Vigabatrin: 2008 Update

Authors


Address correspondence to L. James Willmore, M.D., Department of Neurology and Psychiatry, Saint Louis University School of Medicine, 1402 South Grand Blvd. (M226), St. Louis, MO 63104, U.S.A.
E-mail: willmore@slu.edu

Summary

Vigabatrin (VGB) is a structural analogue of γ-aminobutyric acid (GABA) that irreversibly inhibits GABA-transaminase (GABA-T), increasing brain levels of GABA. VGB is under assessment for treatment of infantile spasms (IS) and refractory complex partial seizures (CPS). Response can be rapid with spasm cessation following approximately 2 weeks of therapy. Patients with symptomatic tuberous sclerosis (TS) and other patients have achieved spasm cessation. Comparison with ACTH has been performed. Patients with refractory CPS have responded as well. Adverse effects and structural findings on imaging occur with VGB treatment. T2 hyperintensities within brain have been observed. Psychotic disorders or hallucinations have occurred rarely. A specific adverse effects is associated VGB, with a peripheral visual field defect (VFD) detected in some patients. Prevalence and incidence of the VGB-induced peripheral VFD varied depending on the age of the patient and the extent of exposure to VGB, with 25% to 50% prevalence in adults; the prevalence in children was 15% and retinal defect in infants ranged from 15% to 31%. A bilateral nasal defect may be the first clinical indication and may progress to a concentric, bilateral field defect observed in many affected patients; central visual acuity is almost always preserved. The earliest finding of the first abnormal field examination in adults was after 9 months of treatment; with a mean duration of VGB exposure of 4.8 years. In children, the earliest onset of a first abnormal field examination was after 11 months, with a mean time to onset of 5.5 years. The earliest sustained onset of the VGB-induced retinal defect in infants was 3.1 months.

Recommendation: Cognitive, age-appropriate visual field testing is required at baseline and then repeated at intervals in patients who continue therapy. Infants are tested at baseline and at 3-month intervals for the first 18 months of treatment, and then every 6 months thereafter. Adults with CPS are tested at baseline and at 6-month intervals. To select patients who are appropriate for VGB therapy, physicians must consider the benefits of fewer seizures and improved quality of life versus the potential risk of developing a VGB-induced peripheral VFD. Effectiveness of VGB can be detected within 12 weeks of initiating therapy. There appears to be minimal risk associated with a 2- to 3-month trial of VGB to evaluate effectiveness before there is a demonstrable risk of developing the VGB-induced peripheral VFD. If patients do not have a clinical benefit from VGB within 12 weeks of treatment initiation, VGB should be discontinued. If patients have a meaningful reduction in seizures or achieve seizure freedom, then the physician and patient or caregiver must determine if the benefits outweigh the potential risk of developing a peripheral VFD. When VGB is prescribed, the patient must be closely monitored for visual field changes. In cases where spasm or seizure improvement is not achieved within 12 weeks of initiation, VGB should be discontinued. In cases where complete spasm cessation, seizure control, or meaningful improvement is achieved within 12 weeks, continued treatment with VGB is warranted; subsequent periodic monitoring for the peripheral VFD is necessary and should be used to mitigate the risk of the defect. The risk of developing the peripheral VFD with short-term exposure seems to be low, therefore, VGB is an appropriate option for patients with IS or refractory CPS who receive a clinical benefit from its effectiveness, given the clinical consequences of uncontrolled seizures and spasms.

Data have been submitted to the U.S. Food and Drug Administration (FDA) to support the approval of vigabatrin (VGB). An update regarding the benefits and risks associated with VGB is important for potential prescribers to understand the appropriate use of this drug. Such information will help maximize clinical benefits while minimizing potential risks of VGB-induced peripheral visual field defect (peripheral VFD). VGB, a drug specifically designed to treat epilepsy, has been used successfully to treat patients with infantile spasms (IS) and refractory complex partial seizures (CPS) in many countries for almost two decades. Although it is currently available in more than 50 countries worldwide, the FDA withheld approval in the United States pending further research on VGB-induced peripheral VFD. That work recently has been completed, and two new drug applications for VGB have been submitted to the FDA for consideration of approval of use in refractory CPS in adults and IS in children.

Pharmacologic Properties

VGB (4-amino-5-hexenoic acid; γ-vinyl GABA), a structural analog of γ-aminobutyric acid (GABA) with a vinyl appendage designed as an anticonvulsant, irreversibly inhibits the major degradative enzyme for GABA, GABA-transaminase (GABA-T). Increased brain levels of GABA result. VGB also may stimulate GABA release (Ben-Menachem, 2002).

VGB-induced elevation of GABA levels in the brain (>1.8 mmol/kg) has been associated with a 2-fold decrease in seizure frequency (Petroff et al., 1996). Plasma concentration of VGB does not correlate with effectiveness.

