Adult myopia progression in high myopes aged 18–25 years: evidence from a population-based cohort

Abstract

Aims

To investigate the prevalence and severity of adult myopia progression among highly myopic individuals aged 18–25 years, identify its associated factors and evaluate its association with pathologic myopia (PM).

Methods

We retrospectively included participants aged 18–25 years with consecutive follow-ups from the Shanghai Child and Adolescent Large-scale Eye Study database. Annual changes of cycloplegic spherical equivalent (ΔSE) and axial length (ΔAL) were calculated. Adult myopia progression is defined as ΔSE ≤−0.50 D/year or ΔAL≥0.10 mm/year. PM was diagnosed based on META-Analysis for Pathologic Myopia criteria (category≥2) and optical coherence tomography (OCT) images.

Results

Among all included 345 eyes, the mean ΔSE was −0.28 D/year (95% CI −0.31 to −0.20; p<0.001) and 64 (18.6%) were classified as having adult myopia progression. The mean ΔAL was 0.09 mm/year (95% CI 0.09 to 0.10; p<0.001), with axial elongation observed in 138 (40.0%) eyes. Both age (ΔSE: β=0.045, p=0.005; ΔAL: β=−0.015, p<0.001) and baseline AL (ΔSE: β=−0.044, p=0.028; ΔAL: β=0.020, p<0.001) were identified as factors associated with ΔSE and ΔAL. No differences in ΔSE (95% CI –0.08 to 0.11; p=0.759) or ΔAL (95% CI –0.04 to 0.01; p=0.190) were found between PM and non-PM groups, and neither metric improved PM discrimination.

Conclusions

Among high myopic individuals aged 18–25 years, 18.6% exhibited adult myopia progression and 40.0% showed axial elongation. Age and baseline AL were associated with adult myopia progression, but no clear link was found between adult myopia progression and the development of PM.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Myopic progression typically occurs during childhood and adolescence, but may persist into adulthood. Individuals with high myopia are more susceptible to adult myopia progression, which may further elevate their long-term risk of myopia-related complications.

WHAT THIS STUDY ADDS

• This retrospective analysis of a population-based cohort examined annual changes in cycloplegic spherical equivalent (SE) and axial length (AL) among 182 high myopes (345 eyes) aged 18–25 years. Adult myopia progression (ΔSE≤–0.50 D/year) was observed in 18.6% of eyes and axial elongation (ΔAL≥0.10 mm/year) in 40.1%. Age and baseline AL were identified as factors associated with myopia progression, with both the rates of myopic shift and axial elongation declining rapidly with age. No clear link was found between myopia progression in this age group and the development of pathologic myopia.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• A notable proportion of highly myopic adults aged 18–25 years continue to experience myopic shift and axial elongation. The persistence of myopia progression in this age group supports ongoing monitoring and management for individuals with documented changes. Given the variability in adult myopia progression, personalised strategies may help reduce the long-term risk of myopia-related complications. Further evidence is needed to improve understanding of adult myopia progression and guide more effective control approaches.

Introduction

Myopia is a common disorder affecting billions of people worldwide.^{1 2} Its prevalence has been steadily increasing in many countries, posing a significant public health concern.^{3 4} Beyond reducing uncorrected visual acuity, myopia can lead to vision-threatening complications that significantly reduce quality of life.⁵ At the societal level, myopia imposes a substantial economic burden through both direct healthcare expenditures and productivity losses.6,8

While myopia progression is typically most rapid during childhood and adolescence, a notable proportion of individuals either develop new-onset myopia or continue to experience myopia progression during young adulthood.9,11 The International Myopia Institute (IMI) has concluded that clinically meaningful myopia progression continues in early adulthood and may average 1.00 D between 20 and 30 years.¹²

High myopia represents an advanced stage of the condition, strongly associated with pathologic myopia (PM) and poor visual prognosis.¹³ At this stage, every increase in refraction can significantly raise the risk of sight-threatening complications.¹⁴ Meanwhile, evidence suggests that individuals with higher baseline myopia are more likely to experience adult myopia progression, further increasing their risk of irreversible vision loss.¹⁵ Understanding adult myopia progression among high myopes is essential for generating evidence to optimise intervention strategies and improve long-term clinical management.

