AIMS—To evaluate the reproducibility of the retardation values (change in polarisation) obtained with the scanning laser polarimeter in a series of normal subjects and glaucoma patients. To improve the analysis of the raw data by devising and evaluating a blood vessel removal algorithm. METHODS—Scanning laser polarimetry was performed on 10 normal subjects and 10 glaucoma patients. A series of six images was obtained from each eye. The normal subjects were re-imaged 3 months after their initial assessment. The retardation values obtained from each eye were analysed using the authors' own methods, including the use of an algorithm to remove blood vessels from the polar profiles. The reproducibility of these measurements and the performance of the blood vessel removal algorithm were assessed. RESULTS—The "individual point" coefficient of variation was approximately 12.5% for normal subjects and 17.0% for glaucoma patients. The "integral" coefficient of variation for these groups was approximately 5.5% and 9.5% respectively. The reproducibility of the measurements did not improve with an increased number of measurements. There was no difference in the reproducibility of the measurements in normal subjects over time. The blood vessel removal algorithm improved the reproducibility of the measurements when the shape of the profile was assessed. CONCLUSION—The intraoperator reproducibility of retardation values obtained with the scanning laser polarimeter is satisfactory for its use as a clinical tool. The use of a blood vessel removal algorithm improves the reproducibility of the measurements and also assists the clinician in the interpretation of the polar profiles. Furthermore, it allows the construction of normal database polar profiles, thereby enabling the identification, location and quantification of retinal nerve fibre layer damage in an "at risk" individual's polar profile.
