Comparative Analysis of Cycloplegic and Non-Cycloplegic Refraction in Children and Adolescents: Implications for Accurate Assessment of Refractive Errors

Abstract

Purpose: This retrospective study aimed to compare the efficacy of cycloplegic (CR) versus non-cycloplegic refraction (NCR) methods in detecting refractive errors among children and adolescents. Methods: Electronic data from pediatric ophthalmology clinics at the University Hospital “Sveti Duh”; Zagreb, Croatia, from January 2008 to July 2023, were analyzed. Comprehensive eye examinations, including Logarithmic Visual Acuity tests, subjective refraction, cycloplegic retinoscopy, slit lamp, and fundus examinations, were conducted. Results: The dataset included 1075 individuals, with 180 undergoing NCR and 895 undergoing CR. In premyopes, the NCR group had a longer follow-up (5.04 vs. 3.45 years; p < 0.001) with similar SE progression. In low myopia, NCR showed more negative first visit SE (−1.86 D vs. −1.35 D; p < 0.001) and faster progression (p = 0.01). In high myopia, follow-up was longer in NCR (5.08 vs. 2.08 years; p = 0.03) with no other significant differences. SE progression was highest in 4–6-year-olds and significantly faster in NCR (−0.61 vs. −0.40 D/year; p = 0.05). Conclusions: Cycloplegic refraction is essential for accurately assessing refractive status, especially in cases of low myopia, as it prevents misclassification and ensures precise evaluation in children and adolescents, thereby facilitating the appropriate diagnosis and treatment of refractive errors.

1. Introduction

As the global myopia epidemic continues to grow, the choice between cycloplegic refraction (CR) and non-cycloplegic refraction (NCR) becomes increasingly important, as it directly impacts the accuracy of refractive error detection and subsequent clinical management. Cycloplegic and non-cycloplegic refraction methodologies offer distinct insights into refractive errors, each presenting its advantages and considerations. To identify refractive errors in young individuals, it is essential to conduct refraction measurements under cycloplegic conditions due to the significant variability in accommodation observed in children [1]. Cycloplegia, which temporarily immobilizes accommodation, is an objective refraction measurement method in young children. Their active accommodative system often results in inaccuracies during non-cycloplegic refraction [2], leading to overestimated myopia and underestimated hyperopia [3]. In the clinical realm, cycloplegic refraction is proven as the superior method for accurately assessing refractive errors, due to its high precision [4]. Nonetheless, the invasive nature and time consumption of this procedure, involving the application of eye drops, can be uncomfortable for many patients, potentially causing distress, particularly among younger children [5].

In spite of these associated side effects and considerations related to time and cost, both the Royal College of Ophthalmologists (UK) and the International Myopia Institute (IMI) recommend cycloplegic refraction for children under the age of 12 years to ensure accurate refractive error measurement [6]. Additionally, the College of Optometrists (UK) recommends conducting a cycloplegic examination when testing young children to achieve precise refraction and optimize fundus examination, which can help identify underlying causes promptly and facilitate timely treatment [7]. While non-cycloplegic refraction is less time consuming and without associated side effects, it does not fully address the influence of accommodation, which becomes a significant factor in ocular examinations of children [8]. Especially when determining the minimal negative prescription for a myopic child, caution is warranted, as excessive prescription may stimulate accommodation and axial length growth, potentially contributing to myopic progression [9]. Misclassifying individuals due to myopic overestimation within various refractive categories, especially when coupled with inaccurate estimations of risk factors, can result in a failure to identify groups with potentially rapid myopia progression [8]. While less likely to produce false-positive identifications, these errors still pose challenges to the accurate detection of relevant risk factors.

This retrospective study aimed to provide a comprehensive analysis of the refractive characteristics and temporal trends among children and adolescents prescribed spectacle correction, by comparing those who underwent cycloplegic retinoscopy prior to prescribed spectacle correction with those who did not. The subgroup of children examined under cycloplegic conditions was previously described in detail in the CroMyop study [9], where the standardized administration of 1% tropicamide ensured full cycloplegia and accurate refractive assessment. One key knowledge gap this study addresses is the lack of comprehensive understanding regarding age-specific differences in cycloplegic (CR) and non-cycloplegic refraction (NCR) discrepancies. While previous research has shown that NCR tends to overestimate myopia, there is limited exploration of how these discrepancies manifest across different age groups, particularly in younger children with stronger accommodative responses [3,10]. This gap is critical, as accommodation-related biases in refraction can lead to the misclassification of refractive errors, which may delay appropriate intervention [11]. Based on the sample size and demographic characteristics of the children without cycloplegic spherical equivalent (SE) data, stratified random sampling of premyopic, low myopic, and high myopic individuals was performed using data from the cohort with cycloplegic SE, thereby enabling a matched comparison across refractive categories. This approach facilitates a clearer understanding of how the choice of refraction method influences refractive error classification, highlights age-related trends in CR/NCR discrepancies, and underscores the clinical implications of using NCR in underdiagnosed populations [3,4]. This study seeks to examine how cycloplegic refraction may enhance refractive error assessment and reduce the likelihood of misclassification, particularly in younger children with stronger accommodative responses, thereby supporting more informed clinical decision-making.