Based on a retrospective analysis of an ongoing population-based cohort study, we aimed to investigate the prevalence and severity of adult myopia progression among highly myopic individuals aged 18–25 years. We also examined the associated factors of dioptre and axial length (AL) changes and explored whether myopia progression during this age period is associated with the development of PM.

Methods

Study participants

The data for this study were obtained from the Shanghai Child and Adolescent Large-scale Eye Study (SCALE-HM). SCALE-HM is a population-based, prospective cohort study (ClinicalTrials.gov identifier: NCT03666052) initiated in 2018 and expected to continue until 2038, with annual follow-ups. All participant information and examination data in the cohort were recorded using an online data acquisition and management system developed by a third party (Gaussinfomad, Beijing, China). The accuracy and completeness of the uploaded data were automatically verified, with modifications requiring supervisor authorisation. Study data were uploaded to the system, where accuracy and completeness were automatically verified. Any data modifications were recorded, requiring supervisor authorisation. The inclusion and exclusion criteria for SCALE-HM have been described in published literature.¹⁶

As of 6 November 2024, records in the SCALE-HM database were manually reviewed. Participants aged 18–25 years at the first visit and with at least one consecutive follow-up visit were included, yielding a total of 220 individuals. Data from both eyes were analysed. The exclusion criteria were as follows: eyes with cycloplegic spherical equivalent (SE)>−6.00 D; best-corrected visual acuity worse than 0.1 logMAR; absence of cycloplegia during visits; history of strabismus, amblyopia, glaucoma or other ocular diseases; and use of any myopia control interventions within the past year. All included examination records were manually reviewed by researchers to ensure data integrity and validity.

Ophthalmic examinations

The data for all participants in this study were obtained using the examination methods employed in SCALE-HM, which was adequately described in published literature.¹⁶ Specifically, axial length (AL) was measured using IOL Master 700 (Carl Zeiss Meditec, Germany). Refraction and corneal curvature radius (CR) were obtained with an autorefractor (KR-8900, Topcon, Japan). Cycloplegia was induced using 1% cyclopentolate (Cyclogyl, Alcon). OCT examination was conducted using a swept-source optical coherence tomography (Triton, Topcon, Japan), and fundus images were captured using a fundus camera (TRC-50DX, Topcon, Japan). All measurements were performed using the same equipment and protocols throughout follow-up visits. Family history of high myopia was collected via questionnaires.

Fundus examination and pathologic myopia classification

Fundus images obtained during the second visit were independently conducted by two researchers following the META-Analysis for Pathologic Myopia classification system, with categories≥2 considered indicative of PM.¹⁷ OCT images were also reviewed to identify any signs of posterior staphyloma and myopic maculopathy; if present, these findings were also classified as PM.¹⁸ Choroidal thickness (ChT) was defined as the distance between Bruch’s membrane and the choroid–sclera interface, and was measured within a 1 mm diameter circular region centred on the fovea.

Annual change of axial length and cycloplegic spherical equivalent

Since SCALE-HM follows an annual follow-up schedule, the interval between consecutive visits was approximately 1 year, ranging from 10 to 16 months in this study. The change in AL and cycloplegic SE was calculated by subtracting the first visit’s values from the second and was then adjusted based on the exact time interval (in years). The results were recorded as annual change of AL (ΔAL, mm/year) and annual change of SE (ΔSE, D/year). A simple flowchart of myopia progression calculation is provided as online supplemental figure S1.

Statistical analysis

Adult myopia progression is defined as ΔSE≤−0.50 D/year, and axial elongation is defined as ΔAL≥0.10 mm/year, following clinical practice.

Quantitative data are presented as mean±SD and categorical data as percentages. Categorical and continuous variables were compared using the χ² test or t-test. Pearson correlation was used to analyse the relationship between ΔAL and ΔSE. Linear mixed-effects models (LMM) were employed to identify the factors associated with ΔAL and ΔSE. Receiver operating characteristic (ROC) curves were applied to assess the differentiating performance for PM. Statistical analyses were conducted using R software, V.4.4.3. Statistical significance was set at a two-tailed p value<0.05. Data analysis was performed from November 2024 to August 2025.