Administered orally, VGB is almost completely absorbed, widely distributed, and primarily eliminated through renal excretion (Ben-Menachem, 2002). VGB does not bind to plasma proteins (Ben-Menachem, 2002). VGB is associated with a 16–20% mean reduction in total phenytoin plasma levels (Browne et al., 1987; Ovation Pharmaceuticals, Inc., 1993a). An increase of at least 10% in serum carbamazepine levels has been reported in 70% of patients with epilepsy who received adjunctive VGB therapy (Jedrzejczak et al., 2000); dose adjustment is usually not necessary.

Clinical Trial Data in Infantile Spasms

Chiron et al. (1991) reported that VGB was associated with a ≥ 50% reduction in seizures in 68% of patients with intractable IS. See Table 1 for other studies of VGB in children with IS.

Table 1.   Summary of clinical trials of vigabatrin in infantile spasms
Authors and dateStudy designStudy durationNResponder rate (≥50% seizure reduction)
Patients (%)
Spasm cessation
Patients (%)
Time to spasm cessation (days)
  1. ACTH, adrenocorticotropic hormone; IS, infantile spasms; VGB, vigabatrin; VPA, valproic acid.

Chiron et al. (1991)Refractory IS; open study; VGB add-on (50–200 mg/kg/day)24 weeks706843 
Vigevano and Cilio (1997)Newly diagnosed IS; VGB (100-mg/kg/day) or ACTH (10 IU/day) as first-line monotherapy40 days42 VGB: 48
ACTH: 74
VGB response ≤14 days; ≤3 days for 67% patients
Chiron et al. (1997)Patients with IS due to TS; VGB (150 mg/kg/day) or HC (15 mg/kg/day); crossover between treatment arms after 1 month2 months22 After 1 month:
VGB: 100
HC: 45
p < 0.01
 
Villeneuve et al. (1998)VGB first-line monotherapy (10–150 mg/kg/day)15 days70 543.5 (mean)
Granström et al. (1999)Newly diagnosed IS; VGB as first-line monotherapy (40–150 mg/kg/day); ACTH or VPA as add-on therapy; population-based study between November 1993 and November 1997Treatment duration was for 9 days in the first year; 12–15 days in years 2 and 342 VGB: 26
VGB/ACTH (n = 22): 50
VGB/VPA (n = 4): 25
≤7 for 82% of VGB patients
Cossette et al. (1999)VGB (100–150 mg/kg/day) or ACTH (110 IU/m2/day); parallel group15 days56 VGB: 67
ACTH: 67
 
Appleton et al. (1999)Newly diagnosed IS; VGB (50–150 mg/kg/day) as a first-line monotherapy or placebo5-day treatment phase; 6 months open-label with VGB40 Double-blind phase:
VGB: 35
Placebo: 10
p = 0.063
 
Elterman et al. (2001)
Ovation Pharmaceuticals, Inc. (2005)
New-onset IS; high-dose VGB (100–148 mg/kg/day) vs low-dose VGB (18–36 mg/kg/day); open-label flexible dosing long-term follow-up up to 3 years14–21 days221 Spasm-free for 7 days within the first 14 days:
High dose: 16
Low dose: 7
p = 0.0375
 
Lux et al. (2004)VGB (100 mg/kg/day) or hormone [prednisone 40 mg/day) or tetracosactide (40 IU alternate days)]; parallel group; open study14 days107 VGB: 54
Hormone: 73
 
Lux et al. (2005)VGB or hormone (prednisone or tetracosactide); follow-up at 12–14 months 107 VGB: 76
Hormone: 75
 

Appleton and colleagues used VGB as a first-line monotherapy in a placebo-controlled trial in infants with newly diagnosed IS. When seizures were evaluated over a 24-h period, more VGB-treated patients achieved complete spasm cessation compared with placebo. Spasm frequency was significantly reduced with VGB versus placebo (78% and 26%, respectively, p = 0.02) (Appleton et al., 1999).

The VGB Infantile Spasms Study Group (VISSG) trial treated infants (N = 221) with low-dose VGB (18–36 mg/kg/day) or high-dose VGB (100–148 mg/kg/day) for 14–21 days. Patients could not receive adrenocorticotropic hormone (ACTH), corticosteroids, valproic acid, felbamate, or any investigational drug other than VGB before or during the 14-day randomized phase. The primary efficacy end point was the proportion of patients who achieved spasm cessation for seven consecutive days with onset within the first 14 days of treatment (Elterman et al., 2001; Ovation Pharmaceuticals, Inc., 2005). Because attainment of the primary end point was only possible if spasm cessation by clinical assessment was confirmed by closed circuit television electroencephalography (EEG) within 3 days of the seventh day, the endpoint was especially stringent and logistically difficult to achieve. Many patients who achieved the clinical endpoint were unable to obtain electrocardiography (EEG) within the 3-day period and were thus evaluated as treatment failures. Nevertheless, significantly more patients in the high-dose group met the primary endpoint compared with patients in the low-dose group (16.0% and 7%, respectively, p = 0.0375) (Ovation Pharmaceuticals, Inc., 2005). Patients continued under evaluation during an open-label follow-up of up to 3 years.

