Validity of noncycloplegic refraction in the assessment of refractive errors: the Tehran Eye Study

To determine the sensitivity and specificity of noncycloplegic autorefraction for determining refractive status compared to cycloplegic autorefraction.

The target population was noninstitutionalized citizens of all ages, residing in Tehran in 2002, selected through stratified cluster sampling. From 6497 eligible residents, 70.3% participated in the study, from August to November 2002. Here, we report data on 3501 people over the age of 5 years who had autorefraction with and without cycloplegia (two drops of cyclopentolate 1.0% 5 min apart, with autorefraction 25 min after the second drop).

Overall, the sensitivity of noncycloplegic autorefraction for myopia was 99%, but the specificity was only 80.4%. In contrast, the sensitivity for hyperopia was only 47.9%, but the specificity was 99.4%. At all ages, noncycloplegic autorefraction overestimated myopia and underestimated hyperopia. Overestimation of myopia was highest in the 21-30 and 31-40 year groups. Underestimation of hyperopia was high up to the age of 50 (20-40%), but decreased with age, to about 8% after the age of 50, down to almost 0% after 70. The difference in mean spherical equivalent with and without cycloplegia fell from 0.71 dioptres (D) in the 5-10 age group to 0.14D in those over 70.

Use of noncycloplegic autorefraction in epidemiological studies leads to considerable errors relative to cycloplegic measurements, except in those over 50-60. The difference between cycloplegic and noncycloplegic measurements varies with age and cycloplegic refractive category, and there is considerable individual variation, ruling out adjusting noncycloplegic measurements to obtain accurate cycloplegic refractions.