2. Materials and Methods

2.1. Study Population

This retrospective study included children and adolescents aged ≥4 to <19 years diagnosed with primary myopia and/or compound myopic astigmatism, who underwent eye examinations at the pediatric ophthalmology clinics of the University Eye Department, University Hospital “Sveti Duh” in Zagreb, Croatia, between 1 January 2008, and 1 July 2023. To be eligible, individuals had to have at least two visits separated by a minimum of 6 months. The dataset included 1075 children and adolescents, comprising 180 individuals without cycloplegic spherical equivalent (NCR group) and 895 individuals with cycloplegic SE (CR group). The detailed demographic characteristics are presented in

Table 1. Sample demographic characteristics of individuals without cycloplegic SE (NCR group) and individuals with cycloplegic SE (CR group).

		n	Gender (Female, %)	Age of Diagnosis (Mean ± SD, y)
Premyopia	NCR Group	17	71.2	9.41 ± 2.56
	CR Group	51	60.8	11.37 ± 3.59
Low myopia	NCR Group	156	58.2	12.46 ± 3.15
	CR Group	813	58.7	11.18 ± 3.53
High myopia	NCR Group	7	31.6	13.02 ± 3.55
	CR Group	31	48.4	11.44 ± 4.35

CR—cycloplegic refraction; NCR—non-cycloplegic refraction; n—number of individuals; SD—standard deviation; y—years.

.

The exclusion criteria were individuals with eye comorbidities such as mixed astigmatism, strabismus, corneal diseases, retinopathy of prematurity, or amblyopia. Additionally, individuals aged 19 years or older and those with allergies to cycloplegic drugs were excluded from the study. Ethical approval was obtained from the ethics committees of University Hospital “Sveti Duh” (Ur. No. 01-03-632/4, dated 12 February 2021) and the School of Medicine of the University of Zagreb (Ur. No. 380-59-10106-22-111/23, Class: 641-01/22-02/01, Zagreb, 21 February 2022). The study adhered to the tenets of the Declaration of Helsinki.

2.2. Refraction Protocol and Visual Acuity Assessment

The patients underwent a Logarithmic Visual Acuity (LogMAR) test using the Lea Symbols and Numbers chart at both near (40 cm) and distant (3 m) conditions, with measurements taken monocularly and binocularly. This testing followed standardized guidelines to ensure consistency and reliability [12]. The choice of chart was adapted as appropriate for the child’s age and developmental level to obtain the most accurate measurements. Assessments were conducted both with and without correction, with refractive errors recorded in diopters of spheres and cylinders on specific axes measured in degrees. Uncorrected visual acuity (UCVA) and best-corrected visual acuity (BCVA) were measured using subjective refraction.

2.3. Cycloplegia Protocol

Retinoscopy during cycloplegia was performed using tropicamide 1% (Mydriacyl^®, Alcon Laboratories Inc., Geneva, Switzerland), which was administered three times at 15 min intervals to induce optimal mydriasis and cycloplegia, as per standard clinical protocols [13]. To rule out eye comorbidities and secondary causes of myopia, slit lamp and fundus examinations were performed. Additional patient data, including information from legal guardians and self-reported medical history, were recorded for analysis.

2.4. Definition of Refractive Status

Spherical equivalent was calculated as the spherical power plus half of the cylindrical power. In accordance with the IMI definition [14], premyopia was defined as a condition where the spherical equivalent refractive error of an eye is >−0.50 D and ≤0.75 D when ocular accommodation is relaxed. Myopia was classified as a condition in which the spherical equivalent refractive error of an eye is ≤−0.50 D and >−6.00 D, whereas high myopia is defined as a spherical equivalent refractive error ≤ −6.00 D when ocular accommodation is relaxed. The spherical equivalent of the worse eye was used to classify individuals into premyopia, myopia, and high myopia.

