Incidence and Progression of Myopia in Early Adulthood

This cohort study investigates the 8-year incidence of myopia and change in ocular biometry in young adults and their association with the known risk factors for childhood myopia.

Key Points

Question

How common is myopia progression and onset during early adulthood?

Findings

In a cohort study of 691 young adults from a general population, significant increases were observed over the 8-year study period in myopia and axial length by 0.04 diopters per year and 0.02 mm per year, respectively. Of the 526 participants without myopia at baseline, myopia incidence from age 20 to 28 years was 14%.

Meaning

In this study, there was a high incidence of myopia and prevalence of myopia progression in the third decade of life.

Abstract

Importance

Myopia incidence and progression has been described extensively in children. However, few data exist regarding myopia incidence and progression in early adulthood.

Objective

To describe the 8-year incidence of myopia and change in ocular biometry in young adults and their association with the known risk factors for childhood myopia.

Design, Setting, and Participants

The Raine Study is a prospective single-center cohort study. Baseline and follow-up eye assessments were conducted from January 2010 to August 2012 and from March 2018 to March 2020. The data were analyzed from June to July 2021. A total of 1328 participants attended the baseline assessment, and 813 participants attended the follow-up assessment. Refractive information from both visits was available for 701 participants. Participants with keratoconus, previous corneal surgery, or recent orthokeratology wear were excluded.

Exposures

Participants’ eyes were examined at ages 20 years (baseline) and 28 years.

Main Outcomes and Measures

Incidence of myopia and high myopia; change in spherical equivalent (SE) and axial length (AL).

Results

A total of 516 (261 male [50.6%]) and 698 (349 male [50.0%]) participants without myopia or high myopia at baseline, respectively, were included in the incidences analyses, while 691 participants (339 male [49%]) were included in the progression analysis. The 8-year myopia and high myopia incidence were 14.0% (95% CI, 11.5%-17.4%) and 0.7% (95% CI, 0.3%-1.2%), respectively. A myopic shift (of 0.50 diopters [D] or greater in at least 1 eye) occurred in 261 participants (37.8%). Statistical significance was found in longitudinal changes in SE (−0.04 D per year; P < .001), AL (0.02 mm per year; P <.001), and lens thickness (0.02 mm per year; P < .001). Incident myopia was associated with self-reported East Asian vs White race (odds ratio [OR], 6.13; 95% CI, 1.06-35.25; P = .04), female vs male sex (OR, 1.81; 95% CI, 1.02-3.22; P = .04), smaller conjunctival ultraviolet autofluorescence area (per 10-mm² decrease, indicating less sun exposure; OR, 9.86; 95% CI, 9.76-9.97; P = <.009), and parental myopia (per parent; OR, 1.57; 95% CI, 1.03-2.38; P = <.05). Rates of myopia progression and axial elongation were faster in female participants (estimate: SE, 0.02 D per year; 95 % CI, 0.01-0.02 and AL, 0.007 mm per year, 95 % CI, 0.00.-0.011; P ≤ .001) and those with parental myopia (estimate per parent: SE, 0.01 D per year; 95% CI, 0.00-0.02 and AL, 95% CI, 0.002-0.008; P ≤ .001). Education level was not associated with myopia incidence or progression.

Conclusions and Relevance

These findings suggest myopia progression continues for more than one-third of adults during the third decade of life, albeit at lower rates than during childhood. The protective effects of time outdoors against myopia may continue into young adulthood.

