Introduction

Myopia is an epidemic disease in eastern Asia, and the prevalence rate of high myopia is high in schoolchildren1,2. In Taiwan, the prevalence of myopia in schoolchildren has dramatically increased in the past 30 years2. The age of myopia onset has also decreased2. Therefore, effective treatments in retarding myopia progression are warranted3,4.

Atropine, an antimuscarinic antagonist, was effective in myopia control in past randomized control trials (RCTs)5,6,7,8,9,10. From ATOM1 and ATOM2 to the LAMP study, daily use of low-concentration atropine also showed effectiveness but fewer side effects, and rebound myopia compared with higher concentrations of atropine5,8,9. Although atropine concentrations greater than 0.05% are associated with increased side effects and a higher risk of rebound myopia, in Taiwan, 0.125% atropine remains the most commonly prescribed dosage for myopia control, which has also been shown effective in myopia control5,11. This is primarily because it is the lowest concentration reimbursed by the National Health Insurance (NHI) before 2023.

Despite its widespread use globally, the underlying mechanism of atropine in slowing myopia progression is still not fully understood. The retina, a light-sensitive layer of the eye that receives visual input, has been proposed to be involved in multiple signalling pathways of atropine in controlling myopia. Atropine was proven in past studies to increase dopamine and reduce protein levels of GABA transporter-1 in the retina, facilitating neuro-transmitting pathways in inhibiting myopia progression12,13,14. Nevertheless, there were concerns over the side effects of atropine on the retina, as existing evidence is conflicting. Several studies have shown that 0.1% atropine may alter neural activities of the inner layer of retina, as evidenced by an electroretinogram15,16. Conversely, research on 0.01% low-dose atropine over 14 days reported no atropine-related changes in the pattern ERG17. Because the use of low- and moderate-dose atropine ( < = 0.125%) on myopia control is usually long-term, with lots of medical providers in Taiwan adopting the continuous use of atropine from myopia onset to adolescence (around 18 years old)18, it is important to elucidate the mechanism and address long-term safety issues of atropine on the retina.

Optical coherence tomography (OCT) and optical coherence tomography angiography (OCTA), a non-invasive imaging technique, have emerged to be effective methods to study the structure and microvasculature of the retina and provide information for clinicians to localize retina pathology. There are some articles on OCT/OCTA of high myopic subjects, but only one longitudinal study investigating the effect of atropine on the retina in young myopes using this imaging modality19. However, the study had a relatively short follow-up period of 3 months with only a very low concentration of atropine 0.01% and less focus on the foveal parameters. The effects of higher-dose atropine, such as dose over 0.125%, on the retina have not been fully studied. As the foveal region is specialized for high-resolution colour vision and is often involved in pathologic myopic patients, we explore the longitudinal effect of 0.125% atropine on the foveal parameters, including area and perimeter of the foveal avascular zone (FAZ), the acircularity index (AI), and foveal vessel density via OCT/OCTA, to identify the effect of atropine on the foveal retina in a 12-month follow-up period.

Results

Baseline characteristics of the study population

A total of 63 eyes from 36 children were included in our study. Table 1 shows the baseline characteristics of our study participants with a mean age of 8.9 ± 2.2 (range, 6.0–15.3) years. There were 21 boys and 15 girls in this study, in which 33 right eyes and 30 left eyes were analysed. The mean follow-up duration was 8.8 ± 3.1 months. The mean axial length (AL) at baseline was 23.8 ± 0.9 mm. As for the foveal parameters, mean FAZ area, FAZ perimeter, AI, and foveal vessel density-300 (FD-300) were also shown in Table 1.

Table 1 Baseline patient characteristics.
Full size table

Atropine effect on AL in young myopes

During the 12-month longitudinal follow-up period, the estimated mean of AL showed a significant reduction on 1-month, 3-month, and 6-month follow-ups when adjusted by the age of our subjects, with a p value of 0.001, < 0.001, and 0.003, respectively. Detailed results are shown in Fig. 1.

Fig. 1
figure 1

The changes of estimated means of AL during longitudinal follow-upa. AL in myopic children treated with 0.125% atropine changed significantly 1, 3, and 6 months after initiating atropine using a linear mixed model. aEstimated means of AL were adjusted by age. *Significant estimated mean compared with baseline. AL: axial length.

