Effect of cycloplegia on the measurement of refractive error in Chinese children

BACKGROUND: To compare the results of cycloplegic and non‐cycloplegic refractive error measurement in Chinese children, and to assess the relationship between age and the difference in refractive error measured with and without cycloplegia. METHODS: This was a prospective study that recruited 224 healthy Chinese children at an ophthalmology clinic from November 2016 to February 2017. Refraction before and after cycloplegia were measured using an auto‐refractor. Then spherical equivalent M, J(0), and J(45) were calculated. The enrolled children were allocated into three groups according to M: myopia, emmetropia, and hyperopia. The distribution of the refraction was further analysed by stratifying by age: four to six years, seven to 11 years, and 12 to 16 years. RESULTS: Mean non‐cycloplegic M, J(0), and J(45) were −1.68 ± 2.00 D, 0.05 ± 0.40 D, and 0.01 ± 0.35 D, while mean cycloplegic M, J(0), and J(45) were −1.16 ± 2.17 D, 0.02 ± 0.40 D, and −0.01 ± 0.35 D. Significant differences were found between cycloplegic and non‐cycloplegic M (p = 0.009), whereas there were no significant differences between cycloplegic and non‐cycloplegic J(0) and J(45) (p = 0.486 and p = 0.594, respectively). The differences between cycloplegic and non‐cycloplegic M were statistically significant in the four to six years group (p = 0.002) and seven to 11 years group (p = 0.023), whereas there was no significant difference between cycloplegic and non‐cycloplegic M in the 12 to 16 years group (p = 0.151). The proportion of myopia decreased from 78.1 per cent before cycloplegia to 71.4 per cent after cycloplegia, while the proportion of hyperopia increased from 12.1 per cent before cycloplegia to 21.4 per cent after cycloplegia. CONCLUSION: Non‐cycloplegic auto‐refraction is found to be inaccurate and not suitable for studies of refractive error in Chinese children.