2.5. Statistical Analysis and Data Modeling

Data analysis was performed using R software (version 4.0.3, https://www.r-project.org/ accessed on 10 February 2025). An independent samples t-test was applied to compare the differences between groups. Myopia progression was defined as the difference in SE between subsequent measurements, with negative values representing myopia progression. Stratified random sampling of premyopic, low myopic, and high myopic individuals was performed from the cohort with cycloplegic SE measurements (CR group) to match the distribution of individuals without cycloplegic SE data (NCR group). This process generated a randomized cycloplegic SE group (rCR group), allowing for balanced and representative subgroup comparisons. Matching was performed based on the number of individuals within each myopia subgroup, age at diagnosis, and gender distribution, in order to minimize potential confounding factors. The stability and reliability of the sampling approach were assessed by analyzing the mean spherical equivalent, age at diagnosis, and proportion of females across 20 iterations. The standard deviation of the mean SE remained within ±0.05 diopters, and the coefficient of variation for age was consistently below 2.5% across all refractive categories. Additionally, the variation in the proportion of females remained within ±5% across iterations.

3. Results

Following stratified random sampling, a subset of individuals from the CR group (rCR group) was selected to match the NCR group in terms of refractive category distribution, age of diagnosis, and gender composition—

Table 2. Sample demographic and refractive characteristics of randomly selected individuals with cycloplegic SE (rCR group) following stratified sampling.

	n	Gender (Female, %)	Age of Diagnosis (Mean ± SD, y)	First Visit Cycloplegic SE (Mean ± SD, D)	Cycloplegic SE 11–24 Months Progression Rate (Mean ± SD, D/y)
Premyopia	17	68.1	11.37 ± 3.59	−0.11± 0.1	−0.36 ± 0.25
Low myopia	156	57.3	11.18 ± 3.53	−1.53 ± 0.96	−0.15 ± 0.12
High myopia	7	36.2	11.44 ± 4.35	−7.61 ± 1.89	−0.25 ± 0.19

D—diopters; n—number of individuals; SE—spherical equivalent; SD—standard deviation; y—years.

. The rCR group included 17 premyopic, 156 low myopic, and 7 high myopic individuals. The gender distribution and mean ages in the rCR group were closely matched to those in the NCR group, supporting the comparability of the two groups.

The refractive and visual acuity characteristics of the NCR and rCR groups are presented in

Table 3. Sample refractive characteristics of individuals without cycloplegic SE (NCR group) and randomly selected individuals with cycloplegic SE (rCR group) following stratified sampling. Independent samples t-test was used to test the significance of difference between two groups.

	Follow-Up Period (Mean ± SD, Years)	First Visit Correction SE (Mean ± SD, D)	Correction SE 11–24 Months Progression Rate (Mean ± SD, D/y)	UCVA (Decimal)	BCVA (Decimal)
Premyopia (NCR Group)	5.04 ± 3.08	−0.05 ± 0.2	−0.37 ± 0.21	0.91 ± 0.10	0.95 ± 0.04
Premyopia (rCR Group)	3.45 ± 1.20	−0.04 ± 0.16	−0.38 ± 0.23	0.89 ± 0.10	0.94 ± 0.05
p-value, 95% CI	* <0.001, −1.49 to −0.09	0.46, −0.93 to 0.42	0.64, −0.81 to 0.53	0.52, −0.04 to 0.08	0.41, −0.01 to 0.03
Low myopia (NCR Group)	3.87 ± 1.54	−1.86 ± 1.01	−0.29 ± 0.24	0.55 ± 0.15	0.95 ± 0.03
Low myopia (rCR Group)	3.41 ± 1.12	−1.35 ± 1.03	−0.19 ± 0.21	0.59 ± 0.12	0.97 ± 0.02
p-value, 95% CI	0.56, −0.84 to −0.39	* <0.001, 0.25 to 0.71	* 0.01, −0.20 to 0.24	0.56, −0.08 to 0.11	0.60, −0.05 to 0.08
High myopia (NCR Group)	5.08 ± 3.55	−7.32 ± 1.71	0.51 ± 0.14	0.30 ± 0.22	0.90 ± 0.05
High myopia (rCR Group)	2.08 ± 1.15	−6.34 ± 1.56	0.40 ± 0.25	0.38 ± 0.20	0.91 ± 0.06
p-value, 95% CI	* 0.03, −2.39 to −0.07	0.28, −0.48 to 1.66	0.19, −1.82 to 0.36	0.43, −0.27 to 0.11	0.62, −0.04 to 0.06

BCVA—best-corrected visual acuity; CI—confidence interval; CR—cycloplegic refraction; D—diopters; NCR—non-cycloplegic refraction; n—number of individuals; rCR—randomized cycloplegic refraction; SD—standard deviation; SE—spherical equivalent; UCVA—uncorrected visual acuity; y—years; * Statistically significant (p ≤ 0.05).

