Erratum: "Abstracts of the 29th BCLA Annual Clinical Conference, Brighton, 2005" (Contact Lens & Anterior Eye (2005) vol. 28 (3) (137-142) 10.1016/j.clae.2005.05.001)

Purpose: To examine the spectral radiance output of slitlamp cobalt blue illumination and the spectral transmission of yellow barrier enhancement filters to assess the optimisation of fluorescein image viewing and capture. Method: The spectral radiance of seven different models of slit-lamp cobalt luminance was measured with a Spectrascan PR-650 photometer and the illumination outputs from three slit-lamps of the same model were measured. The spectral transmission of three barrier filters (two of which were integrated into the slit-lamp optics) were measured with a Cary 2300 Spectrophotometer. Results: The spectral radiance of the cobalt blue illumination was typically between 404 and 504 nm, with a peak at between 452 and 484 nm. All the slit-lamps examined showed a similar light distribution (S.D. of difference 0.20-1.13%). Comparison between three slitlamps of the same model showed a wide range of maximum light intensity (438-634 cd/m2), but a similar spectral radiance distribution. The barrier filters had a cut-off at 510-520 nm, achieving 90.0-98.5% transmission above 580 nm. However, the external cardboard mounted filter excluded less light below 500 nm (7.4%) compared to the two integrated filters (<0.1%). Conclusion: For optimal fluorescence, the light source should have maximum power at the wavelengths absorbed by the fluorophore (∼485 nm) and little power at the wavelengths emitted by the fluorophore (∼510-520 nm). Of the slit-lamps examined, only between 8.9 and 29.3% of the illumination output was optimised for >80% fluorescein excitation and 1.9-10.6% of the illumination overlapped with that emitted by the fluorophore. A barrier filter enhances observation by obscuring the non-fluorescing reflected light. As the fluorescein excitation and emission spectra at the average pH of the eye overlap at 500 nm, yellow barrier enhancement filters should ideally have a lower cut-off.