Macular pigment optical density measurement in autofluorescence imaging: comparison of one- and two-wavelength methods

Background: Measurement of macular pigment (MP) can be performed by analysis of autofluorescence (AF) images. These can be obtained by standard 488-nm argon-imaging alone (one wavelength, 1-Lambda) or with additional digital subtraction of a second image at 514 nm (two wavelengths, 2-Lambda). The analyses are easy to perform, and we present a comparison of both methods and investigate their reliability and repeatability. Methods: Inter-individual variability of MP optical density (MPOD) measurements was assessed in single eyes of 120 subjects with a modified Heidelberg retina angiograph (HRA). MPOD values obtained with one (488 nm) Lambda (MPOD1 Lambda) were compared with those obtained with two (488 nm and 514 nm) Lambda (MPOD2 Lambda). To test the repeatability of the two methods, 20 subjects were subjected to five repeated measurements. Results: Among 120 individuals, mean MPOD1 Lambda at 0.5 degrees eccentricity was 0.59 (range 0.06-1.32), mean MPOD2 Lambda was 0.5 (range 0.01-1.21). Apart from this systematic difference, 1-Lambda and 2-Lambda measurements at 0.5 degrees agreed well across the range of MPOD values (beta=0.964, 95% CI 0.891-1.096; R=0.83). At 2 degrees around the fovea, a systematic difference (0.11) was accompanied by declining agreement at higher MPOD values (beta=0.669, 95% CI 0.519-0.844; R=0.48). Among 20 subjects with five repeated measurements, the reliability ratio was 0.97 for 1-Lambda and 0.94 for 2-Lambda at 0.5 degrees and 0.93 and 0.94, respectively, at a distance of 2 degrees. Conclusions: Both methods showed a high repeatability with little influence of measurement error. They agree well at the fovea centre in terms of ranking individuals according to their MPOD, but provide increasingly deviating results at a distance of 2 degrees around the fovea, probably because the 1-Lambda method, in contrast to the 2-Lambda method, cannot compensate for disruptive influences and for heterogeneous distributions of the lipofuscin fluorophores. The 1-Lambda method can be performed by standard HRA and could therefore be used for screening in multicentre studies, but only approaches the actual amounts of MP. The 2-Lambda method remains the more precise method for MPOD measurement in autofluorescence imaging.