Limiting Retinal Toxicity of Vigabatrin in Children With Infantile Spasms

Commentary

Vigabatrin Retinal Toxicity in Children With Infantile Spasms: An Observational Cohort Study.

Westall CA, Wright T, Cortese F, Kumarappah A, Snead III OC, Buncic JR. Neurology 2014;83:2262–2268.

OBJECTIVES: To determine time to vigabatrin (VGB, Sabril; Lundbeck, Deerfield, IL) induced retinal damage in children with infantile spasms (IS) and to identify risk factors for VGB-induced retinal damage (VGB-RD). METHODS: Observational cohort study including 146 participants (68 female, 81 male) with IS, an age-specific epilepsy syndrome of early infancy, treated with VGB. Participants ranged from 3 to 34.9 months of age (median 7.6 months). The median duration of VGB treatment was 16 months (range 4.6–78.5 months). Electroretinograms (ERGs) were performed according to the Standards of the International Society for Clinical Electrophysiology of Vision. Inclusion required baseline (pre-VGB or within 4 weeks of starting VGB treatment) and at least 2 follow-up ERGs. Significant reduction from baseline of the 30Hz ERG flicker amplitude on 2 consecutive visits identified VGB-RD. Kaplan-Meier survival analyses depicted the effect of duration of VGB on VGB-RD. RESULTS: These data represent the largest survival analysis of children treated with VGB who did not succumb to retinal toxicity during the study. Thirty of the 146 participants (21%) showed VGB-RD. The ERG amplitude reduced with duration of VGB treatment (p = 0.0004) with no recovery after VGB cessation. With 6 and 12 months of VGB treatment, 5.3% and 13.3%, respectively, developed VGB-RD. There was neither effect of age of initiation of VGB treatment nor sex of the child on survival statistics and no significant effect of cumulative dosage on the occurrence of VGB-RD. CONCLUSIONS: Minimizing VGB treatment to 6 months will reduce the prevalence of VGB-RD in patients with IS.

Infantile spasms (West syndrome) is a catastrophic epilepsy affecting infants. Prompt recognition and treatment of infantile spasms (within 3–4 weeks after onset) using effective therapy are essential for good seizure and developmental outcome. Effective treatments include hormonal agents (ACTH or prednisolone) and vigabatrin (1). Vigabatrin (VGB) is particularly effective in children with infantile spasms (IS) due to tuberous sclerosis and preferred over hormonal treatments in this setting (2). Vigabatrin use is associated with risk of developing visual field loss, which first begins as a bilateral nasal wedge-shaped field defect, progressing eventually to bilateral concentric loss of vision. Typically, the affected individual is unaware of the field defect until it becomes severe. Ophthalmologic monitoring is recommended at baseline, every 3 months during treatment, and 3 to 6 months after completing therapy. In adults, the interval should be at least every 6 months (3, 4). It is uncommon for a patient to develop a visual field defect on VGB in less than 3 months; therefore, VGB should be discontinued if there has not been meaningful seizure reduction after 3 months of treatment (3).

Willmore et al. reviewed data from Vigabatrin clinical trials for the treatment of infantile spasms and refractory complex partial seizures in adults. Confirmed visual field defects were noted in 25%–50% of adults and in 15% of children following 9–11 months of treatment (3). Maguire et al. examined the prevalence of visual field defects in 1,678 vigabatrin-treated patients from 32 published studies (included adults and older children) using perimetry. Visual field loss occurred in 44% of adults, 34% of children, and 7% of control subjects. The relative risk for visual field loss was 4.0; larger mean cumulative dose of VGB and increasing age were associated with a higher proportion of visual field loss (4). In a retrospective study of 93 adult patients taking VGB for =6 months, visual field constriction occurred in 52.7% of patients and as early as 1.1 years. There was no correlation of visual field defects with either the maximum or cumulative VGB dosage or duration of exposure. The authors suggested that the retinopathy was an idiosyncratic drug response due to a genetic predisposition (5). Visual field defects also occur in patients with refractory epilepsy not on VGB, especially those taking gabaergic AEDs such as phenobarbital or tiagabine (6).

Visual field testing in young children is often difficult. White Sphere Kinetic Perimetry has been done successfully in developmentally delayed children using a larger object (6⁰ vs 1.7⁰ for Goldman Kinetic Perimetry); 8/28 children or 29% were found to have visual field defects in this study (7). However, 20 to 25 percent of patients are unable to complete perimetry; in these situations, other methods—electroretinogram (ERG) or optical coherence tomography—are useful.

