Modelling the topography of absolute defects in patients exposed to the anti-epileptic drug vigabatrin and in normal subjects using automated static suprathreshold perimetry of the entire 80° visual field

Abstract

(i) To map the spatial distribution of absolute visual field loss (AL) in patients exposed to vigabatrin and to compare the findings with AL in normal individuals, and (ii) to describe the relationship between the major risk factors for absolute loss (gender, age, cumulative dose), and the severity of field loss, expressed as the number of locations with AL per eye. Visual field plots were retrospectively reviewed from 428 individuals. Perimetry was done with the Tübingen Automated Perimeter (using a threshold-oriented, marginally supraliminal strategy, 80° eccentricity, 99 test locations). Three hundred and sixty-one individuals were ophthalmologically normal, and 67 were patients exposed to vigabatrin who had performed reliably during perimetry prior to neurosurgery for epilepsy. Two hundred and sixty-six of the 361 normals were included on the empirical basis that they manifested AL at no more than 19 stimulus locations per eye (the 74th percentile). The frequency per eye of AL at each of 99 stimulus locations was determined for the normals and patients. The effects of age, gender and cumulative dose of vigabatrin on the number of ALs per individual was assessed by an analysis of covariance. A stochastic model was developed to determine the 50% probability of AL at each location. Thirty-nine of the 67 vigabatrin patients exhibited at least 20 locations with AL. The number of ALs was independent of age (p = 0.7603). The frequency of AL was 14.8% in the peripheral field and 1.0% in the central field (odds ratio 16.7; 95% CI 15.0-18.6%; p < 0.0001). Those exposed to vigabatrin exhibited a frequency of 17.1% ALs, compared to 5.2% for the normal individuals (odds ratio 3.77, 95% CI 3.6-4.0%; p < 0.0001). According to the modelled data, males exhibited 20.9% (95% CI 1.3-44%; p = 0.0360) more absolute losses than did females. The number of absolute losses per person doubled with an increase in cumulative dose of vigabatrin of 936 g (95% CI 775-1181g). According to this retrospective study, the spatial configuration of absolute defects attributable to vigabatrin indicates sparing of the temporal field up to approximately 60° eccentricity. Such a finding is likely to explain, at least in part, the initially asymptomatic nature of the defect. This study reconfirms a (cumulative) dose effect of vigabatrin on the extent of absolute field loss, with a greater risk for male gender.