Patients responded rapidly to VGB, in that spasm cessation occurred dramatically following approximately 2 weeks of VGB therapy. The stringent 3-day EEG requirement led to an underrepresentation of treatment responders at the primary endpoint. The proportion of responders in the high-dose group increased from 30% at 2 weeks, to 43% at 4 weeks, 51% at 2 months, and 58% at 3 months. In the low-dose group, the proportion of responders was 13%, 21%, 33%, and 44% at the same time points, respectively (Ovation Pharmaceuticals, Inc., 2005). When response rates were evaluated by etiology in both dose groups combined, 74% of symptomatic tuberous sclerosis (TS) patients, 72% of cryptogenic patients, and 50% of symptomatic other patients achieved spasm cessation for seven consecutive days and remained spasm-free for the duration of the study (Ovation Pharmaceuticals, Inc., 2005). Of the 171 patients who attained spasm cessation for 7 days, with or without closed circuit television EEG confirmation, 23% relapsed, but then 72% of those achieved spasm cessation again and 79% maintained spasm cessation for the remainder of their follow-up (Ovation Pharmaceuticals, Inc., 2005). Patients who regained spasm cessation did so with increasing the VGB dose, with administration of one or more concomitant antiepilepsy drugs (AEDs), or with switching to an alternative AED (Ovation Pharmaceuticals, Inc., 2005).

Several comparator studies evaluated VGB against other treatments used for IS (Table 1). In 22 patients with IS due to TS, a greater number of patients were spasm-free after 1 month on VGB (150 mg/kg/day) as compared with treatment with a low dose of hydrocortisone (HC, 15 mg/kg/day; 100% of VGB-treated patients vs. 45% with HC, p < 0.01). All patients who crossed over from HC to VGB achieved spasm cessation; because all patients treated initially with VGB achieved spasm cessation, none of these patients were crossed over to receive HC (Chiron et al., 1997). Vigevano and Cilio reported higher spasm cessation rates initially with ACTH compared with VGB (74% and 48%, respectively); however, after 3 months, 32% of ACTH-treated patients relapsed, compared with 4% of VGB-treated patients (Vigevano & Cilio, 1997). In the United Kingdom Infantile Spasms Study (UKISS), Lux and colleagues administered either VGB or hormone therapy (prednisolone or tetracosactide, a synthetic analogue of ACTH) to children with IS (N = 107); children with TS were excluded from the study (Lux et al., 2004). After 14 days, patients who received hormone therapy had higher rates of spasm cessation compared with VGB (73% and 54%, respectively); however, follow-up evaluation at 12–14 months showed equivalent effectiveness for VGB and hormone treatments (76% and 75%, respectively) (Lux et al., 2005).

Whereas the joint guidelines issued by the American Academy of Neurology and the Child Neurology Society state that VGB is only “possibly effective” (Mackay et al., 2004), the National Institute for Clinical Excellence (2004) and the Scottish Intercollegiate Guidelines Network (2005) note that VGB “is effective” for first-line treatment of IS.

Clinical Trial Data in Refractory CPS

Rimmer and Richens used VGB for the treatment of patients with refractory CPS and reported a ≥50% reduction in seizures in 67% of patients (Rimmer & Richens, 1984). Several double-blind studies conducted by European groups and a group of Canadian investigators found VGB effective in patients with refractory CPS (Table 2). Using responder rates (patients with a ≥50% seizure reduction), two large US studies also reported the effectiveness of VGB as an adjunctive agent for patients with refractory CPS (Table 2).

Table 2.   Summary of double-blind, placebo-controlled add-on trials of vigabatrin in refractory complex partial seizures
Authors and dateStudy designStudy duration (weeks)NResponder rate (≥50% seizure reduction)
Patients (%)
≥75% Seizure reduction
Patients (%)
  1. VGB, vigabatrin.

