationwide observational study from 2012 to 2021. Clin Obes. 2024;14:e12636. doi:10.1111/cob.12636

13.

Rigo

Grivas

. “Rehabilitation schools for scoliosis” thematic series: describing the methods and results. Scoliosis. 2010;5:27. doi:10.1186/1748-7161-5-27

14.

West

Bittel

Russell

Evans

Mama

Conroy

. A systematic review of physical activity, sedentary behavior, and substance use in adolescents and emerging adults. Transl Behav Med. 2020;10:1155-1167. doi:10.1093/tbm/ibaa008

15.

Hefti

. Pathogenesis and biomechanics of adolescent idiopathic scoliosis (AIS). J Child Orthop. 2013;7:17-24. doi:10.1007/s11832-012-0460-9

16.

Sanders

Browne

McConnell

Margraf

Cooney

Finegold

. Maturity assessment and curve progression in girls with idiopathic scoliosis. J Bone Joint Surg Am. 2007;89:64-73. doi:10.2106/JBJS.F.00067

17.

Willner

. Moire topography--a method for school screening of scoliosis. Arch Orthop Trauma Surg. 1979;95:181-185. doi:10.1007/BF00460117

18.

Beausejour

Roy-Beaudry

Goulet

Labelle

. Patient characteristics at the initial visit to a scoliosis clinic: a cross-sectional study in a community without school screening. Spine. 2007;32:1349-1354. doi:10.1097/BRS.0b013e318059b5f7

19.

Karachalios

Sofianos

Roidis

Sapkas

Korres

Nikolopoulos

. Ten-year follow-up evaluation of a school screening program for scoliosis. Is the forward-bending test an accurate diagnostic criterion for the screening of scoliosis? Spine. 1999;24:2318-2324. doi:10.1097/00007632-199911150-00006

20.

Fong

Lee

Cheung

, et al. A meta-analysis of the clinical effectiveness of school scoliosis screening. Spine. 2010;35:1061-1071. doi:10.1097/BRS.0b013e3181bcc835