Purchase this article with an account.
DT Edwards, GB Bartley, DO Hodge, EA Bradley; Eyelid position Measurement in Graves' Ophthalmopathy: A Comparison of Photographic and Clinical Techniques . Invest. Ophthalmol. Vis. Sci. 2002;43(13):350.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Purpose: To evaluate intraobserver reliability of a computer-based digital image measurement of eyelid position in Graves ophthalmopathy (GO) patients, and to compare digital image measurements with clinical measurements that coincided with the photography date. Methods: Digital images were created from 35mm color slides of 84 eyes (42 patients) with GO and were projected on a 15 inch flat-screen computer monitor. EKG calipers were used to measure interpalpebral fissure (IPF), margin-reflex distance (MRD), and inferior scleral show bilaterally. A ruler mounted at the lateral canthus at the time of photography allowed quantification of measurements to the nearest 0.5mm. Measurements were taken one week apart and intraobserver reliability was evaluated using a paired t-test. Agreement between digital image measurements and clinical measurements was summarized using mean differences and 95% confidence intervals. Results: Excellent agreement was observed between the two measurements taken from the digital images of clinical photographs (p≷0.05). No significant difference was found between the digital image and clinical measurements for the right IPF (mean difference -0.3mm, 95% CI -1.04, 0.40), left IPF (-0.4mm, 95% CI -1.15, 0.36), right inferior scleral show (-0.3mm, 95% CI -0.58, -0.06), and left inferior scleral show (-0.1mm, 95% CI -0.37, 0.15). A significant difference was found between the digital image and clinical measurements for both the right MRD (1.25mm, 95% CI 0.67, 1.83) and the left MRD (1.19 mm, 95% CI 0.62, 1.75). Conclusion: Measurement of several eyelid parameters in GO patients from computer-based digital images is repeatable and correlates well with clinical measurements for all parameters except MRD. The discrepancy in MRD measurement may result from short-term biologic variability of this parameter in GO patients. Alternatively, the observed difference may reflect a lack of validity in one of the techniques employed. Further studies examining the source of this discrepancy and evaluating interobserver variability of lid position measures are needed before either technique is used in prospective studies of GO.
This PDF is available to Subscribers Only