Rimmer and Richens (1984)Crossover
VGB
3 g/day
9 per arm24VGB: 67
VGB: 3.5 seizures/week
Placebo: 6.2 seizures/week
VGB: 29
Gram et al. (1985)Crossover
VGB
3 g/day
12 per arm21VGB: 44
Median seizure frequency:
Placebo = 16
VGB = 9
p < 0.05
VGB: 17
Loiseau et al. (1986)Crossover
VGB
3 g/day
10 per arm23VGB: 58
Seizure frequency: VGB vs. placebo p < 0.01
 
Tartara et al. (1986)Crossover
VGB
2 or 3 g/day
7 per arm23VGB: 60
Seizure frequency VGB vs. placebo p < 0.01
VGB: 20
Tassinari et al. (1987)Crossover
VGB
2–3 g/day
12 per arm31VGB: 33 
French et al. (1996)Parallel group
VGB 3 g/day
24182VGB: 43
Placebo: 19
p < 0.001
 
Dean et al. (1999)Parallel group
VGB: 1, 3, or 6 g/day
26174VGB:
1 g/day: 24
3 g/day: 51a
6 g/day: 54a
Placebo: 7
aP<0.0001 vs. placebo
 
Bruni et al. (2000)Parallel group36111VGB: 48
Placebo: 26
 

The two US clinical trials enrolled highly refractory patients aged 18–60 years with uncontrolled CPS in an 8-week pretreatment evaluation, followed by 16 or 18 weeks of treatment with VGB or placebo (Ovation Pharmaceuticals, Inc., 1993a, 1993b; French et al., 1996; Dean et al., 1999). The majority of patients (62% and 53% in study 1 and study 2, respectively) were treated with two or more concomitant AEDs at trial entry. Patients had a long history of epilepsy (mean duration, 22 ± 9 and 22 ± 10 years in study 1 and study 2, respectively); 92.4% had used at least two AED classes (including carbamazepine, hydantoins, barbiturates, valproic acid, and benzodiazepines), and 78% had used at least three AED classes, and 51.3% had used at least four AED classes. The primary efficacy endpoint was change in median monthly seizure frequency during the last 8 weeks of the study when compared with baseline.

The first US trial enrolled 182 patients. They found significantly greater reduction in monthly seizure frequency with VGB (3 g/day) compared with placebo (−3.0 and −0.8, respectively, p = 0.0002) (French et al., 1996). VGB treatment also resulted in a greater number of responders (43% with VGB and 19% with placebo, p < 0.001) (French et al., 1996). Response to VGB was rapid: a statistically significant difference in seizure reduction between VGB and placebo was seen on first observation at 14 days of therapy (p = 0.014) and maintained for the duration of the study (Ovation Pharmaceuticals, Inc., 1993a). In addition, 5.4% of patients treated with VGB were seizure-free compared with no seizure-free patients in the placebo group (Ovation Pharmaceuticals, Inc., 1993a).

In the second US study, doses were VGB 1, 3, or 6 g/day or placebo for 174 patients (Ovation Pharmaceuticals, Inc., 1993b; Dean et al., 1999). Significantly greater reductions in monthly seizure frequency were found in the two higher doses of VGB compared with placebo (−4.3 for VGB 3 g/day, −4.5 for VGB 6 g/day, −0.2 for placebo; p = 0.0001 vs. placebo) (Dean et al., 1999). VGB treatment at the two higher doses resulted in a significantly greater number of responders: 24% responders in the VGB 1-g/day group, 51% in the VGB 3-g/day group, 54% in the VGB 6-g/day group, and 7% in the placebo group (p < 0.0001 vs. placebo) (Dean et al., 1999). No significant difference in effectiveness was observed between the VGB 3- and 6-g/day groups.

Response to VGB was rapid: a statistically significant difference in seizure reduction between VGB 3 and 6 g/day versus placebo was observed after 14 days (p = 0.0017) and maintained for the duration of the study (Dean et al., 1999). Seizure freedom throughout the last 8 weeks of the study was achieved by 9.3% and 12.2% of patients in the VGB 3- and 6-g/day groups, respectively, compared with no patients in the VGB 1-g/day and placebo groups (Dean et al., 1999).

Safety of VGB

Clinical trials conducted in Europe, Canada, and the United States reported VGB as generally well tolerated, with adverse events similar to those experienced with other AEDs. However, three VGB-specific side effects have been reported: intramyelinic edema (IME) in animals, psychosis, and VGB-induced peripheral VFD.

Human trials of VGB in the United States were temporarily suspended in 1983 because of the emergence of IME in rodents and dogs given VGB (Krauss & Miller, 1999). Human trials were allowed to resume after autopsy and surgical brain samples demonstrated no evidence of IME, and after magnetic resonance images (MRI) were normal in humans treated with VGB at doses from 1–6 g/day for periods ranging from 3 months to 12 years. In an estimated 350,000 patient-years of VGB exposure (approximately 175,000 patients treated for 2 years at a mean dose of 2 g/day), no definite case of VGB-induced IME has been identified (Cohen et al., 2000). Numerous toxicologists and a European Medicines Agency evaluation concluded that IME is a species-specific adverse effect.