Introduction

The global myopia epidemic is well reported^1,2 and the rate of myopia-associated complications is expected to similarly rise as younger generations with high myopia prevalence approach middle age and older age.^3,4 Myopia typically develops and progresses fastest during childhood, and it has been reported that myopia stabilizes (defined as change of less than 0.5 diopters [D]) at around age 15 to 16 years.^5,6 However, longitudinal studies involving university students have demonstrated that myopia may progress and even start to develop during young adulthood. In 118 university students who were observed for 3 years in Portugal⁷ (mean age, 21 years at baseline), prevalence of myopia and hyperopia increased by 5% and decreased by 9%, respectively, while mean spherical equivalent decreased by 0.3 D. Another study in Norway⁸ found a 3-year myopia incidence of 33% among university students (mean age, 21 years at baseline), with mean spherical equivalent decreasing by 0.6 D. Similar longitudinal findings were reported in university students in Denmark⁹ and the US.^10,11

With the modern emphasis on education, a known risk factor for myopia,¹² myopia may continue to progress or onset during young adulthood. With the rise in indoor jobs in the past century¹³ and the increase in automation of many manual or outdoor labor occupations,¹⁴ individuals are likely to spend less time outdoors, which could further drive myopia progression during young adulthood,^15,16,17 even after formal education is completed. However, there are limited data in the literature on myopia development or progression in young adulthood, and often studies have been conducted in select populations.

This study aimed to (1) describe the 8-year incidence of myopia and high myopia and (2) examine the 8-year within-person change in refractive measures in young adults from a general population. Within both aims, we explored risk factors for myopia development or progression during young adulthood and tested the hypothesis that the 3 known major risk factors of childhood myopia—higher level of education, lower time spent outdoors, and parental myopia—are also associated with myopia development and progression during young adulthood.

Methods

Study Sample

The Raine Study¹⁸ has observed a cohort of participants since their prenatal periods in 1989 to 1991, when more than 2900 pregnant women were recruited from the King Edward Memorial Hospital and surrounding obstetric clinics in Perth, Australia. From these women, 2868 offspring were born, forming the original study cohort who are now in young adulthood, and have since been undergoing a series of regular medical and health examinations.

At the 20-year follow-up (age 18 to 22 years), participants underwent their first Raine Study eye examination.¹⁹ Participants were invited to return for an eye examination at the 28-year follow-up. All follow-up assessments of the Raine Study were conducted in compliance with the Declaration of Helsinki and have been approved by the University of Western Australia’s Human Research Ethics Committee. All participants were given a full explanation of the nature of the study and provided written informed consent prior to participating in each follow-up assessment. No incentives were provided for participation, apart from reimbursement of parking fees for the visit.

Eye Examination

The 20-year and 28-year follow-up assessments were conducted from January 2010 to August 2012 and from March 2018 to March 2020, respectively.^19,20 In brief, both eye examinations included conjunctival ultraviolet autofluorescence (CUVAF) photography, ocular biometry (IOLMaster V.5; Carl Zeiss Meditec AG), postmydriatic autorefraction/keratometry (Nidek ARK-510A; NIDEK), and lens thickness measurement (Oculus Pentacam; OculusOptikgerate GmbH), among others. Autorefraction was performed at least 20 minutes after instillation of 1 drop of tropicamide, 1%. CUVAF photography is an objective method of measuring ocular sun exposure and has a strong correlation with self-reported time spent outdoors in adults.²¹ The same refraction and ocular biometric measurement protocol and instrument models were used in both follow-up assessments.

A participant was considered to have myopia or high myopia if either or both eyes had a spherical equivalent of 0.50 D or less or 6.00 D or less, respectively.²² A refraction shift was defined as a change of 0.50 D or more in spherical equivalent in either direction (myopic/hyperopic).

Questionnaire

In a self-administered questionnaire, participants indicated their highest level of education as (1) up to secondary school; (2) vocational qualification (including technical college, vocational training, or other certification courses); (3) undergraduate degree; or (4) postgraduate degree, and years of education. Self-reported parental myopia, race, and ocular history were also obtained. Race was categorized as East Asian, White, and others/mixed (Aboriginal Australian, South Asian, Southeast Asian, Torres Straits Islander, or a combination of these with East Asian or White). East Asian participants were analyzed in their category in view of the observed high prevalence of myopia in this demographic.²³ Ocular history information included previous operations and keratoconus. For participants who had laser refractive surgery, we further asked if they remembered their approximate refraction prior to surgery (eg, what their contact lens prescription was prior to surgery).

