Purchase this article with an account.
F. W. Chang, M. Eller, S. B. Steinman, P. Gunvant; Accuracy of the Heine Lambda 100 Retinometer. Invest. Ophthalmol. Vis. Sci. 2008;49(13):2538.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Patients undergoing retinometer testing before cataract surgery can show lower acuity than Snellen acuity. We investigated causative factors by comparing acuity for projected Snellen letters, the Heine Lambda 100 retinometer, and red-black gratings on a computer screen.
Snellen acuity was measured by standard clinical techniques. Equivalent psychophysical procedures were used for retinometer and computer acuity measurements (n=58). The orientations of the grating bar orientations were randomized at each acuity level. Computer-based gratings subtended visual angles close to those of the retinometer gratings. Acuity was recorded as the last bar width in which the subject could correctly identify at least 3 of 5 grating orientations, i.e., > 60% correct detection. In a further experiment (n=30), both grating orientation and bar width were randomly selected on each presentation and a psychometric function was generated.
All subjects were ≥ 20/20 as measured by the Snellen letters and computer-generated gratings. A statistically significant number of subjects had lower retinometer acuities. Correct detection remained close to 100%, decreasing for < 20/25 stimuli, but only dropping to 91% for 20/25 stimuli.
The lower visual acuity measured with the retinometer is not attributable to differences in criterion level. Observers were encouraged to guess, forcing them to set a uniform criterion level. The same threshold % correct was used on both retinometer and computer grating acuity. If the smallest retinometer bars were really smaller than 20/25, % correct would decrease with a high slope for that bar width alone, but % correct decreased prior to the 20/25 stimulus. If all gratings had a smaller width or decreased contrast, % correct would fall below 100% for bar widths much larger than 20/25, yet % correct detection only dropped to 91% for 20/25 stimuli. A threshold near 20/20 would result if smaller bar sizes were present in the retinometer. Lower acuities on the retinometer can therefore be attributed to the lack of a 20/20 target. The acuity values obtained by the retinometer are acceptable because the acuity would remain a conservative estimate of post-operative vision.
This PDF is available to Subscribers Only