More recently, during a retrospective chart review of 15 patients aged 5 months to 20 years, Pearl and colleagues identified new-onset T2 hyperintensities in three patients on VGB (Pearl et al., 2006). These hyperintensities were localized within the basal ganglia, thalami, anterior commissure, corpus callosum, or midbrain, and resolved following discontinuation of VGB in two patients; the third patient was undergoing VGB taper. Ovation Pharmaceuticals has recently conducted a multicenter retrospective epidemiologic study to better characterize the observed findings.

In US and non-US clinical studies of 4,077 patients treated with VGB, <1% reported either psychotic disorder or hallucinations as an adverse event; no deaths were attributed to psychotic disorders (Ovation Pharmaceuticals, Inc., 2007a). Psychiatric episodes tended to abate with slow introduction of VGB, by reducing dose, or by administering psychotropic drugs (Levinson & Devinsky, 1999; Veggiotti et al., 1999). If psychosis persisted despite these measures, VGB therapy was discontinued because psychosis resolves after VGB discontinuation (Sander et al., 1991; Xavier et al., 2000; Garcia Pastor et al., 2000).

Eke et al. (1997) described VGB-induced peripheral VFD, and the FDA rescinded the approvable letter granted to Hoechst Marion Roussel in 1997.

VGB-Induced Peripheral Visual Field Defect

VGB is associated with a peripheral VFD in some patients. Whereas most patients affected are asymptomatic (Ovation Pharmaceuticals, Inc., 2007b), some patients notice a constriction of their visual fields. VGB has not affected central visual acuity or color vision (Ovation Pharmaceuticals, Inc., 2007b,c).

Forty-six centers in France, South Korea, Italy, Spain, and Australia began enrolling 735 patients in March 1999. To be included, patients had to be at least 8 years old and have refractory CPS for at least 1 year. Patients with known secondary ophthalmic disorders were excluded. Patients were stratified into two age groups (children aged 8–12 years and adults older than 12 years) and three treatment groups: group 1 contained patients currently being treated with VGB, with a mean duration of treatment of 4 years; group 2 contained patients who were previously treated with VGB and had discontinued the drug, with a mean duration of treatment of 1.8 years in children and 2.5 years in adults; group 3 contained patients who had never been treated with VGB but were eligible to be treated. Patients in groups 1 and 2 were to have received VGB for at least 6 months; patients in group 2 were to have discontinued VGB for at least 6 months (Wild et al., 2007). Of 735 patients, 524 met the criteria to be evaluated, including having one or more conclusive perimetry examinations (Ovation Pharmaceuticals, Inc., 2007b). Prevalence and incidence of the VGB-induced peripheral VFD varied depending on the age of the patient and the extent of exposure to VGB (Table 3). In various studies, the prevalence of the VGB-induced peripheral VFD ranged from 25–50% in adults; the prevalence in children was 15%. The prevalence of the VGB-induced retinal defect in infants ranged from 15–31% (Table 3).

Table 3.   Prevalence, incidence, earliest time of onset, and mean time of onset of the VGB-induced peripheral VFD in adults and children with CPS, and in infants with IS
 Study 4020
Final analysisa,b
(n = 524)f
Glasgow Studyc
(N = 105)
Toronto Studyb,d
(N = 197)
Kinirons Studye
(N = 93)
  1. Confirmed visual field defect (VFD): ≥2 abnormal visual fields that were categorized as a bilateral concentric peripheral constriction; sustained abnormality: ≥2 abnormal electroretinography (ERG) findings based upon the 30-Hz flicker amplitude; first conclusive VFD: appearance of the bilateral concentric peripheral constriction at first viable perimetry.

  2. aDeToledo et al. (2006).

  3. bOvation Pharmaceuticals, Inc. (2007b).

  4. cOvation Pharmaceuticals, Inc. (2007c).

  5. dUniversity of Toronto (2007).

  6. eKinirons et al. (2006).

  7. fRepresents the evaluable population from within the locked database of 735 patients.

PrevalenceConfirmed VFD
 Children: 15.3%
 Adults: 24.6%
Adults:
30% to 50%
Infants
Sustained abnormality:
 30-Hz flicker amplitude = 31.1%
 Cone b-wave amplitude = 14.7%
 
Incidence
(number of cases per 100 patient-years)
Confirmed VFD
 Children: 6.1
 Adults: 4.4
 Infants
Sustained abnormality:
 30-Hz flicker amplitude = 15.3
 Cone b-wave amplitude = 6.2
 
Earliest time of onsetFirst conclusive VFD
 Children:
 11 months
 Adults:
  9 months
 Infants
Sustained abnormality
 9.9 months
Adults
 1.1 years