Statistical Analysis

For aim 1 (describing the 8-year incidence of myopia), participants were included in the analysis if they had postmydriatic refraction data at both assessments and if they had no myopia at the 20-year follow-up (baseline). A similar process was applied to obtain the 8-year incidence of high myopia. Logistics regression was used to explore the risk factors of myopia development, including sex, race, education, and ocular sun exposure.

To address aim 2, all participants who had refraction data at both follow-up assessments were included, regardless of myopia status. Linear mixed-effect models were used with random intercept and slope for participants to account for the within-participant correlation between 2 eyes.²⁴ In multivariable analyses, the main effects of sex, race, highest level of education, CUVAF area (as an objective measure of ocular sun exposure), and parental myopia as well as interaction effects with age on refractive measures were evaluated.

Participants who wore orthokeratology lenses or had a history of cataract or corneal surgery were removed from the analyses. Participants with keratoconus, defined as having a Belin/Ambrósio enhanced ectasia display score of 2.6 or more in either eye based on Scheimpflug imaging²⁵ at the 28-year follow-up assessment, were additionally excluded. To maximize the sample size, we included participants who underwent laser refractive surgery between ages 20 and 28 years by adding their self-provided estimated presurgical spherical equivalent (if known) to their 8-year refraction data as obtained during the eye examination. However, these participants were removed from analyses with keratometry as the outcome measure. Participants who were not able to provide or recall their estimated refraction data prior to surgery were excluded.

All analyses were conducted using R version 3.6.2 (The R Foundation for Statistical Computing Platform), and the level of significance was set at P < .05. Because of the multiple comparisons in aim 2, the level of significance was set as P < .025 for aim 2 with the Bonferroni correction, in consideration of the 2 main refractive outcome measures (spherical equivalent and axial length; this adjustment was not done for aim 1 as it only had 1 outcome measure).

Results

Of 1344 participants who attended the baseline visit, 1328 had refractive data, including 342 participants and 19 participants with myopia and high myopia, respectively, giving a prevalence of 25.8% (95% CI, 23.5%-28.2%) and 1.4% (95% CI, 0.9%-2.2%). Of the 801 participants who attended the follow-up, 783 had refractive data, included 260 participants with myopia (prevalence, 33.2%; 95% CI, 30.1%-36.7%) and 12 participants with high myopia (1.5%; 95% CI, 0.9%-2.7%).

Among the 1344 participants who attended the baseline visit, there was no significant difference in race, baseline spherical equivalent, axial length, or CUVAF area between those who returned and did not return for the follow-up visit. However, there were more male participants than female participants who did not attend the follow-up visit (male: did not attend, 360 [53.9%] vs attended, 261 [48.9%]; P = .05).

8-Year Incidence of Myopia

After excluding participants who had no refraction data at baseline or follow-up, those with myopia at baseline, keratoconus, or recent use of orthokeratology contact lenses, a total of 516 participants (50.6% male) were included in the myopia incidence analysis (Figure). The cumulative 8-year myopia incidence was 14.0% (95% CI, 11.5%-17.4%), with 72 participants developing myopia. In univariable logistic regression, myopia incidence was significantly associated with female sex, East Asian race (relative to White participants), less sun exposure (as indicated by smaller CUVAF areas), and parental myopia (Table 1). Participants who reported vocational training as their highest level of education had lower odds of incident myopia relative to those who reported up to secondary school. In the multivariable analyses, all these factors, except for education, remained significantly associated with incident myopia (eAppendix in the Supplement).

Figure. Sample Size for Incidence Analysis (Aim 1).

^aFewer participants attended the 28-year follow-up partly because data collection had to cease early because of the COVID-19 pandemic.

