Abstract
Purpose: :
In population-based studies of older adults refractive error is frequently measured without cycloplegia. However, as younger adults are studied, it is important to know the measurement error associated with accommodative ability. The purpose of this paper is to compare refractive error measured on pharmacologically dilated versus undilated eyes.
Methods: :
Preliminary data from the Beaver Dam Offspring Study provided refractive error for both undilated and pharmacologically dilated eyes of 2399 participants. Refractive error was defined by the spherical equivalent (SE), calculated from autorefractor readings. Differences for each eye were calculated as the SE for the pharmacologically dilated eye minus the SE for the undilated eye. Generalized estimating equations, stratified by age groups, estimated mean differences, while accounting for the correlation between eye measurements from the same person. Myopia was defined as SE ≤ -1, emmetropia as -1< SE< 1, and hyperopia as SE ≥ 1.
Results: :
The mean age was 49 years (range 22 to 84). Among those aged 22 to 39 years, the estimated mean difference in SE was largest among the hyperopes (0.89 diopters, 95% CI 0.53, 1.25), followed by the emmetropes (0.55, 95% CI 0.48, 0.61), and then the myopes (0.28, 95% CI 0.23, 0.32). Differences between pharmacologically dilated and undilated SEs did not vary greatly by these classifications for older participants, but did decrease with age. The estimated mean difference was 0.41 diopters (95% CI 0.39, 0.44) for those aged 40 to 49, 0.14 (95% CI 0.12, 0.16) for those 50 to 59 years old, and 0.09 (95% CI 0.06, 0.12) for those 60 and older. Across all age groups, the agreement was high between classifications based on pharmacologically dilated SEs and on undilated SEs (84% to 91% agreement).
Conclusions: :
On average, spherical equivalents for pharmacologically dilated eyes were slightly greater (more hyperopic) than those for undilated eyes, particularly among hyperopes under 40 years of age. However, the magnitude of differences in refractive error (mean of <1 diopter) and impact on estimates of myopia and hyperopia suggest that cycloplegia may not be essential in epidemiologic studies of adults.
Keywords: clinical (human) or epidemiologic studies: biostatistics/epidemiology methodology • refraction • accomodation