Full size image

Significant decrease in FAZ area and FAZ perimeter within the first month; significant decrease in FD-300 and significant increase in AI within first 3 months of atropine use in young myopes

Figure 2 shows the changes in the estimated means of foveal parameters compared with the baseline during the follow-up period. Compared with the baseline, the FAZ area and perimeter decreased significantly at 1 month, with p values of 0.014 and 0.041, respectively. On the other hand, the increase of AI and decrease of FD-300 were significant at 3 months, with p values of 0.047 and 0.035, respectively.

Fig. 2
figure 2

The changes of estimated means of foveal parameters during longitudinal follow-upa. The changes in FAZ area, FAZ perimeter, AI, and FD-300 in myopic children treated with 0.125% atropine. FAZ area (a) and perimeter (b) significantly reduced one month after the initiation of atropine but remained stable at baseline at other follow-up periods. The increase of AI (c) and reduction of FD-300 (d) were significant three months after the initiation of atropine but also remained stable to baseline at other follow-up periods. Significance was estimated with the linear mixed model. aEstimated means of foveal parameters were adjusted by AL. *Significant estimated mean compared with baseline. AL: axial length; FAZ: foveal avascular zone; AI: acircularity index; FD-300: foveal vessel density-300; F/U: follow-up.

Full size image

Discussion

In this study, we examined the longitudinal effect of 0.125% atropine on the foveal microvasculature, with the focus within the 1-mm region centered on the fovea. Our result showed that FAZ area and perimeter revealed a significant decrease, especially in the first month after the application of atropine for myopia control in children. Moreover, AI significantly increased 3 months after atropine use. We also found that the reduction of FD-300 was significant, only 300 μm outside of FAZ in the first 3 months of atropine therapy. However, the changes were not significant throughout the follow-up period. To the best of our knowledge, this is the first study providing an in-depth evaluation of the effect of atropine 0.125% on foveal microvasculature in a 12-month follow-up period.

The baseline parameters in our study were comparable with past research. Wang et al. demonstrated the 3-month short-term effect of 0.01% atropine in young children. Their study showed remarkable myopic change with axial elongation within 3 months19. On the contrary, our study showed a slight reversal in AL at 1-month, 3-month, and 6-month follow-ups. The difference suggested a more robust efficacy of 0.125% than 0.01% atropine in myopia control and decreased initial AL. According to the ATOM 2 study, 0.1% atropine had significantly better efficacy in AL retardation than 0.01% atropine5. Furthermore, in the most recent network meta-analysis of randomized controlled trials on atropine myopia control, moderate to high concentration atropine was more effective than 0.01% atropine regarding refraction changes or axial elongation20. As for short-term AL reversal, several studies in the past also demonstrated similar results21,22,23,24. Our past study on myopia control with 0.125% atropine showed that the shortest AL occurred at the 4.8-month follow-up, which also echoed the current study, with remarkable shortening in AL shown in 1-month, 3-month, and 6-month follow-ups22.

FAZ and foveal vessel density were important OCTA foveal parameters. FAZ was an avascular area with the highest density of cone photoreceptors, supplied by the surrounding foveal/parafoveal capillary networks. The combined analysis of foveal vessel density and FAZ provided crucial information on the vascular status of the fovea25. Several studies investigated the relationship between foveal vessel density, FAZ, and myopia19,26,27,28,29,30,31,32. However, the correlation between the two was still controversial, with some studies showing larger FAZ areas in high myopia patients28,29,30 and some showing no significant correlation19,27,31,32. Regarding foveal vessel density, most studies evaluated the effect of myopia on macular or parafoveal vessel density. Only one study studied the effect of myopia on foveal vessel density, finding that AL was not correlated with foveal vessel density in myopic patients29. This was also the only cross-sectional OCTA study evaluating foveal parameters on myopic children, with the mean age of study subjects around 13 years old, regardless of myopia control modalities.