^bIncludes 5 with history of laser refractive surgery but known refractive error prior to surgery.

Table 1. Risk Factors Associated With Incident Myopia Between Age 20 and 28 Years.

Characteristic	No. (%)			Analyses
	All participants	Developed myopia
				Univariable		Multivariablea
		Yes	No	OR (95% CI)	P value	OR (95% CI)	P value
No.	516	72	444	NA	NA	NA	NA
Sex
Male	261 (50.6)	26 (36.1)	235 (90.0)	1 [Reference]	NA	NA	NA
Female	255 (49.4)	46 (63.9)	209 (82.0)	1.99 (1.19-3.33)	.009	1.81 (1.02-3.22)	.04
Raceb
East Asian	6 (1.2)	3 (4.2)	3 (6.8)	6.63 (1.31-33.63)	.02	6.13 (1.06-35.25)	.04
White	458 (88.8)	60 (83.3)	398 (89.6)	1 [Reference]	NA	NA	NA
Other/mixedc	52 (10.1)	9 (12.5)	43 (9.7)	1.39 (0.64-2.99)	.40	1.45 (0.65-3.26)	.55
Parental myopiab,d
None	373 (72.3)	40 (55.6)	333 (75.0)	1 [Reference]	NA	NA	NA
1 Parent	102 (19.8)	24 (33.3)	78 (17.6)	NA	.001	1.57 (1.03-2.38)c	.05
Both parents	30 (5.8)	7 (9.7)	23 (5.2)	1.86 (1.28-2.70)c	NA		NA
No response	9 (1.7)	2 (1.4)	10 (2.3)	NA	NA	NA	NA
Highest educationb
Up to secondary school	96 (18.6)	17 (23.6)	79 (17.9)	1 [Reference]	NA	NA	NA
Vocational training	158 (30.6)	14 (19.4)	144 (32.4)	0.45 (0.21-0.96)	.04	0.57 (0.13-1.02)	.22
Undergraduate degree	168 (32.3)	25 (34.7)	143 (32.2)	0.81 (0.41-1.60)	.55	0.64 (0.31-1.32)	.44
Postgraduate degree	77 (14.9)	15 (20.8)	62 (14.0)	1.12 (0.52-2.43)	.77	1.00 (0.44-2.25)	.80
No response	18 (3.3)	1 (1.4)	16 (3.6)	NA	NA	NA	NA
CUVAF area per 10-mm2 increase, median (IQR)	40.7 (21.6- 64.5)	34 (14.8-52.7)	42.8 (23.3- 65.4)	9.87 (9.78-9.97)	.009	9.86 (9.76-9.97)	.009

Abbreviations: CUVAF, conjunctival UV autofluorescence; NA, not applicable; OR, odds ratio.

^a Includes all variables in the table.

^b Self-reported.

^c Other includes Aboriginal Australian, South Asian, Southeast Asian, Torres Strait Islander, or a combination of these with East Asian or White.

^d OR with each additional parent with myopia.

Eyes with incident myopia had lower spherical equivalent, longer axial lengths, and thinner lens at baseline (age 20 years) than those that did not become myopic (Table 2). There was no significant difference in baseline corneal radius between groups (Table 2).

Table 2. Baseline (Age 20 Years) Ocular Measures According to Incident Myopia.

Measure	Developed myopia, median (IQR)		Group difference statistical outcomea
	Yes	No	F statistic	P value
Eyes, No.	122	925	NA	NA
Spherical equivalent, D	0 (−0.13 to 0.25)	0.50 (0.25-0.75)	F1,514 = 50.2	<.001
Axial length, mm	23.52 (23.10-23.88)	23.30 (22.81-23.79)	F1,512 = 7.2	.007
Central corneal radius, mm	7.72 (7.55-7.85)	7.74 (7.57-7.92)	F1,512 = 2.0	.16
Lens thickness, mm	3.46 (3.33-3.58)	3.51 (3.38-3.64)	F1,474 = 5.0	.03

Abbreviations: D, diopter; NA, not applicable.

