Using Atropine for Cycloplegic Refraction May Avoid Overestimation of Premyopia in Children

After cycloplegia, children who received cyclopentolate eye drops showed more myopic SE, a higher prevalence of premyopia and a lower prevalence of low hyperopia and moderate to high hyperopia compared with the atropine group. However, the study raises questions regarding the practicality of using atropine as a cycloplegic agent, particularly due to potential adverse events, such as prolonged blurred vision and discomfort associated with its extended duration of action. Additionally, there remains uncertainty about whether the higher cycloplegic refractive error achieved with atropine correlates with true accuracy in diagnosing refractive errors. Click image to enlarge. As myopia gains recognition as a significant global public health concern, the use of cycloplegic eye drops has increased in young children to help promote early detection and intervention. Cyclopentolate is often the agent of choice for cycloplegic refraction due to its rapid onset and short duration; however, a recent study published in JAMA Ophthalmology sought to evaluate how atropine—a more potent agent with a delayed onset of action and prolonged effects—influences the accuracy of refraction and diagnosis outcomes among children aged three to seven. It found that the use of atropine was associated with higher levels of hyperopia and a lower prevalence of premyopia (8.7%) vs. cyclopentolate (21.6%), though questions remain regarding the accuracy of refractive error determinations and prolonged adverse effects of using atropine as a cycloplegic agent. This investigation was a post hoc analysis comparing data from two population-based studies: the Preschool Children Refractive Development Pattern and Influencing Factors Study (2024), which comprised the atropine group, and the Elaborative Shanghai Childhood Ocular Refractive Development Study (2013-2014) for the cyclopentolate group. A total of 1,761 children (3,048 eyes) aged three to seven were analyzed, and propensity score matching was used to ensure a near-equal distribution of age and sex between both groups. Those in the atropine group were administered the agent twice daily for four days, while the cyclopentolate group received two drops, five minutes apart.The results indicated that the mean noncycloplegic spherical equivalent was comparable between groups (0.30D in the atropine group and 0.31D in the cyclopentolate group), while the mean cycloplegic spherical equivalent was greater in the atropine group at 1.56D compared to the cyclopentolate group at 0.97D. Notable differences in refractive states were also observed: moderate to high hyperopia was found in a greater percentage of the atropine group than in the cyclopentolate group (7.2% vs 2.7%), while low hyperopia was also more prevalent in the atropine group (82.8% vs. 74.0%). Conversely, the cyclopentolate group had a higher prevalence of premyopia and myopia than the atropine group (8.7% vs. 21.6% for premyopia; 1.3% vs. 1.8% for myopia). According to the authors of the study, these findings suggest that “using atropine for cycloplegia in preschool children yields less myopic refraction than cyclopentolate, potentially avoiding overestimation of premyopia prevalence.” However, it’s important to note that the study did not assess each cycloplegic agent in the same group of children.While these findings suggest that atropine results in more accurate refraction measurements, a recent commentary on the study, also published in JAMA Ophthalmology, posits that the findings might indicate that atropine simply measures more hyperopia or less myopia rather than providing accurate refractive error determinations; in other words, a higher cycloplegic refractive error with atropine does not necessarily correlate with better accuracy.“Atropine may be better at fully relaxing the accommodative system, which is important in those with hyperopia,” the commentators wrote. “However, for a child with myopia whose accommodation should be completely relaxed, a stronger eye drop may not necessarily lead to a more accurate determination of refractive error.” They further explained, “Certainly, the strength and dynamic nature of a child’s accommodative system clearly warrants cycloplegia for a more stable accommodative system; however, once the child’s system is fully relaxed, does adding more strength help?” In future research, the commentators suggest that assessing residual accommodative tone using an open-field autorefractor could clarify which eye drop is more effective at fully paralyzing the accommodative system, rather than merely focusing on the resulting refractive error.The commentary also highlights a significant oversight in the study, that being the absence of data on the prevalence and duration of adverse effects associated with atropine use, such as pupil size and accommodation changes. They argue that suggesting atropine over cyclopentolate without this information “makes it challenging to evaluate whether the benefits outweigh the risks.”In summary, while there is a theoretical advantage to identifying premyopia, the commentators argue that the negative side effects of blurred vision that lasts for a week and the time required to achieve full cycloplegia may lead many to find a daily regimen of 1% atropine eye drops impractical. They illustrate this situation using a metaphor: “Could you kill a bug with a hand grenade? Sure, but is there a way without so much collateral damage?” Thus, they surmise, “Many clinicians may continue to use cycloplegic agents with relatively shorter duration of action, such as cyclopentolate or tropicamide.”Click here for the study and here for the commentary. Wu H, Wang Y, Lu Q, et al. Atropine or cyclopentolate to diagnose premyopia in preschool children. JAMA Ophthalmol. September 25, 2025. [Epub ahead of print].Weise KK, Khanal S. Invited commentary: Atropine vs cyclopentolate for cycloplegic refraction in children. JAMA Ophthalmol. September 25, 2025. [Epub ahead of print].This article was developed by the editorial staff in conjunction with experts in the field. In the process, AI may have been among the editorial tools used to meet the goals of human editors, who approved all content.