Abstract
PURPOSE: To determine the utility of root mean squared (RMS) deviations of steady-state accommodation as a noncycloplegic marker for spasm of near reflex (SNR) vis-à-vis regular refractive errors. METHODS: Binocular steady-state responses of accommodation, pupil, and vergence of 20 patients with accommodative spasm subtype of SNR (SNR-A; 9-23 years) and 91 with regular refractive errors (29 emmetropes, 41 myopes, 21 hyperopes; 19-38 years) was recorded in the uncorrected refractive error state for 120 seconds using a dynamic (50 frames per second), infrared photorefractor. Mean and RMS deviation of raw data was calculated for three 20-second-long epochs and their diagnostic utility was determined using standard ROC curves. RESULTS: RMS deviations of accommodation increased with mean refractive error in SNR-A (y = -0.23x + 0.38; r2 = 0.69; P < 0.001) and regular refractive error (y = -0.02x + 0.10; r2 = 0.14; P = 0.002) cohorts, albeit with steeper slope and higher y-intercept in the former rather than the latter cohort. RMS deviation of 0.19D reliably distinguished SNR-A from regular refractive errors with a sensitivity and specificity of 95.2% and 92.2%, respectively [mean (±1 SEM) area under ROC curve: 0.98 ± 0.01]. The sensitivity, specificity, and area under ROC curve for RMS deviations of pupil (66.7%, 80%, and 0.70 ± 0.09) and vergence (52.4%, 84.6%, and 0.68 ± 0.08) were smaller than accommodation. CONCLUSIONS: RMS deviations of steady-state accommodation is a robust noncycloplegic marker for differentiating SNR-A from regular refractive errors. Pupil and vergence fluctuations have limited utility in this regard. TRANSLATIONAL RELEVANCE: RMS deviations of accommodation may be easily obtained using commercial photorefractors, and the cut-off values reported herein may be implemented to identify SNR-A during refractive error screening.