Results

After screening the records, a total of 345 eyes (182 participants) were included in the study. The mean age was 19.44±1.21 years at the first visit and 20.50±1.22 years at the second visit. The cycloplegic SE was −9.37 ± 2.06 D at the first visit and −9.66 ± 2.15 D at the second. AL increased from 27.27 ± 1.16 mm to 27.37 ± 1.19 mm. No differences in SE or AL were observed between the right and left eyes at any visit. Other characteristics are summarised according to the presence or absence of PM in table 1.

Table 1. Participants’ demographic characteristics and ocular measurements stratified by PM status.

	Overall	Non-PM	PM	95% CI*	P value*
Eyes (n)	345	286	59	–	–
Right eye, n (%)	177 (51.3)	149 (52.1)	28 (47.5)	–	0.613
Sex, male, n (%)	175 (50.7)	147 (51.4)	28 (47.5)	–	0.683
Age (years)	19.44±1.21	19.35±1.15	19.88±1.36	−0.91 to −0.15	0.007
SE (D)	−9.37±2.06	−9.14±1.86	−10.49±2.58	−9.14 to −10.49	<0.001
AL (mm)	27.27±1.16	27.12±1.07	27.99±1.32	−1.23 to −0.50	<0.001
CR (mm)	7.87±0.29	7.85±0.29	7.99±0.27	−0.22 to −0.06	0.001
IOP (mm Hg)	15.40±3.06	15.39±3.11	15.42±2.84	−0.85 to 0.79	0.939
RT (μm)	270.90±11.08	271.85±10.61	266.33±12.20	2.12 to 8.91	<0.001
ChT (μm)	171.93±45.57	179.93±43.55	133.16±33.91	36.63 to 56.90	<0.001
ΔSE (D/year)	−0.28±0.28	−0.28±0.28	−0.30±0.30	−0.08 to 0.11	0.759
ΔAL (mm/year)	0.09±0.08	0.09±0.08	0.10±0.07	−0.04 to 0.01	0.190

^*P values and 95% CI were calculated on the basis of t-tests for continuous variables and χ2 test for categorical variables.

AL, axial length; ChT, choroidal thickness, measured within a 1-mm-diameter circular region centered on the fovea; CR, corneal curvature radius; IOP, intraocular pressure; PM, pathologic myopia; RT, retinal thickness, measured within a 1-mm-diameter circular region centered on the fovea; SE, spherical equivalent; ΔAL, annual change of axial length; ΔSE, annual change of cycloplegic spherical equivalent.

Individual SE change and prevalence of adult myopia progression

The mean ΔSE across all included eyes was −0.28±0.28 D/year (95% CI −0.31 to −0.25 D/year; p<0.001). 178 eyes (51.6%) exhibited –0.25≤ΔSE<0.00 D/year, 99 eyes (28.7%) exhibited −0.50≤ΔSE<–0.25 D /year, 51 eyes (14.8%) exhibited −1.00≤ΔSE<–0.50 D /year and 13 eyes (3.8%) exceeding –1.00 D/year. Four eyes exhibited ΔSE>0 D/year. Figure 1A illustrates the density distribution of ΔSE in the overall sample. Using a threshold of –0.50 D/year, 64 eyes (18.6%) were identified as having adult myopia progression.

Figure 1. Distribution of ΔSE and ΔAL. Solid lines represent density estimates of the data. ΔAL, annual change of axial length; ΔSE, annual change of cycloplegic spherical equivalent.

Individual AL change and prevalence of axial elongation

The mean change in CR between the two visits across all included eyes was –0.00±0.04 mm (95% CI –0.00 to 0.01 mm; p=0.703). The mean ΔAL across all included eyes was 0.09±0.08 mm/year (95% CI 0.09 to 0.10 mm/year; p<0.001). 112 eyes (32.5%) exhibited 0≤ΔAL<0.05 mm/year, 90 eyes (26.1%) exhibited 0.05 ≤ΔAL<0.10 mm /year, 101 eyes (29.3%) exhibited 0.10≤ΔAL<0.20 mm/year and 37 eyes (10.7%) exceeding 0.20 mm/year. Five eyes showed ΔAL<0 mm/year. Figure 1B illustrates the density distribution of ΔAL in the overall sample. Using a threshold of 0.10 mm/year, 138 eyes (40.0%) were identified as having axial elongation.

