Objective: To determine the optimum perimetric criteria for short-wavelengt
h automated perimetry on the basis of probabilistic maps that best discrimi
nate between normal and glaucoma.
Design: Comparative cross-sectional study.
Participants: Ninety-five glaucomatous eyes and 128 normal eyes were includ
ed in the study.
Interventions: The subjects underwent retinal nerve fiber layer photographi
c evaluation and short-wavelength automated perimetry.
Main Outcome Measures: A probabilistic map of differential points for short
-wavelength automated perimetry was elaborated, and the number of altered p
oints was calculated for each visual field and level of significance (95%,
98%, 99%, and 99.5%).
Results: The presence of a cluster of four or more points outside 95% norma
l probability limit showed a good combination of sensitivity and specificit
y. At deeper levels, P < 1% and P < 0.5%, the best combination is achieved
with three points outside the normal probability limits.
Conclusions: The optimum criterion to define glaucomatous abnormalities in
short-wavelength automated perimetry is the presence of a cluster of four p
oints lower than P < 5% or a cluster of three points lower than P < 1%, Oph
thalmology 2001;108:285-289 (C) 2001 by the American Academy of Ophthalmolo
gy.