Infants and young children unable to perform perimetry should be assessed by an ERG. In a photopic ERG, the leading edge of the a-wave arises predominantly from cone photoreceptor hyperpolarization to a flash stimulus, while the b-wave peak represents predominantly ON-center bipolar depolarization. With the termination of the light stimulus, the OFF-bipolar cells depolarize in response to photoreceptor glutamate release; this predominant depolarization is the d-wave of the ERG (8, 9). VGB-related visual field loss is associated with abnormalities of cone function manifesting as reduction in the photopic flash b-wave and flicker amplitude. Light-adapted 30Hz flicker ERG showed reduction in the d-wave amplitude in 9/51 children (17.5%) who were taking VGB (9). The 30Hz flicker ERG has been recommended for primary screening with testing at baseline (pre-VGB) or within 4 weeks of starting VGB treatment and at 3-month intervals (10, 11). A different study also using 30Hz flicker response was unable to reproduce these results and recommended caution in use of the ERG for screening (12). The electrooculogram (EOG) of patients treated with VGB records the electrical potential of the retina. A decrease in the Arden ratio—the ratio of the electrical potentials to light and darkness—indicates damage to the retinal pigmentary epithelium (13).

Recently, Westall and colleagues reported their experience in an observational cohort study of 146 children with IS ranging in age from 3 to 35 months (median 7.6 months) who had baseline and at least 2 follow-up ERGs. The light-adapted 2.29 flicker amplitude was taken as the average of the left and right eye responses. The median duration of VGB treatment was 16 months, and retinal damage occurred in 30/146 (21%). After 6, 12, and 30 months of VGB treatment, 5.3%, 13.3%, and 38% of children, respectively, showed evidence of retinal damage by ERG. The median response reduced during the first 2 years and then leveled off, reaching a plateau that persisted even after VGB discontinuation. During the first 12 months of VGB treatment, there was no difference in the cumulative dose between those who developed retinal toxicity and those who did not. After 12 months of treatment, the cumulative dose was higher for those developing retinal toxicity. The authors suggested that VGB treatment for 6 months or less reduced the prevalence of VGB associated retinal damage in children with IS.

The Westall study provides useful guidance to the clinician regarding the duration of VGB treatment in an infant, which should not exceed 6 to 9 months in most cases. VGB should be discontinued at the first indication of toxicity. Ophthalmological monitoring with ERG is helpful to identify toxicity earlier than clinical examination. As the cumulative dose does not appear to be a major risk factor, one should use an adequate dose of VGB.

In the last decade, optical coherence tomography (OCT) has become available for examining the optic nerve and layers of the retina. Thinning of the retinal nerve fiber layer reflects ganglion cell axon loss and correlates strongly with the extent of visual field loss (14). VGB-associated visual loss is irreversible; those with minimal loss may show improvements in ERG amplitude and EOG ratios after stopping VGB, suggesting that electrophysiological changes precede irreversible vision loss.

Vigabatrin is an irreversible inhibitor of GABA transaminase, thereby raising levels of GABA, an inhibitory neurotransmitter. There are three subtypes of GABA receptors: GABA_A, which gates the chloride ionophore and also has binding sites for benzodiazepines and barbiturates. GABA_B receptors belong to the G-protein coupled receptor family and are coupled to calcium and potassium ion channels. GABA_C receptors are concentrated in the retina, especially in the amacrine cells, bipolar cells, photoceptors, Müller and ganglion cells, as well as axon terminals. Compared to GABA_A receptors, GABA_C receptors are 10 times more sensitive to GABA. Vigabatrin accumulates in the amacrine and Müller cells at concentrations 5 to 18 times higher than in the brain, leading to decreased retinal GABA transaminase activity and a fivefold increase in GABA concentrations. Intensely activated, GABA receptors become excitatory, leading to excitotoxic effects caused by osmotic imbalance resulting from influx of chloride, sodium, and water, causing cell lysis and death. GABA may also decrease cerebral and ocular blood flow, possibly exacerbating the neurotoxicity of VGB (15).

Animal experiments show that exposure to bright light damages stressed photoreceptor cells by generating free radicals. VGB-treated albino rats exposed to light experienced retinal damage, whereas rats maintained in darkness had significantly less damage—although this may not necessarily apply to the pigmented human retina (15). Taurine depletion resulting from VGB treatment has been implicated in development of retinal toxicity in rats (16). Taurine has antioxidant and membrane-stabilizing properties and may modulate cellular signaling through protein phosphorylation and calcium uptake. Taurine may also have anticonvulsant properties (15). There are no controlled trials in humans using taurine to prevent VGB-associated retinal damage.