ΔSE and ΔAL showed a negative correlation among all participants, with a Pearson’s correlation coefficient of –0.67 (95% CI –0.72 to –0.60; p<0.001). Online supplemental figure S2 presents the scatter plot of ΔSE vs ΔAL, along with the fitted correlation line.

Associated factors of ΔSE and ΔAL

Univariable LMMs were first applied, identifying age, sex, cycloplegic SE, AL and ChT as potential associated factors of ΔSE or ΔAL. These variables were subsequently included in the multivariable LMMs, where age and baseline AL were identified as associated factors of both ΔSE (age: β=0.045, p=0.005; AL: β=−0.044, p=0.028) and ΔAL (age: β=−0.015, p<0.001; AL: β=0.020, p<0.001). The results of the multivariable models for ΔSE and ΔAL are presented in table 2. The marginal coefficient of determination (R²), conditional R² and adjusted intraclass correlation coefficient for the ΔSE model were 0.090, 0.681 and 0.649, respectively, while those for the ΔAL model were 0.206, 0.707 and 0.623.

Table 2. Univariable and multivariable regression analysis of factors associated with ΔSE and ΔAL.

	Univariable LMM			Multivariable LMM
Variables	Coefficient*	Standard error*	P value	Coefficient*	Standard error*	P value
LMM for ΔSE
Age (years)	4.44	1.59	0.006	4.50	1.60	0.005
Sex, male (n)	−7.93	3.84	0.041	−4.22	4.13	0.307
IOP (mm Hg)	−1.02	0.55	0.064	–	–	–
Cycloplegic SE (D)	2.57	0.76	<0.001	1.67	1.03	0.105
AL (mm)	−5.38	1.48	<0.001	−4.35	2.07	0.028
ChT (μm)	0.06	0.03	0.046	0.04	0.03	0.208
Parental high myopia (n)	−7.36	3.99	0.067	–	–	–
LMM for ΔAL
Age (years)	−1.42	0.44	0.001	−1.50	0.42	<0.001
Sex, male (n)	2.54	1.10	0.019	0.24	1.09	0.829
IOP (mm Hg)	0.21	0.20	0.107	–	–	–
Cycloplegic SE (D)	−1.20	0.20	<0.001	−0.46	0.27	0.092
AL (mm)	2.93	0.40	<0.001	1.99	0.57	<0.001
ChT (μm)	−0.03	0.01	0.002	−0.01	0.01	0.211
Parental high myopia (n)	1.74	1.10	0.115	–	–	–

^*Coefficient and standard error were all multiplied by 102.

AL, axial length; ChT, choroidal thickness; IOP, intraocular pressure; LMM, linear mixed-effects models; SE, spherical equivalent; ΔAL, annual change of axial length; ΔSE, annual change of cycloplegic spherical equivalent.

Age-based subgroup analysis was conducted to further examine ΔSE and ΔAL across different age groups. Figure 2 presents scatter plots of age and ΔSE or ΔAL, along with the proportions of adult myopia progression and axial elongation in each age group. Overall, the proportion of participants exhibiting adult myopia progression and axial elongation declined rapidly with increasing age. The rate of adult myopia progression decreased from 26.6% at age 18 to 3.8% at age 21, with no participants aged 22–25 identified as having progression. Similarly, the proportion of axial elongation dropped from 44.3% at age 18 to 19.2% at age 21, and no cases were identified in the 22–25 age group.

Figure 2. Scatter and bar plots illustrating age-related changes in ΔSE and ΔAL. (A) Scatter plot of age vs ΔSE. Eyes with adult myopia progression are shown in blue, and eyes without progression are shown in grey. (B) Scatter plot of age vs ΔAL. Eyes with axial elongation are shown in red, and eyes without axial elongation are shown in grey. (C) Bar chart showing the proportion of adult myopia progression across age groups. Among eyes aged 18 years (n=158), 43 eyes (26.6%) were identified as having adult myopia progression and 115 eyes (73.4%) as having no progression. At age 19 (n=96), 15 eyes (15.6%) were identified with progression and 81 eyes (84.4%) without. At age 20 (n=53), 6 eyes (11.3%) were identified with progression and 47 eyes (88.7%) without. At age 21 (n=26), 1 eye (3.8%) was identified with progression and 25 eyes (96.2%) without. In the 22–25 age group (n=12), no eyes were identified as having adult myopia progression. (D) Bar chart showing the proportion of axial elongation across age groups. Among eyes aged 18 years (n=158), 70 eyes (44.3%) were identified as having axial elongation and 88 eyes (55.7%) as having no axial elongation. At age 19 (n=96), 42 eyes (43.8%) were identified with axial elongation and 54 eyes (56.2%) without. At age 20 (n=53), 21 eyes (39.6%) were identified with axial elongation and 32 eyes (60.4%) without. At age 21 (n=26), 5 eyes (19.2%) were identified with axial elongation and 21 eyes (80.8%) without. In the 22–25 age group (n=12), no eyes were identified as having axial elongation. ΔAL, annual change of axial length; ΔSE, annual change of cycloplegic spherical equivalent.