. Among premyopic individuals, the mean follow-up period was significantly longer in the NCR group compared with the rCR group (5.04 ± 3.08 years vs. 3.45 ± 1.20 years; p < 0.001, 95% CI −1.49 to −0.09). No significant differences were observed in the first visit correction SE (p = 0.46, 95% CI −0.93 to 0.42) or in SE progression rates between 11 and 24 months (p = 0.64, 95% CI −0.81 to 0.53). The UCVA and BCVA values were similar between groups, with no significant differences (p = 0.52 for UCVA and p = 0.41 for BCVA).

In the low myopia subgroup, the NCR group showed a significantly more negative first visit correction SE compared with the rCR group (−1.86 ± 1.01 D vs. −1.35 ± 1.03 D; p < 0.001, 95% CI 0.25 to 0.71). SE progression rates were also significantly higher in the NCR group compared with the rCR group (p = 0.01, 95% CI −0.20 to 0.24). UCVA and BCVA were comparable between the groups, with no significant differences observed.

For high myopia, the NCR group demonstrated a longer follow-up period (5.08 ± 3.55 years vs. 2.08 ± 1.15 years; p = 0.03, 95% CI −2.39 to −0.07), but there were no significant differences in first visit correction SE, SE progression rates, UCVA, or BCVA between the NCR and rCR groups.

Correction SE progression rates stratified by age group are shown in

Table 4. Average correction SE progression rates (in diopters) between 11 and 24 months stratified by different age groups in individuals without cycloplegic SE (NCR group) and randomly selected individuals with cycloplegic SE (rCR group) following stratified sampling. Independent samples t-test was used to test the significance of difference between two groups.

			Age (Years)
			4–6	7–9	10–12	13–15	16–18
NCR Group	Progression Rate	n	17	25	62	43	33
		Mean (D/y)	−0.61	−0.42	−0.28	−0.13	−0.21
		SD	1.05	0.33	0.24	0.18	0.30
rCR Group	Progression Rate	n	17	25	62	43	33
		Mean (D/y)	−0.40	−0.34	−0.19	−0.14	−0.10
		SD	0.42	0.41	0.27	0.29	0.22
		p-value	* 0.05	0.44	0.24	0.55	0.08
		95% CI	−0.22 to 0.14	−0.25 to 0.11	−0.01 to 0.10	−0.01 to 0.03	−0.02 to 0.33

CI—confidence interval; D—diopters; NCR—non-cycloplegic refraction; n—number of individuals; rCR—randomized cycloplegic refraction; SD—standard deviation; y—years; * Statistically significant (p ≤ 0.05).

. In the NCR group, younger children (4–6 years) exhibited the fastest progression rate (−0.61 D/year), which gradually decreased with age, reaching −0.21 D/year in the 16–18 years group. Similarly, in the rCR group, progression rates were highest in the youngest age group (−0.40 D/year) and decreased to −0.10 D/year in the oldest group. Significant differences in progression rates between the NCR and rCR groups were observed only in the 4–6 years age group (p = 0.05, 95% CI −0.22 to 0.14).

4. Discussion

This study evaluated the impact of cycloplegic versus non-cycloplegic refraction on the classification and progression of myopia among children and adolescents. Consistent with previously published findings [2,3,11], our results suggest that assessing refractive error without cycloplegia may lead to an overestimation of myopia, particularly in cases of low myopia. Such overestimation could influence reported spherical equivalent values at the time of diagnosis, complicating the assessment of myopia progression and potentially misclassifying individuals at a higher risk of rapid progression.

As emphasized by the International Myopia Institute (IMI) [14], cycloplegia is recommended as the standard method for refractive assessment in research to ensure accurate classification within study cohorts. Most clinical trials investigating myopia control interventions have incorporated cycloplegic refraction as a fundamental requirement [9,15]. In our cohort, the retrospective analysis showed that the majority of children and adolescents (895 individuals) underwent cycloplegic refraction, whereas 180 individuals were prescribed spectacle correction based solely on non-cycloplegic measurements.

Using stratified random sampling, we matched individuals based on demographic and refractive characteristics to ensure balanced comparisons between the NCR and rCR groups. Within this matched framework, we observed significantly longer follow-up durations among premyopic and high myopic individuals in the NCR group. However, this finding should be interpreted with caution, as longer follow-up does not necessarily indicate a higher rate of myopia progression. It is plausible that children who did not undergo cycloplegic refraction were more likely to return for annual follow-up visits due to the less invasive nature of the examination, whereas those receiving cycloplegia each year may have been less inclined to attend follow-up appointments due to associated discomfort or parental concerns. These behavioral and procedural factors may have contributed to the observed differences in follow-up duration and should be considered when interpreting group-level outcomes.

