Related Experiment Videos
Analysis of tear-protein patterns as a diagnostic tool for the detection of dry eyes
F H Grus1, A J Augustin, N G Evangelou
1Department of Ophthalmology, University of Bonn, Germany. 100024.1664@Compuserve.COM
Purpose:
To analyze the electrophoretic patterns of tear proteins from diabetic (DIDRY) and non-diabetic (DRY) patients with dry-eye disease in comparison with the patterns in healthy subjects (CTRL). The patterns were classified using multivariate statistical methods.
Methods:
A total of 119 eyes were examined in this study (50 DIDRY, 39 DRY, 30 CTRL). The patients were primarily grouped according to the results of the basic secretory test (BST) and subjective symptoms (burning, foreign body sensations, tearing, "dryness" of the eyes). Patients with values < or = 10/5' plus subjective symptoms were classified as dry-eye patients. Tear proteins were separated by sodium-dodecyl-sulfate poly-acrylamide gel electrophoresis (SDS-PAGE). Digital image analysis was done using the ScanPacK (Biometra, Gottingen, Germany). A data set was created from each electrophoretic pattern. The data were analyzed by multivariate analysis of discriminance, k-means cluster analysis and with a factor analysis before k-means cluster analysis as a data reduction tool.
Results:
The protein patterns of the three groups were significantly different (Wilks' lambda: 0.1425; P < 0.01): P < 0.05 (CTRL-DRY), P < 0.00005 (CTRL-DIDRY), and P < 0.0005 (DIDRY-DRY). There were more peaks/electrophoretic lane in the DRY and DIDRY groups (P < 0.05) than CTRL. Classification of all samples as DRY and CTRL gave the following results: known patterns 97% correct; unknown patterns 71.4% correct. Classification in DIDRY, DRY or CTRL was 92% correct for known patterns and 43% for unknown patterns. Using k-means cluster analysis, 72% of patients previously classified as "dry-eye" according to the BST value were classified as "dry-eye" based on their electrophoretic data too.
Conclusions:
Analysis of protein patterns and statistical evaluation are suitable for the detection of dry eyes. Tear-protein pattern analysis can provide more information on the pathogenesis of the disease. Thus, this new method might be more reliable than the BST value for both diagnostic and therapeutic purposes.