Diagnostic performance of adult myopia progression for pathologic myopia

Among all 345 highly myopic eyes, 59 eyes (17.1%) were diagnosed with PM. No differences were found in mean ΔSE (95% CI –0.08 to 0.11; p=0.759) or mean ΔAL (95% CI –0.04 to 0.01; p=0.190) between the PM and non-PM groups. After adjusting for age and baseline AL, the differences in mean ΔSE and ΔAL remained non-significant.

We further evaluated the differentiating performance of ΔSE and ΔAL for identifying PM. When used individually, the area under the curves (AUCs) for ΔSE and ΔAL were 0.53 (95% CI 0.45 to 0.61) and 0.56 (95% CI 0.48 to 0.64), respectively. ChT demonstrated the highest AUC, significantly outperforming both ΔSE (95% CI –0.38 to –0.19; p<0.001) and ΔAL (95% CI –0.48 to –0.27; p<0.001) using DeLong test. For multivariable prediction strategies, a multivariable model incorporating age, AL, ChT and retinal thickness achieved the highest AUC of 0.85 (95% CI 0.79 to 0.91). Adding ΔSE or ΔAL to this model did not improve performance (ΔSE: 95% CI –0.01 to 0.01; p=0.79; ΔAL: 95% CI –0.01 to 0.01; p=0.87). Figure 3 illustrates the ROC curves for both univariable and multivariable models.

Figure 3. Receiver operating characteristic curves for univariate and multivariate models in identifying pathologic myopia. Combined method: a linear mixed-effects model incorporating age, AL, ChT and RT. AL, axial length; AUC, area under the curve; ChT, choroidal thickness; RT, retinal thickness; ΔAL, annual change of axial length; ΔSE, annual change of cycloplegic spherical equivalent.

Discussion

Based on a retrospective analysis of a population-based cohort, we examined annual changes in cycloplegic SE and AL among high myopes aged 18–25 years, as well as the prevalence of adult myopia progression and axial elongation. Overall, myopic shift remained common among high myopes in this age group, with 18.6% identified as having adult myopia progression (ΔSE≤−0.50 D/year). While CR showed no change, ΔAL change was found to be correlated with ΔSE, with 40.0% of individuals exhibiting axial elongation (ΔAL≤0.10 mm/year). Both the rates of myopic shift and axial elongation declined rapidly with age. No clear association was found between myopia progression during this age range and the development of PM. These findings highlight the need for continued monitoring and management of myopia progression in high myopes beyond adolescence.

Growing evidence has positioned adult myopia progression as a hot topic in myopia research as it may further exacerbate the already substantial burden of myopia, especially in Asian countries.^{9 19 20} After reviewing existing evidence, IMI concluded that myopia progression in adults is common and requires ongoing management. IMI also emphasised the need for more high-quality research to better characterise the extent of myopic shift and axial elongation during adulthood.¹²

Many previous studies have identified baseline myopia severity as a promotive factor associated with adult myopia progression.¹⁵ In clinical practice, patients with high myopia are more likely to experience continued dioptre changes and axial elongation during adulthood, further increasing the risk of vision-threatening complications. A Japanese study involving 1877 highly myopic individuals with a mean age of 62.1 years reported an average ΔAL of 0.05 mm/year.²¹ Another study of 563 highly myopic but non-pathologic participants with a mean age of 31.4 years reported an average ΔAL of 0.03 mm/year.²² More recently, a study including highly myopic adults aged 18–40 years found a higher rate of 0.07 mm/year.²³

Our research specifically focused on adult myopia progression in high myopes aged 18–25 years. At this college-age stage, individuals are generally younger and often exposed to intensive near work and prolonged screen time. A faster rate of myopia progression in this subgroup is therefore expected. Our findings confirmed a higher rate of myopic shift and axial elongation than previously reported, with a clear age-related decline.

