Risk factors for rapid axial length elongation with low concentration atropine for myopia control

Three hundred and twenty-eight myopic children, randomized to use either 0.01% (N = 166) or 0.02% (N = 162) atropine were enrolled in this study. Gender, age, body mass index(BMI), parental myopia status, atropine concentration used, pupil diameter, amplitude of accommodation, spherical equivalent refractive error (SER), anterior chamber depth (ACD) and axial length (AL) were collected at baseline and 1 year after using atropine. Rapid AL elongation was defined as > 0.36 mm growth per year. Univariate analyses showed that children with rapid AL elongation tend to be younger, have a smaller BMI, use of 0.01% atropine, narrow ACD, lower SER, shorter AL, smaller change in pupil diameter between 1 year and baseline (all P < 0.05). Multivariate logistic regression analyses confirmed that rapid AL elongation was associated with children that were younger at baseline (P < 0.0001), use of 0.01% atropine (P = 0.04), a shorter baseline AL (P = 0.03) and a smaller change in pupil diameter between 1 year and baseline (P = 0.04). Younger children with shorter AL at baseline, less change in their pupil diameter with atropine treatment and using the lower of the two atropine concentrations may undergo rapid AL elongation over a 12 months myopia control treatment period.


Discussion
In this prospective study, the axial length of the younger children, using lower concentration atropine, with shorter axial length at baseline and smaller change in pupil diameter may still increase rapidly while receiving atropine treatment in mainland China. These factors seem to suggest that these children may be less responsive to the effects of low concentration atropine.
Rapid AL elongation despite the use of low concentration atropine treatment has been described in other different atropine concentrations and populations studies. In the ATOM2 study 5,6 , where children were randomized to atropine concentrations of 0.01%, 0.1% and 0.5%, the AL change after twelve months was 0.24 ± 0.19 mm, 0.13 ± 0.18 mm and 0.11 ± 0.17 mm, respectively. In a study of 0.01%, 0.025% and 0.05% atropine concentrations, Yam et al. 12,15 found that the AL change after 1 year was 0.36 ± 0.29 mm, 0.29 ± 0.20 mm and 0.20 ± 0.25 mm, after 2 years was 0.59 ± 0.38 mm, 0.50 ± 0.33 mm, and 0.39 ± 0.35 mm, respectively. In a one year study of Korean myopic children 10 , the AL elongation was about 0.44 ± 0.32 mm, 0.30 ± 0.24 mm and 0.23 ± 0.25 mm in 0.01%, 0.025% and 0.05% atropine, respectively. In a study of Chinese myopic children 14 , the AL elongation after one year was about 0.32 ± 0.19 mm and 0.41 ± 0.09 mm in 0.01% atropine and placebo, respectively. By comparison, Note: # due to: Spherical equivalent refractive error > -1.25D (N = 8) or < -6.0D (N = 2), astigmatism ≤ -2.0D (N = 3), intraocular pressure >21mmHg (N = 2), had cataract surgery (N = 1). 24 Lost to follow -up 10  www.nature.com/scientificreports/ in the present study the AL elongation was 0.36 ± 0.19 mm with 0.01% and 0.30 ± 0.22 mm with 0.02% atropine, 0.02% atropine had a better effect on myopia control than 0.01% atropine 13 . There seems to be considerable variation in the range of AL change results with some children changing very little and others increasing very quickly. Other reports of low concentration atropine for controlling myopia progression in children measured refraction and not AL 8,9,11,18 . They defined myopia progression of more than 0.50D or 1.0D at 1 year as rapid myopic progression. Again rates of refractive progression varied considerably between studies, risk factors for faster myopic progression in low concentration atropine studies include younger age at baseline 11,18 , higher initial myopia 8 , and a family history of high myopia 10 . This confirms the importance of conducting appropriately controlled studies to explore independent risk factors associated with rapid AL elongation in myopic children in mainland China treated with low concentration atropine.
In the current study, baseline age had a significantly negative correlation with AL increase in myopic children using 0.01% or 0.02% atropine. The younger children were at baseline, the more the AL elongation was evident at the end of the study period. This was consistent with three other studies which explored the same relationship between baseline age and AL increase in myopic children using atropine and whose baseline profiles were similar to the current study. Joachimsen et al. 11 , Wei et al. 14 and Lee et al. 18 found that younger children in German, mainland China and Taiwan may still progress quickly while using 0.01%, 0.01% and 0.05% atropine, respectively. It is well established that there is slowing of physiological change in AL in older children 19 . Conversely, Moon et al. 10 and Wu et al. 8 reported no relationship between baseline age and AL change in Taiwanese children using 0.05% atropine and in Korean children using 0.01%, 0.025% or 0.05% atropine. These contrasting results may arise due to different study populations, different baseline characteristics and methodological and analytical differences.