As for longitudinal OCTA studies on atropine-treated myopes, Wang’s study with 0.01% atropine was the only longitudinal study that also evaluated foveal parameters. They compared the mean change of FAZ size over 1-month and 3-month follow-ups and found that the change was insignificant19. In the current study, we found an initial decrease in the area and perimeter of FAZ in myopes within the first month after 0.125% atropine treatment, which had never been proposed in the past as far as we knew. One possible explanation was the vasodilation effect of atropine on ocular vessels. Atropine, as a muscarinic receptor antagonist, could increase the release of nitric oxide from vasodilator nerves33 and stimulate the release of dopamine, which could also increase the diameters of retinal vessels34,35. With the dilation of para-FAZ vessels, the area and perimeter of FAZ would decrease correspondingly. Another explanation was the initial AL reversal in our cohort. Studies have shown that retinal vessel calibre narrows with increasing AL, indicating altered retinal blood flow in highly myopic patients and contributing to FAZ enlargement36,37. Conversely, the initial shortening of AL might reduce mechanical stretching of the globe, leading to a decrease in FAZ area and perimeter. These two explanations might play important roles in the change of FAZ. Nevertheless, the observed decrease in FAZ area/perimeter becoming nonsignificant after three months of follow-up could potentially be attributed to tachyphylaxis, warranting further investigation.

In terms of foveal vessel density, the study from Wang et al. did not reveal a significant change in foveal density parameters19. However, we found that the area 300 μm outside of FAZ showed a significant decrease in density in a 3-month follow-up in 0.125% atropine-treated myopia children. A prior study indicated that ciliary muscle paralysis induced by atropine has the potential to alter the mechanical forces within the eyeball, leading to a transition from a prolate to a more oblate shape23. Although this study focused on higher doses of atropine (1%) as investigated in the ATOM1 study, it offers a potential explanation for the observed reduction in FD-300 during the initial three months of our study. However, further large-scale studies are warranted to confirm these findings. Such a phenomenon could elucidate the observed reduction in FD-300 during the initial three months of our study. As the retinal thickness was thinner within 300 μm outside of FAZ, the change in the shape of the eyeball could result in a more significant change in vessel density. However, detailed experiments on the pathophysiology of atropine were still needed to corroborate our hypothesis. The current study also found a significant increase in AI during the 3-month follow-up. This was more likely attributed to the fact that FD-300 decreased during the corresponding period with the changes in the surrounding capillary network, which may decrease the circularity of FAZ after the temporary application of atropine in young myopes.

Our study is subject to certain limitations. First, our study had no control group; however, we mitigated the risk of bias by employing a consistent measurement device and measuring patients continuously. Therefore, the patients’ baseline data can be used as a control for changes afterward. Second, the accuracy of vessel density measurement may be affected by magnification errors attributed to AL variations38. To address this concern, we employed the linear mixed model with an adjustment for AL to correct potential magnification errors, thus enhancing the accuracy of our data. Third, our investigation is characterized as a small prospective cohort study, with relatively short follow-up period and some patients missing follow-up visits. Larger, more comprehensive prospective studies with longer follow-up period are warranted to validate our findings.

In summary, our study highlights the effects of 0.125% atropine on the foveal microvasculature. While there were significant initial changes in FAZ area, FAZ perimeter, AI, and FD-300, these changes became insignificant over the 12-month follow-up period, suggesting the general long-term safety of atropine on the fovea. However, given that concentrations of 0.125% atropine have been associated with changes in foveal microvasculature, future studies should explore correlations between these microvascular changes and visual function, including contrast sensitivity and color perception.

Methods

We enrolled our study subjects prospectively from June 11, 2020 to February 23, 2023. Patients (1) who were aged from 5 to 18 years old, (2) with baseline cycloplegic spherical equivalent between − 0.5D to -6.0D, (3) with baseline astigmatism less than 2D, (4) with anisometropia less than 1.5D and (4) received 0.125% atropine treatment in single or both eyes as the first myopia controlling methods, were included in our study. Eligible participants in our study were required to meet the following exclusion criteria: we excluded potential participants who had (1) any history of systemic disease or ocular pathology, (2) incomplete chart records, (3) less than six months of follow-up after the initiation of 0.125% atropine treatment, and (4) adhering to atropine treatment < 80% of the time during the follow-up period, as documented in their medical records.

We performed AL measurement, completed the OCT/OCTA examination, and recorded comprehensive medical records, including detailed ophthalmology examinations and treatments before and after initiation of atropine treatment. The study was approved by the Chang Gung Memorial Hospital Institutional Review Board (IRB No.: 202000452B0C602) and followed the tenets of the Declaration of Helsinki. Informed consent for participation was obtained from parents and/or legal guardians.

