Abstract

Aims: To compare event-based glaucoma progression analysis (GPA) I with new GPA II software and pattern deviation-based trend analyses (visual field index [VFI]) to detect progression in a glaucoma population.

Methods: A retrospective study that included 90 eyes of 90 patients with a minimum of five reliable visual field tests and a follow-up period of at least 2 years.

Results: Event-based GPA II detected progression in 16.7% of eyes in which trend analysis VFI failed. GPA detected progression 6.8 months earlier than VFI. GPA I and II showed excellent agreement (k=0.94). Agreement between VFI and mean deviation (MD) linear analysis and with GPA criteria was k=0.52 and k=0.48, respectively. Mean rates of progression of MD and VFI were -0.41 dB and -1.30% annually, respectively (rho=0.824; P < 0.0001). Using VFI, mean follow-up time was 6.12 and 4.89 years (P= 0.004) and mean number of visual field tests was 7.33 and 6.01 (P= 0.023) in eyes with and without progression, respectively.

Conclusions: Event-based software GPA I and II had an excellent agreement. Event analysis showed earlier and greater sensitivity for detecting progression than VFI analysis and both had only moderate agreement. Trend analysis VFI is likely to detect progression in patients with a greater number of visual field tests and a longer follow-up time. The VFI analysis seems to be more accurate than MD analysis for determining rate of progression.