We also found that the children with 0.01% atropine had a more rapid AL elongation and a worse myopic control effect than 0.02% atropine 13 . Other reports have also found a concentration-dependent response in myopia control with low concentration atropine 6,10,12,14 . Side effects and adverse effects are reportedly similar in 0.01% and 0.02% atropine 13,20 , suggesting that if children using 0.01% atropine are not achieving adequate myopia control, changing to the 0.02% concentration could be considered. It is important to use an individualized atropine concentration to control myopia development. Table 1. Baseline characteristics of participants mean ± SD or n (%). Rapid change in AL defined as > 0.36 mm/year. ※ Outdoor activity = outdoor exercise + outdoor leisure activity. # Nearwork = 3* (homework + reading + playing on cell phone) + 2* (using computer + playing video game) + 1* (watching TV). www.nature.com/scientificreports/ The children with a smaller change in pupil diameter from baseline to 1 year had rapid AL elongation in our study. The underlying reasons why these children did not respond well are unclear. So far, it is unclear how low concentration atropine acts to inhibit myopia progression. One theory 21 suggests that increased ultraviolet exposure (secondary to pupil dilation) may increase collagen cross-linking within the sclera, thereby limiting scleral growth during myopia progression and AL elongation. The greater change in pupil diameter, may be due to the better absorption of the drug, greater collagen cross-linking within the sclera and a superior effect on controlling myopia progression. Recently, studies have shown that adding 0.01% atropine eye drops to OK therapy was more effective than OK alone for controlling myopia progression in children 22,23 . One of the possible mechanisms may be low concentration atropine's mydriatic effect, as other studies have shown that myopia control effect of OK lens was also affected by pupil diameter 23,25 . Children with larger pupil diameter would receive a greater proportion of peripheral myopic defocus associated with OK lens wear, which in turn may be relate to better myopia control 23 . This is a possible mechanism that warrants further investigation.
We found that a rapid AL elongation related significantly to shorter AL at baseline, but not to initial levels of myopia. In other words, the shorter AL at baseline, the more the AL elongation was found with low concentration atropine, irrespective of baseline refractive error. Lin 26 however found that highly myopic children had less increase in AL after using 0.125% atropine. Wu et al. 8 conversely found that higher initial myopia corresponded significantly to a higher chance of myopia progression irrespective of atropine concentration. Moon et al. 10 found that baseline refraction and AL were not associated with the rate of myopia progression irrespective of atropine concentration, however only age, family history and AL were selected for inclusion in the multivariate logistic regression analyses of this study. Several studies 27,28 have established a strong association between the eye's AL and its refractive error. Myopia progression usually occurs due to excessive AL elongation of the eye, as evidenced by the strong association between changes in myopia progression and changes in AL growth 19,29,30 . There was a strong association between the change in AL and change in SER after 1-year treatment in the present study. Another study 14 found that 0.01% atropine reduced myopia progression by 34.2% and axial elongation by Table 2. Univariate and multivariate logistic regression analyses showing risk factors associated with rapid axial length elongation over 12 months. The bold represents P value less than 0.05. Note: # due to: Spherical equivalent refractive error > -1.25D (N = 8) or <-6.0D (N = 2), astigmatism ≤-2.0D (N = 3), intraocular pressure > 21 mmHg (N = 2), had cataract surgery (N = 1). OR odds ratio; CI confidence interval. a Adjusted: body mass index (BMI), spherical equivalent refractive error (SER), accommodation amplitude (AMP), anterior chamber depth (ACD), outdoor activity and nearwork. b Adjusted: age, parental myopia status, SER, AMP, corneal curvature, ACD and nearwork. c Adjusted: gender, age, BMI, SER, AMP, corneal curvature, ACD, outdoor activity and nearwork. d Adjusted: age, BMI, SER, AMP, pupil diameter, corneal curvature, ACD and nearwork. ※ Outdoor activity = outdoor exercise + outdoor leisure activity. # Nearwork = 3* (homework + reading + playing on cell phone) + 2* (using computer + playing video game) + 1* (watching TV).

Risk factors
Non-adjusted OR, (95% CI) P value Adjust OR, (95% CI) P value www.nature.com/scientificreports/ 22.0% compared with the placebo group. Howerver, in the ATOM2 study 6 , while the degree of myopia was stable, the AL continued to increase between 8 and 24 months after using 0.01% atropine. For assessing the efficacy of interventions to control the progression of myopia in children, we recommend that the change in AL is not be used interchangeably with the change in refraction. This study did not consider the rate of change in AL using age matched emmetropes, however previous studies 31,32 describing the growth curve for AL with age showed that the rate of AL elongation in emmetropes was slower than in myopes, and the AL growth slowed down in emmetropes as the children became older, but little or no decrease in the rate of change of AL in myopic children is observed. Other limitations were that the study measured axial elongation over a 1 year period only and the potential bias introduced by loss to follow-up of approximately 15% in each low concentration atropine group. Further investigations including age matched emmetropes and using a study design with a longer follow-up time are required to confirm the present findings. An additional limitation of the present study, is that the age at onset of myopia was not available. Previous studies 33,34 have identified that age at onset of myopia is a risk factor for myopia progression and this factor should included in future studies.