^a Group difference analyzed using linear mixed-effect models to account for the within-person correlation between 2 eyes and adjustment for sex and race.

8-Year Incidence of High Myopia

There were 683 participants (338 male [49.5%]) available for the high myopia incidence analysis (Figure). This included 5 participants with prior laser refractive surgery who were able to provide their estimated refraction prior to surgery, either directly obtained from their optometrist (n = 1) or the participants recalled their presurgery contact lens prescription (n = 4).

The incidence of high myopia was 0.7% (95% CI, 0.3%-1.2%), with 5 participants progressing to high myopia. None of these participants had a history of laser refractive surgery. eTable 1 in the Supplement presents the refractive error, parental myopia, highest level of education, and CUVAF area for these 5 participants. Most of these 5 participants had myopia of −5.00 D or worse in at least 1 eye at the 20-year follow-up visit and progressed by less than 2.00 D in that 8-year period (progression rate of −0.08 to −0.22 per year).

8-Year Change in Refractive Measures

Of 701 participants who had refractive data at both follow-up visits, 6 participants who had keratoconus, 3 participants who had prior refractive surgery with unknown prescription prior to surgery, and 1 participant who wore orthokerothalogy lenses a few days before the eye examination were excluded from the analysis. This left 691 participants available for this analysis. The 5 participants who had prior refractive surgery but had presurgery refraction information were excluded only from the keratometry analyses.

There were 261 participants (37.8%) who experienced a myopic shift (0.50 D or greater) in at least 1 eye over the 8 years, including 152 participants with a myopic shift in both eyes (Table 3). The spherical equivalent in most participants (361 [52.2%]) remained stable in both eyes (within 0.50 D) between their baseline and 8-year follow-up visits (Table 3). As shown in eTable 2 in the Supplement, the 8-year change in axial length, but not lens thickness or corneal radius, was significantly different between those with a myopic shift compared with those with no refractive change. Myopia progression of 0.25 D or less per year (generally the minimum detectable change in refraction) in at least 1 eye was observed in 19 participants (2.7%).

Table 3. Refraction Shift Over 8 Years Among 691 Participants.

Refractive shifta	Left eye, No. (%)
	Myopic shift	No change	Hyperopic shift
Right eye
Myopic shift	152 (22.0)	56 (8.1)	0
No change	53 (7.6)	361 (52.2)	45 (6.5)
Hyperopic shift	0	10 (1.4)	14 (2.0)

^a Myopic/hyperopic shift defined as a change in refraction of 0.50 diopters or more.

There was a significant longitudinal change in spherical equivalent, axial length, and lens thickness after correcting for sex, race, and the major known risk factors of myopia (Table 4; eTables 3 to 6 in the Supplement). Corneal radius did not change significantly over time.

Table 4. Estimated Annual Change in Myopia-Related Parameters.

Measure	Estimate (97.5% CI)a	Statistical outcomeb
		F statistic	P value
All participants
Spherical equivalent	−0.041 (−0.055 to −0.027)	F1,1972 = 27.6	<.001
Axial length	0.020 (0.014 to 0.025)	F1,1962 = 87.3	<.001
Corneal radius	0 (−0.002 to 0.002)	F1,1965 = 18.5	.91
Lens thickness	0.020 (0.017 to 0.024)	F1,1819 = 170.7	<.001
Men
Spherical equivalent	−0.018 (−0.036 to 0.001)	F1,959 = 17.7	.03
Axial length	0.010 (0.004 to 0.016)	F1,956 = 53.1	<.001
Corneal radius	−0.001 (−0.003 to 0.002)	F1,956 = 1.3	.50
Lens thickness	0.019 (0.015 to 0.022)	F1,887 = 160.2	<.001
Women
Spherical equivalent	−0.044 (−0.063 to −0.025)	F1,1005 = 6.1	<.001
Axial length	0.022 (0.015 to 0.030)	F1,999 = 9.1	<.001
Corneal radius	0 (−0.002 to 0.002)	F1,1001 = 0.3	.97
Lens thickness	0.021 (0.016 to 0.026)	F1,924 = 10.0	<.001

^a 97.5% CI shown for significance at P < .025 (with the Bonferroni correction for the 2 main outcome measures [spherical equivalent and axial length; .05/2]).