Myopia control and follow-up protocol

AL was recorded on every patient during the first visit. Atropine treatment was prescribed to the patients with risks of high myopia development (e.g., rapid increase in AL) after the second follow-up visit. The follow-up period was scheduled at 1 month, 3 months, 6 months and 12 months after the initiation of 0.125% atropine treatment. During follow-ups, all examinations were performed by experienced technicians under the services of Dr. C.F. Liu.

Ocular examination

All participants received standardized ocular examination, including cycloplegic refraction, was measured 30 min after cycloplegia, which was achieved using 1% tropicamide (Mydriacyl; Alcon Vision, LLC, Fort Worth, TX, USA) and 10% phenylephrine hydrochloride (Phenylephrine Eye Drops; Wu Fu Laboratories Co., Ltd, Yilan, Taiwan) three times at 10-minute intervals. After confirming no light reflex in the enlarged pupil, cycloplegic refraction was assessed using an Auto Ref/Keratometer (ARK-1a/ARK-1; Nidek Co., Ltd, Gamagori, Japan), and the average of three successive measurements was taken as the objective cycloplegic refraction error during every visits. A cycloplegic spherical equivalent was used to evaluate the cycloplegic refraction error and was calculated as the sum of one half of the cylinder power plus the sphere power. AL measurements (IOL Master® 500; Carl Zeiss Meditec AG, Jena, Germany) and SD-OCT (AngioVue, Optovue RTVue XR Avanti; Optovue Inc., Fremont, CA, USA) were also performed after cycloplegia.

OCTA images of study participants were acquired through a 304 × 304 A-scan covering an area of 3 × 3 mm2 centred on the fovea in the macular region. The foveal region was defined as the 1-mm diameter circle centred on the fovea. The ANGIOVUE software (version: A2017,1,0,151, Optovue Inc., Fremont, CA, USA) automatically calculated the FAZ area, FAZ perimeter, AI, and FD-300 of the eye. AI was defined as the perimeter of FAZ to the perimeter of a circle with equal area, and FD-300 was defined as vessel density within the 300-µm wide annulus around FAZ. The foveal parameters were determined by the total layer of the retina, which ranges from the internal limiting membrane to 10 μm below the outer plexiform layer. Figure 3 illustrates these four parameters. All OCTA examinations were conducted using a standardized protocol. The machine’s tracking and centration algorithms ensured consistent foveal alignment during acquisition, and all scans underwent quality control to exclude artifacts or decentration. At least two scans per participant were obtained, with only the highest-quality and best-aligned images included in the analysis.

Fig. 3
figure 3

Illustration of foveal parameters. (a) The anatomy of the FAZ. Foveal area is the 1-mm circle centred on the fovea. The avascular area within the foveal area is the FAZ. FD-300 is the vessel density of the 300-µm circular annulus around FAZ. (b) The acircularity of FAZ. Our study demonstrated the acircularity of FAZ as the AI, which is the perimeter of FAZ to the perimeter of a circle with an equal area. (c) OCTA image in the left eye of a 9-year-old boy at baseline. The FAZ area, FAZ perimeter, AI, and FD-300 were measured using the automated calculation output from the ANGIOVUE software (version A2017.1.0.151, Optovue Inc., Fremont, CA, USA). FAZ: foveal avascular zone; FD-300: foveal vessel density-300; AI: acircularity index; OCTA: optical coherence tomography angiography.

Full size image

Statistical analysis

SPSS (version 26.0; IBM Corp., Armonk, NY, USA) was used to conduct all statistical analyses. Numerical variables, including age, follow-up period, AL, FAZ area, FAZ perimeter, AI, foveal thickness, and FD-300, were summarized as means ± standard deviation (SD), and categorical variables such as gender and eye were summarized as frequency. The effects of atropine treatment on AL, FAZ area, FAZ perimeter, AI, and FD-300 were examined using a linear mixed model, with patient ID included as a random effect to account for within-subject variability. The estimated means of AL were adjusted by age as shown in Figure 1, and AL adjusted the estimated means of other foveal parameters as shown in Figure 2. Statistical significance was assigned as p < 0.05.