A bilateral nasal defect may be the first clinical indication of a VGB-induced retinal defect (Ovation Pharmaceuticals, Inc., 2007b,c). This nasal wedge may progress to the concentric, bilateral field defect observed in many affected patients; central visual acuity is almost always preserved. The magnitude of the bilateral field defect is similar between eyes. Reported onset of the VGB-induced peripheral VFD varied among studies (Table 3. The earliest finding of the first abnormal field examination in adults was after 9 months of treatment; with a mean duration of VGB exposure of 4.8 years. In children, the earliest onset of a first abnormal field examination was after 11 months, with a mean time to onset of 5.5 years (Ovation Pharmaceuticals, Inc., 2007b). In a longitudinal study of children, the earliest sustained onset of the VGB-induced retinal defect in infants was 3.1 months. In infants, sustained onset, defined as two consecutive confirmed electroretinography (ERG) abnormalities, is the requisite measure of a true defect because many infants have abnormal values at baseline that may normalize spontaneously, with or without VGB therapy. Incidence of the VGB-induced retinal defect in the Toronto study was 15.3 cases/100 patient-years, as determined by the 30-Hz flicker amplitude, the most sensitive ERG parameter (Ovation Pharmaceuticals, Inc., 2007b; University of Toronto, 2007). A more rigorous definition requires both abnormal 30-Hz flicker amplitude and abnormal cone B wave amplitudes, as in the adult patients with bilateral concentric peripheral constriction who had simultaneous field and ERG measurements (Harding et al., 2000).

Visual fields in 358 patients were assessed by Goldmann perimetry. Patients exposed to VGB retained a median of 71° in the lateral visual field, with 90° being normal (Fig. 1) (Ovation Pharmaceuticals, Inc., 2007b). Final analysis reported 129 patients with a first conclusive VFD (i.e., the appearance of the bilateral concentric peripheral constriction at the first viable perimetry); 72% (93 of 129) of these patients had a confirmed VFD (≥2 abnormal visual fields that were categorized as a bilateral concentric peripheral constriction). Most patients with the VGB-induced peripheral VFD confirmed by perimetry are asymptomatic (Krauss et al., 1998; Wild et al., 1999; Mauri-Llerda et al., 2000), with a few experiencing symptoms sufficiently severe to hinder daily activities. One patient had severe central vision restriction (15 degrees visual field) 4 years after discontinuing VGB therapy (Ovation Pharmaceuticals, Inc., 2007b).

Figure 1.

 (A) The percentage of patients (N = 275) retaining various degrees (15° to 90°) of lateral vision in the temporal field at final Goldmann perimetry in the Aventis 4020 study. Right and left eye degrees were averaged. Masked ophthalmologists conducted the Goldmann visual field perimetry (Ovation Pharmaceuticals, Inc., 2007b). (B) Patients exposed to vigabatrin experienced a mean reduction in their visual field to 71°(normal = 90°).

Progression of the VGB-induced peripheral VFD while on therapy is minimal, as determined by a mean 2-year period from first to last Goldmann perimetry examination (Ovation Pharmaceuticals, Inc., 2007b). While exposed to VGB, patients experienced a loss of <2 degrees per year on average from the temporal visual field (depending on cohort), except for the four children who had been exposed for >7.5 years, for whom the decrease was 11 degrees per year on average after the 7.5 years. The decreases in the nasal field were <1 degree per year in all subject cohorts, including the four children (Ovation Pharmaceuticals, Inc., 2007b). Another study found that established field defects resulting from VGB showed no significant progression with continued use of the drug in 94% of patients exposed to VGB for 5 years or more (range, 5–12 years) (Best & Acheson, 2005). Relatively slow progression of peripheral visual loss and the moderate degree of impairment on lateral vision (approximately 25°) may explain why patients are asymptomatic. In addition, eye movement and head turning are used unconsciously by patients to compensate for peripheral field loss and may affect perimetry performance.

Although some publications have reported reversibility of visual field loss after discontinuing VGB treatment (Versino & Veggiotti, 1999; Fledelius, 2003; Westall et al., 2003; Kinirons, et al., 2006), nearly all others indicate that the VGB-induced peripheral VFD appears to be irreversible (Hardus et al, 2000; Nousiainen et al., 2001). In a study of 60 adults with partial epilepsy treated with VGB from 7 months to 14 years, 24 patients had VGB-induced peripheral VFD at initial evaluation; however, total duration of VGB treatment was not reported. A follow-up examination was performed after 4–38 months (mean, 15 months) in 55 patients, 29 of whom had discontinued VGB therapy; no significant recovery was observed in visual fields, and there was no progression with continued therapy (Nousiainen et al., 2001, Ovation Pharmaceuticals, Inc., 2007b).