^b Corrected for ethnicity, conjunctival autofluorescence area, education, parental myopia, and sex (where appropriate).

The multivariable analyses showed that men had lower spherical equivalents, longer axial lengths, and flatter corneas than women (eTables 3 to 6 in the Supplement). However, the age by sex interaction results suggested that female participants had higher rates of spherical equivalent decrease and axial elongation than male participants, as shown in Table 4 and eTables 3 to 6 in the Supplement.

On average, East Asian participants had higher longitudinal rates of axial elongation and corneal flattening compared with White participants, albeit a small difference of only 0.014 (95% CI, 0.001-0.027) and 0.008 mm per year (95% CI, 0.004-0.012), respectively. There was no other significant main or interaction effect of race (eTables 3 to 6 in the Supplement). Parental myopia was significantly associated with a faster rate of spherical equivalent decrease (estimate, 95% CI, 0.004-0.020; −0.012 D per year for each additional parent with myopia; P < .001) and axial elongation (estimate, 95% CI, 0.002-0.008; 0.005 mm per year for each additional parent with myopia; P < .001).

Discussion

Several reports on myopia incidence and progression in school-aged children, especially in East Asian countries where myopia rates are the highest in the world, have been published. The annual myopia incidence have been reported to range from 7% to 30% in East Asian children^{26,27,28,29,30,31} and 1% to 3% in White children,^27,32,33 depending on geographical location and age. For example, in East Asian children aged 6 to 7 years at baseline, annual myopia incidence was higher in those living in Singapore, China, or Hong Kong (11% to 24%)^26,28,30 compared with children the same age in Australia (7%).^27,28 Many studies have also reported that myopia incidence in children decreases with older age.^26,28,34

Exploration of adult-onset myopia incidence, on the other hand, has been limited. In the 1990s, studies on university students in their late teens or early 20s have reported that the annual myopia incidence was between 2.5% and 13%.^7,839 While these estimates on university studies cannot be broadly applied to the general population, they provide evidence that it is common for myopia to start developing after childhood and adolescence.

In our study of young adults from a general population, we found an 8-year incidence of 14% and that spherical equivalent progressed by −0.04 D per year on average, with 38% of participants experiencing a myopia shift of 0.50 D or more in at least 1 eye over 8 years, in contrast to a mean age of myopia stabilization at approximately 15 years reported by the Correction of Myopia Evaluation Trial.⁵ With younger generations increasingly pursuing postgraduate education,³⁵ we may expect more at-risk young adults to develop myopia in their 20s or early 30s. Even in nonuniversity students or graduates, individuals are likely to start their first full-time occupation in or just prior to their 20s (students are aged 17 to 18 years when they leave school), and the rise in indoor occupations will inevitably result in the development or progression of myopia in a substantial proportion of the population.

Indeed, we observed an inverse association of increased sun exposure, as quantified using CUVAF area with incident myopia. Similarly, previous studies^{36,37,38,39,40} have noted a protective effect of increased time outdoors against myopia, but findings on whether it reduces myopia progression have been conflicting. The lack of association between ocular sun exposure and refractive measure change may also be partly owing to use of sunglasses or hats in some adults, which filters out incident ultraviolet rays and thus is protective against enlargement of CUVAF area,⁴¹ while still allowing exposure to higher levels of outdoor lighting.

