Predictive value of confocal scanning laser for the onset of visual field loss in glaucoma suspects

To perform a longitudinal multivariate analysis of the ability of Moorfields Regression Analysis (MRA) to predict the onset of glaucoma in a population of patients with suspected glaucoma due to appearance of the optic nerve head.

Prospective longitudinal evaluation of a diagnostic test.

Single, randomly selected eye of prospectively recruited patients with suspected glaucoma based on the optic nerve head appearance on stereophotographs and normal baseline visual field results.

The MRA was evaluated at baseline (Heidelberg Retina Tomograph; Heidelberg Engineering, Dossenheim, Germany), and visual field tests were repeated every 6 months. A longitudinal multivariate proportional hazard ratio (HR) analysis was performed, and likelihood ratios and positive and negative predictive values were compared.

Onset of visual field losses.

The study included 230 eyes that were followed up during a mean period of 62±14 months, ranging from 4 to 7 years. The predicted HR (for onset of visual field losses) of the MRA temporal-inferior sector outside normal limits was 3.65 (95% confidence interval [CI], 2.32-5.75; P < 0.0001). An MRA temporal-superior sector outside normal limits had an HR of 3.43 (95% CI, 2.21-5.32; P < 0.0001).

The temporal-inferior and temporal-superior positions of the MRA are highly predictive for the onset of visual field loss in glaucoma suspects.