Although not well established, smoking and male gender may increase the risk of developing VGB-induced peripheral VFD (Wild et al., 1999; Kalviainen & Nousiainen, 2001; Ovation Pharmaceuticals, Inc., 2007b). Associations between dose and duration of exposure to VGB and risk of developing the peripheral VFD are contradictory. Reports range from no associated risk (Harding et al., 2000) to increasing risk with cumulative exposure, with incidence increasing rapidly during the first 2 years of VGB treatment and within the first 2 kg of VGB intake, and stabilizing at 3 years and after a total VGB dose of 3 kg (Kalviainen & Nousiainen, 2001).

Testing for the VGB-Induced Peripheral VFD

Perimetry by static or kinetic methods, specialized visual-evoked potentials (VEPs), and ERG testing of vision in adults and children have been used to diagnose VGB-induced peripheral VFD (Table 4). Baseline studies must be obtained by the best age-specific method available before starting the drug.

Table 4.   Testing options for peripheral visual field defect
Infants and older8 years and oldera12 years and oldera12 years and oldera
  1. aAge ranges are estimates. Cognitive abilities should be taken into consideration.

Electroretinography (ERG)Confrontation testingKinetic perimetryAutomated perimetry
The only testing option appropriate for infants
Uses an electrode placed on the eye to monitor electrical activity in the retina in response to a flash of light
An objective testing option that usually requires sedation in infants and young children
Examiner moves hand through patient’s peripheral vision to roughly determine the border of the visual field
Widely available and easily performed, but not a sensitive test
Examiner moves stimuli through patient’s peripheral vision and maps visual field defects on a reference grid (Goldmann perimetry testing is an example of kinetic perimetry)
More sensitive than confrontation testing, but also more challenging to perform and less widely available
Uses automated visual field analyzers, such as Octopus and Humphrey visual field analyzers, to project light stimuli, measure reaction, and map the patient’s visual field
An objective testing option

Studies using all three modalities in individual patients have helped define the VEP and ERG parameters that correspond to the VGB-induced peripheral VFD. Cognitive, age-appropriate visual field testing is recommended at baseline and then repeated at intervals in patients who continue therapy. It is recommended that infants should be tested at baseline and at 3-month intervals for the first 18 months of treatment, and then every 6 months thereafter (Ovation Pharmaceuticals, Inc., 2007b). Adults with CPS should be tested at baseline and at 6-month intervals (Ovation Pharmaceuticals, Inc., 2007b). If a change is detected, confirmatory testing should take place within 1 month and results should be discussed with the patient or caregivers.

Perimetry testing with either static or kinetic procedures (suitable for patients aged >9 years) is sufficiently sensitive and specific to establish a baseline and monitor peripheral vision. There are, however, inherent difficulties in performing perimetry tests in children aged ≤9 years and in cognitively impaired patients (Kalviainen & Nousiainen, 2001).

Field-specific VEP is well tolerated, sensitive, and specific for identifying the VGB-induced retinal defect in children aged >2 years (Harding et al., 2002; Spencer & Harding, 2003). It is possible to detect the VGB-induced retinal defect in young infants and developmentally delayed children with techniques such as full-field or multifocal ERG. Although ERG is suitable in young or developmentally delayed patients, the technique usually requires sedation in the youngest patients.

Emerging techniques such as wide-field, multifocal ERG are highly sensitive and more specific than either static or kinetic perimetry (Harding et al., 2000; Ponjavic & Andreasson, 2001; Gonzalez et al., 2004).

VFD requires many months to years of VGB treatment before it develops, is usually asymptomatic, can be monitored effectively, and does not affect central vision. With proper understanding of the prevalence, incidence, timing of onset, detection, monitoring, and severity of the VGB-induced peripheral VFD, patients can be effectively monitored for its development.

Patient Selection and Benefit/Risk Assessment

In patients with IS and refractory CPS, VGB reduces seizure frequency and provides the opportunity for some patients to become seizure-free. In addition, VGB is generally well tolerated and has been an important treatment in many parts of the world for almost two decades.

To select patients who are appropriate for VGB therapy, physicians must consider the benefits of fewer seizures and improved quality of life versus the potential risk of developing a VGB-induced peripheral VFD. Effectiveness of VGB is well established in both adults and infants within 12 weeks of initiating therapy (French et al., 1996; Dean et al., 1999; Elterman et al., 2001). Because the peripheral VFD was not detected until after 9 months of therapy in adults, after 11 months of therapy in children with CPS, and via abnormal ERG findings at 3 months of therapy in infants (Table 3), the effectiveness of VGB could be evaluated with less risk in individual patients before the earliest onset of peripheral VFD.