The continued presence of myopic shift supports ongoing myopia control interventions in high myopes with documented progression.¹² In our data, 18.8% of participants were identified with adult myopia progression and 40.1% with axial elongation, indicating a substantial proportion who may potentially benefit from continued intervention. Our results also offer clearer guidance on when myopia control interventions may be relaxed as progression rates markedly stabilise with increasing age within the 18–25 year range. Additionally, these findings underscore the importance of thoroughly assessing refractive stability before considering refractive surgery in high myopic individuals in this age group.²⁴

An interesting topic for discussion is whether adult myopia progression represents a delayed cessation of axial elongation, predominantly occurring in the retroequatorial region, which is recognised as the main site of structural changes during refractive development in childhood and adolescence.²⁵ In contrast, another perspective suggests that continued axial elongation in adulthood reflects a distinct process mainly driven by morphological changes at the posterior pole.²⁶ This view proposes that adult myopia progression may be caused by the development of posterior staphyloma or choroidal thinning, which may exert greater impact on the normal structure and function of the retina.¹⁸ The Chinese Expert Consensus on Prevention and Control of High Myopia also considers continued myopic progression beyond the age of 18 as one of the characteristic manifestations for pathologic high myopia.²⁷

Our study did not find further evidence linking adult myopia progression in individuals aged 18–25 years with the development of PM. The PM group did not show a faster rate of myopic shift or axial elongation compared with the non-PM group, and neither ΔSE nor ΔAL provided additional value in distinguishing PM cases. However, this conclusion is limited by the relatively short follow-up period. It is also noteworthy that the prevalence of PM in our analysis was higher than in previous observational studies,²⁸ which may be attributable to the higher baseline SE in our participants. This finding may also reflect a degree of selection bias as individuals with more severe ocular conditions were more likely to return for follow-up. It may be more valuable to investigate the relationship between fundus changes and dioptre progression, which was not addressed in our current analysis. Future studies could consider using tessellated fundus density or other quantitative retinal imaging markers to explore this further.^{29 30}

Both younger age and higher baseline myopia were identified as influencing factors of adult myopia progression and axial elongation, which aligned with previous findings.¹⁵ Although prolonged myopia progression has been suggested to be related to genetic predisposition, our study did not find a clear association between family history of high myopia and either ΔSE or ΔAL. This could be attributed to the limited validity of self-reported questionnaire data. Our LMMs suggest that fixed effects accounted for only a limited proportion of the variance in ΔSE and ΔAL, while random effects were more substantial. This highlights considerable individual variability in adult myopia progression that remains unexplained. Incorporating additional variables such as demographic characteristics, including socioeconomic status and education level, as well as behavioural factors such as outdoor activity and screen time, may improve predictive performance. The use of more advanced modelling approaches or the inclusion of structural biomarkers, such as choroidal blood flow data, may also be helpful. Meanwhile, the observed variability underscores the importance of personalised approaches to myopia management.

Additional limitations include the following: (1) the 1-year observation period may have resulted in relatively small changes in refraction and AL, potentially amplifying the impact of measurement error. (2) The short follow-up duration limited our ability to evaluate the relationship between adult myopia progression and long-term development of PM. (3) Environmental factors were not included in our analysis. Additionally, the absence of ultra-widefield SS-OCT or B-scan ultrasonography may have led to missed cases of extensive posterior staphyloma. Further studies with longer follow-up and more comprehensive imaging and behavioural assessments are needed to improve our understanding of adult myopia progression and guide more effective control strategies during this age range.

Conclusion

Among high myopic individuals aged 18–25 years, 18.6% exhibited adult myopia progression and 40.0% showed axial elongation. Age and baseline AL were identified as factors associated with progression, but no clear link was found between adult myopia progression and the development of PM. Further research is needed to better understand adult myopia progression among high myopes.

Data availability statement

Data are available upon reasonable request.