Among the refractive subgroups, only individuals with low myopia in the NCR group demonstrated significantly greater myopia progression between 11 and 24 months compared with their counterparts in the rCR group. While stratified random sampling was used to match for age, sex, and refractive category, other important risk factors for myopia progression—such as family history, near-work activity, outdoor exposure, and urban versus rural living environments [4,16,17,18,19]—were not available in our dataset and could not be controlled for. These unmeasured variables may have contributed to the observed differences in progression rates and represent a potential source of residual confounding.

Previous studies have reported that non-cycloplegic refraction tends to yield spherical equivalent measurements approximately 0.6 to 0.8 diopters more myopic compared with cycloplegic refraction [10,20]. In our study, a similar trend was observed. Among individuals with low myopia, the first visit correction spherical equivalent was more negative in the non-cycloplegic group (−1.86 ± 1.01 D) compared with the cycloplegic group (−1.35 ± 1.03 D; p < 0.001, 95% CI 0.25 to 0.71). Furthermore, the mean cycloplegic SE on the first visit in the cycloplegic group was −1.53 ± 0.96 D. Although causality cannot be directly established from this study, the differences observed highlight the critical importance of the method of refractive assessment in both clinical practice and research settings.

Accurate cycloplegic refraction is particularly important in younger children. In our cohort, among children aged 4 to 6 years, the mean myopia progression rate was significantly greater in the NCR group (−0.61 D/year) compared with the rCR group (−0.40 D/year; p = 0.05), with notably higher variability (SD 1.05 vs. 0.42). These findings are consistent with previous studies demonstrating that non-cycloplegic refractions can overestimate myopia in young children due to their strong accommodative tone [21,22]. Similarly, Zhao et al. [23] reported that non-cycloplegic measurements frequently misclassify refractive error in preschool-aged populations. Together, these results underscore the necessity of cycloplegic refraction for the accurate assessment of refractive status in early childhood.

Moreover, recent developments in predictive modeling for myopia progression have demonstrated the value of incorporating cycloplegic data [24,25]. Huang et al. reported that including cycloplegic SE measurements in a time-aware deep learning model improved prediction accuracy from 78.5% to 87.2%, and reduced the mean absolute error (MAE) from 0.56 D to 0.42 D [25]. These findings underscore the potential utility of cycloplegic refraction not only for diagnosis but also for enhancing individualized risk assessment and management strategies.

Several limitations of this study should be acknowledged. First, the retrospective design precludes direct causal inference regarding the differences between the cycloplegic and non-cycloplegic groups. A paired-design approach, in which each child undergoes both cycloplegic and non-cycloplegic refraction, would have minimized inter-group variability and strengthened the conclusions. The differences in age distribution and follow-up periods between the groups could introduce potential confounding, although stratified random sampling was employed to reduce this bias. Visual acuity assessments were conducted under routine clinical conditions without standardized lighting control, which may have introduced measurement variability. Additionally, refractive error measurements were based on single final values recorded in clinical documentation, with no documentation of repeated measurements per eye, limiting the analysis of intra-examiner variability. Selection bias may also be present, as children in the non-cycloplegic group often did not receive cycloplegia due to parental refusal, cooperation difficulties, or logistical constraints. Furthermore, the small sample size in the premyopia subgroup, particularly in the non-cycloplegic group, limits the generalizability of the findings in this category. Although this study identified statistically significant differences in low myopia progression rates, the clinical relevance of the magnitude of these differences over a three-year period remains modest and should be interpreted cautiously.

Despite these limitations, this study possesses several notable strengths. It includes a large and well-characterized cohort of 1075 children and adolescents, utilizes stratified random sampling to ensure comparability across refractive categories, age of diagnosis, and gender distribution, and benefits from a long 15-year data collection period (2008–2023). Repeated sampling analyses confirmed the stability and reliability of group matching. Age-specific subgroup analyses were performed, providing new insights into how accommodation-related differences may influence the discrepancy between cycloplegic and non-cycloplegic refraction, particularly in younger children. The assessments were conducted by pediatric ophthalmologists, ensuring a high standard of clinical accuracy, and standardized Lea Symbols and Numbers charts were used to measure visual acuity appropriately across different age groups.

5. Conclusions

In conclusion, our current study highlights the limited effectiveness of non-cycloplegic refraction in accurately assessing the refractive status of children’s and adolescent’s eyes, particularly for those with low myopia. Therefore, the use of cycloplegia remains crucial in this regard.