There appears to be minimal risk associated with a 2- to 3-month trial of VGB to evaluate the effectiveness before there is a demonstrable risk of developing the VGB-induced peripheral VFD. If patients do not have a clinical benefit from VGB within 12 weeks of treatment initiation, VGB should be discontinued. If patients have a meaningful reduction in seizures or achieve seizure freedom, then the physician and patient or caregiver must determine if the benefits outweigh the potential risk of developing a peripheral VFD. When VGB is prescribed, the patient must be closely monitored for visual field changes. The risk of developing the peripheral VFD may be mitigated by monitoring because discontinuing the drug likely eliminates the probability of progression, and the decision to discontinue treatment can occur at any time. In accordance with FDA requirements, risk minimization tools for VGB are being developed, and when approved will further guide therapy recommendations.

Conclusions

Uncontrolled epilepsy can cause severe problems, including bodily and brain injury, diminished quality of life, reduced social interaction (Sperling, 2004), and intellectual decline in children (Bjornaes et al., 2001). Patients newly diagnosed with epilepsy have a 42% higher mortality rate compared with those who do not have epilepsy (Mohanraj et al., 2006). In general, mortality rates are four to seven times higher (Sperling, 2004), and the risk of sudden unexpected death is increased in people with uncontrolled seizures compared with those whose seizures are controlled (Nashef et al., 1995). To reduce the burden of epilepsy on patients and society, effective and safe treatments are essential to help patients with medically intractable epilepsies attain a reduction in seizures and potentially become seizure-free.

VGB can reduce or eliminate seizures in patients with IS and CPS and is generally well tolerated. Clinical trials involving thousands of patients show that VGB may provide rapid reductions in seizure frequency and has allowed some patients to become seizure-free.

The data regarding the VGB-induced peripheral VFD will enable physicians to better determine appropriate patients for VGB therapy and to assess the associated benefits and risks. Because of the rapid onset of effect of VGB, physicians have the ability to evaluate on a case-by-case basis the seizure improvement derived from VGB treatment before there is a major risk for developing the peripheral VFD. In cases where spasm or seizure improvement is not achieved within 12 weeks of initiation, VGB should be discontinued. In cases where complete spasm cessation, seizure control, or meaningful improvement is achieved within 12 weeks, continued treatment with VGB is warranted; subsequent periodic monitoring for the peripheral VFD is necessary and should be used to mitigate the risk of the defect. The risk of developing the peripheral VFD with short-term exposure seems to be low; therefore, VGB is an appropriate option for patients with IS or refractory CPS who receive a clinical benefit from its effectiveness, given the clinical consequences of uncontrolled seizures and spasms.

Acknowledgments

We thank Ovation Pharmaceuticals, Inc., for permitting us to review their clinical trial data pertaining to the VGB-induced peripheral VFD and for sponsoring the development of this update. We also thank Elaine Gilhooly, PhD, who compiled all available clinical trial data to prepare an initial working draft of this manuscript; that document provided a working framework from which the authors selected salient data and developed the present update. We confirm that we have read the Journal’s position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.

Disclosure of conflicts of interest: LJW has received research support from Abbott Laboratories, Athena Pharmaceuticals, Dianippon Pharmaceutical, Ortho-McNeil Pharmaceuticals, National Institutes of Health, Novartis, Parke-Davis, GlaxoSmithKline, UCB Pharmaceuticals, and R.W. Johnson Pharmaceuticals; and participates in speaker bureaus for Abbott Laboratories, Elan Pharmaceuticals, GlaxoSmithKline, Novartis, Ortho-McNeil Pharmaceuticals, Parke-Davis, and Pfizer. In addition, LJW is a consultant for Abbott Laboratories, Alza, Ovation Pharmaceuticals, and K-V Pharmaceuticals. MBA has received honoraria from and is a consultant for Ovation Pharmaceuticals. EBM has received honoraria for participating in advisory boards convened by Ovation Pharmaceuticals, Janssen-Cilag, UCB Pharmaceuticals, Bial, and sanofi-aventis, and has research projects funded by UCB Pharmaceuticals, Cyberonics, Pfizer, Janssen-Cilag, and Eisai. In addition, EBM is a consultant for sanofi-aventis. JMP has received honoraria or grants from the National Institutes of Health/National Institute of Neurological Disorders and Stroke, Abbott Laboratories, AstraZeneca, Aventis, Cephalon, Eisai, GlaxoSmithKline, Jazz Pharmaceuticals, King Pharmaceuticals, KV Pharmaceuticals, Marinus Pharmaceuticals, MedPointe, NeuroPace, Novartis, Ortho-McNeil, Johnson & Johnson, Ovation Pharmaceuticals, Pfizer, Questcor, Schwartz Pharma, UCB Pharmaceuticals, and Valeant; all grants, research support, consultant fees and honoraria were paid to Virginia Commonwealth University or to the Physician Practice Plan (MCV Physicians). WDS is a consultant for Ovation Pharmaceuticals.

Ancillary