In conclusion, our study showed that the axial length in young children, using lower concentration of atropine, having a shorter axial length at baseline and smaller change in pupil diameter may still increase rapidly while receiving low concentration atropine treatment in mainland China. If the axial length of these children with these characteristics continues to grow at a rate similar to that before the start of atropine treatment, then a change in atropine concentration or combining with other myopia control therapies to control the progression of axial length may be considered.

Methods
We report a prospective doubled blinded randomized controlled trial of 0.01% and 0.02% atropine in myopic children recruited from the First Affiliated Hospital of Zhengzhou University. This study was part of a larger series of clinical studies which also evaluated a further three non-contemporaneous non randomized patient groups, not reported here, including subjects wearing single-vision spectacles, subjects with low myopia using 0.005% atropine, and a subsequent group evaluating 0.02% atropine used every other day.
In this prospective study, three hundred and twenty-eight Chinese myopic children (right eyes, Han nationality) who presented between July 2016 and October 2017 were recruited. The inclusion criteria were: 6-14 years of age, cycloplegic autorefraction SER from − 6.00 D to − 1.25 D, astigmatism less than 2.0 D, anisometropia of less than 1.0 D, monocular best corrected visual acuity of 16/20 or better, intraocular pressures (IOP) between 10 and 21 mmHg, no other eye diseases and surgery, no ocular and systemic conditions that might affect vision or vision development. Exclusion criteria were: previous use of atropine, pirenzepine, soft contact lenses, rigid gas permeable and orthokeratology contact lens to control myopia progression; inability to comply with the study visit schedule. This study was approved by the Medical Ethics committee of the First Affiliated Hospital of Zhengzhou University and registered in the Chinese Clinical Trial Registry (registration number: ChiCTR-IPD-16008844, first registration in 14/07/2016). This study conformed to the tenets of the Declaration of Helsinki. Written informed consent was obtained from parents before the procedures, and possible risks were fully explained before treatment initiation.
The 0.01% and 0.02% atropine eye drops (pH value of 5.4-5.6, 3-mL sealed bottle, kept away from the light, 15-25 ℃ room temperature storage, discarded one month after opening the bottle) were made by diluting 1% atropine (Eye & ENT Hospital, Affiliated to Fudan University) with saline under sterile conditions, with the addition of the preservative (0.3 mg/ml ethylparaben). Subjects were double-blindly and randomly assigned to 0.01% and 0.02% concentration, where random numbers were generated by computer algorithm. The children were prescribed constant wear of full correction single-vision spectacle lenses (SV) with the highest positive/ least negative power consistent with optimum visual acuity and with one drop of 0.01% or 0.02% atropine eye drops administered into both eyes once nightly before bed time.
Corneal power, anterior chamber depth (ACD) and AL were evaluated using a non-contact partial coherence interferometer (IOLMaster; Carl Zeiss Meditec AG, Germany). On each occasion, five successive measurements were taken and their mean was used as a representative value. Based on other two similar studies 35,36 , rapid AL elongation was defined as > 0.36 mm growth per year (i.e., equivalent myopic progression 1.00 D per year) and children were categorized as either having rapid or slow AL elongation.Outdoor activity time (hours per day) and nearwork time (hours per day) 12 were assessed using a paper questionnaire. Cycloplegic autorefraction was performed after the instillation of four drops of compound tropicamide eye drops 37,38 (0.5% tropicamide and 0.5% neo-synephrine) (Santen, Japan) administered 10 min apart in each of the patients' eyes. Ten minutes after the instillation of the fourth drop, three autorefraction measurements were taken (Topcon RM 8000A, CA) and a mean was obtained. The degree of myopia was expressed as SER.
Children were provided with four bottles of eye drops after the first examination and subsequent each followup. When the child was reexamined, they were asked to return all four bottles and compliance was assessed according to the remaining amount of eye drops. One-drop of eye drops is about 0.04 ml, one child will use more than 2.4 ml each month. If a child's remaining eye drops in any bottle exceeded 10% (about 1 ml) of the total amount of each bottle, then his compliance was not good. In order to improve compliance, we have also adopted two methods: 1. we explained to children and their parents the importance of using eye drops correctly every day for myopia control after they entered the research group. 2. The WeChat group was set up for all the children's parents in the research group, there were two colleagues of the research group answered all kinds of questions encountered by the children in the medication process.
Continuous baseline variables were expressed as mean ± standard deviation (SD) and evaluated by using Student t-test. Categorical variables, such as gender, atropine concentration and parental myopia status, were