Additionally, we did not find a significant association between highest level of education with rate of change in refractive measures. Instead, unmodifiable factors, such as race, sex, and parental myopia, appeared to have stronger associations with the rate of change in refractive measures than environmental factors.

Women were more likely than men to develop myopia and had greater changes in refractive measures between ages 20 and 28 years. Longitudinal studies on school-aged children in East and South Asian countries have similarly reported higher myopia incidence^26,30,34 and faster myopia progression^28,42 in girls compared with boys. Likewise, the Correction of Myopia Evaluation Trial reported that myopia progressed faster in girls than in boys, in terms of spherical equivalent but not axial length.⁴³ In previous studies involving children, the differential effect of sex on myopia progression and axial elongation may be influenced by pubertal growth spurts,⁴⁴ but this is unlikely to be a factor in our young adult sample. Instead, this difference between young men and young women may reflect the modern societal push for higher education in girls and women, as reflected by the increasing proportion of women with higher education than men,⁴⁵ and a tendency for women to work in indoor-based occupations in Australia.⁴⁵ However, the associations between female sex and myopia progression were significant even after correcting for education and CUVAF area. Moreover, higher level of education was not significantly associated with myopia incidence or progression. It is possible that some other lifestyle habits, biological, or hormonal factors may mediate this age and sex interaction effect during young adulthood, and this should be explored in further studies.

Over the 8 years of the study, there were crystalline lens thickening and axial elongation in this sample. The latter is particularly concerning as it is strongly believed that longer axial length increases the risk of myopia-related complications.⁴⁶ Fricke et al⁴ estimated that more than 55 million people (0.6% of the world population) will be visually impaired from myopic macular degeneration alone, including 18 million who will be blind, in the year 2050 if we do not implement interventions to slow myopia progression. Similarly, Cheung et al³ predict a retinal detachment epidemic as a consequence of a surge in prevalence of high myopia. Myopia management strategies targeting control of axial elongation should therefore be considered in young adults exhibiting myopia progression.

Strengths and Limitations

A main strength of the current study is the large sample of community-based young adults, rather than recruiting participants from universities or myopic cohorts, as has been done in previous studies on young adults.^7,8,11,47,48 The Raine Study participants have also been shown to be generally representative of the Western Australian population of the same age.⁴⁹ However, this study has limitations. Our findings may not be generalizable to recent immigrants of Western Australia, which may comprise a higher proportion of people of East Asian decent than the current cohort. There was also a substantial proportion of participants who did not attend the follow-up eye examination. This was partly because we had to cease data collection early due to the COVID-19 pandemic. It is unclear how this may have affected the incidence estimates. Participants with eye issues may be more likely to attend the eye examination, resulting in an overestimation of myopia incidence, but these participants are also likely to have had more frequent optometrists visits and thus did not feel the need to attend the follow-up eye examination, resulting in an underestimation of myopia incidence. We were also unable to ascertain whether there was a differential rate of change within the follow-up period, for example, if progression was faster during the early 20s than mid to late 20s.

Conclusion

Nonetheless, our findings provide evidence that myopia can start to develop and continue to progress during young adulthood. The eye continues to elongate axially in some participants during young adulthood, which may contribute to the increased risk of myopia-related complications as these young adults reach middle and older age. Our findings highlight the need for research into myopia control methods in young adults in addition to those currently being researched in children.

Supplement.

eAppendix. Additional analyses examining the association between education and myopia

eTable 1. Refraction and axial length of participants with incident high myopia between ages 20 and 28 years

eTable 2. Eight-year change in ocular biometric measures between eyes without and with a myopia shift

eTable 3. Multivariable analysis outcomes of spherical equivalent (D)

eTable 4. Multivariable analysis outcomes of axial length (mm)

eTable 5. Multivariable analysis outcomes of corneal radius (mm)

eTable 6. Multivariable analysis outcomes of lens thickness (mm)