Abstract
Objective: To determine whether there is a causal link between vigabatrin treatment and concentric visual field defects and to evaluate the prevalence of these visual field constrictions.
Background: While the GABAergic antiepileptic drug (AED) vigabatrin was being clinically developed, only rare cases (less than 1:1000) of symptomatic visual field constriction and retinal disorders were reported. During 1997 to 1998, concentric visual field constrictions were described in case reports of mostly drug-resistant epilepsy patients receiving vigabatrin concurrently with other AEDs.
Methods: Ophthalmologic tests including Goldmann perimetry were performed on 32 adult patients on long-term successful vigabatrin monotherapy (treatment duration 29 to 119 months) and on 18 patients on carbamazepine monotherapy (treatment duration 32 to 108 months). Eighteen healthy adults served as controls.
Results: None of the patients complained about vision problems when asked to participate into the study. Thirteen out of the 32 (40%) epilepsy patients treated with vigabatrin monotherapy had concentrically constricted visual fields (9% severely, 31% mildly constricted), whereas none of the carbamazepine monotherapy patients or normal controls presented with a visual field defect (chi-square test, p = 0.0001). The extents of the visual fields were significantly constricted in vigabatrin group as compared with the visual fields of the patients in carbamazepine group or healthy controls (analysis of variance, Scheffé F-test, significant at 99%).
Conclusions: The use of vigabatrin seems to increase the risk of a unique and specific pattern of bilateral, mainly asymptomatic visual field constriction. This risk should be considered when using vigabatrin. Visual field testing should also be performed before treatment and during routine follow-up for patients on vigabatrin.
Vigabatrin is a selective irreversible inhibitor of GABA-transaminase,1 inhibition of which produces greater available pools of presynaptic GABA for release in CNS synapses. Currently, vigabatrin is approved in over 60 countries for the adjunctive management of partial epilepsy, which is not satisfactorily controlled by conventional therapy, and for initial monotherapy in the management of infantile spasms.
During the development of the GABAergic antiepileptic drug (AED) vigabatrin, only rare cases (less than 1:1000) of symptomatic visual field constriction and retinal disorders have been reported. During 1997 to 1998, symptomatic and later also asymptomatic visual field constrictions were described in 13 case reports in patients with drug-resistant epilepsy who were receiving vigabatrin concurrently with other AEDs.2-7 More recently, three add-on therapy patients with visual field defects were reported in which electroretinogram demonstrated evidence of bilateral retinal dysfunction.8 Oscillatory potential responses were lost in these patients, suggesting impairment of the highly GABAergic amacrine cells in the retina. Most patients studied have had refractory epilepsy and also used other AEDs, and a causal relationship between the visual field defects and vigabatrin treatment has not yet been established.9
We randomized our newly diagnosed partial epilepsy patients into vigabatrin or carbamazepine monotherapy during 1988 to 1995.10,11 The objective of this study was to determine whether there was a causal link between vigabatrin treatment and concentric visual field defects and also to evaluate the prevalence of these visual field constrictions in our monotherapy patients.
Patients and methods.
Patients.
A total of 135 patients with newly diagnosed partial epilepsy was randomized to either vigabatrin or carbamazepine monotherapy10,11 groups in our clinic during 1988 to 1995. For the present cross-sectional study, we identified 32 patients with vigabatrin therapy and 18 patients with carbamazepine therapy still successfully using their study treatment as monotherapy and attending our clinic. The majority of vigabatrin patients, i.e., 27 out of 32, had vigabatrin as their initial treatment, but we also investigated 5 patients who had started on carbamazepine but had developed a hypersensitivity reaction to it during the titration phases 5 to 9 years earlier, at the time of their epilepsy diagnosis, and were prescribed vigabatrin as their second monotherapy. The majority of the carbamazepine patients (17 out of 18) underwent treatment as initial treatment but we also studied 1 patient who did not respond to vigabatrin monotherapy during the titration phase (1.5 month with vigabatrin 8 years earlier) and was switched to carbamazepine as her second monotherapy.
Ophthalmologic examination.
A full ophthalmologic evaluation was performed to exclude other ophthalmologic disorders. The ophthalmologic examination consisted of a history of visual symptoms, testing the best-corrected visual acuity, phorias and tropias, biomicroscopy, and direct and indirect ophthalmoscopy with the aid of tropicamid 0.8% + phenylephrine 5% drops. Intraocular tension was measured with the Goldmann applanation tonometer.
Kinetic perimetry was performed with the manually operated Goldmann perimeter (Haag-Streit, Bern, Switzerland) using standardized objects I/3 and IV/4 with the background luminance of 31.5 apostilbs.12 An appropriate refractive correction was used for examining the central field of vision. The visual field examinations were performed with normal-sized pupils, i.e., without pharmacologically induced mydriasis or miosis. All the visual field examinations were performed by the same ophthalmologist (I.N.), who also calibrated the perimeter herself. The perimetry results were interpreted by the examiner together with another ophthalmologist (M.M.). The visual field testings and the neuro-ophthalmologic examinations were verified in eight cases (two severely abnormal cases, five mildly abnormal cases, and one normal case) by a neuro-ophthalmologist (E.N.) in another university hospital. All the ophthalmologic evaluations were performed blindly without knowledge of the clinical details (e.g., the AED regimen used).
Additional examinations.
Electroretinography and an MRI were performed on patients who had abnormal visual fields to further evaluate the pathogenesis of the defects. Electroretinography was performed with a Nicolet Ganzfield stimulator according to the recommended standards,13 and patients were scanned with a 1.5 T Magnetom Vision (Siemens, Erlangen, Germany), special attention being paid to orbits, visual tracts, and occipital lobes.
Statistical analysis.
Visual fields were classified into three categories: 1) normal, if the fields extended over 70 degrees in the temporal meridian; 2) mildly abnormal, if the fields extended from 50 degrees to 70 degrees in the temporal meridian; and 3) severely abnormal, if the fields extended below 50 degrees in the temporal meridian. Statistical analysis was performed with the SPSS (Chicago, IL) program using the chi-square test, and with StatView (Berkeley, CA) 512+ program using one-way analysis of variance (ANOVA) with the Scheffé F-test and a linear regression test. p Values of 0.01 or less were considered statistically significant.
Approval of the ethics committee.
The study was approved by the Ethics Committee of the University of Kuopio, and informed written consent was obtained from all the patients.
Results.
Demographic and clinical data.
The ages of the patients and controls did not differ (ANOVA, p = 0.7719) (mean age of the vigabatrin group, 40.4 ± 16.3 years; range, 19 to 73 years; mean age of the carbamazepine group, 43.8 ± 18.1; range, 20 to 70 years; and mean age of the controls, 42.7 ± 15.7, range, 23 to 74 years). The male/female ratio was 12/20 in the vigabatrin group, 6/12 in the carbamazepine group, and 7/11 in the control group. All patients had well-controlled partial epilepsy and had been seizure free for at least one year. On the basis of the history, clinical examination, and previous brain CT/MRI scans, the etiology of epilepsy was cryptogenic in 22 of the 32 vigabatrin patients (69%) and in 15 of the 18 carbamazepine patients (83%). The duration of vigabatrin therapy varied from 29 to 119 months (mean 68.7 ± 29.5) and that of carbamazepine therapy from 32 to 108 months (mean 60.5 ± 26.2).
Ophthalmologic investigations.
None of the patients complained about visual problems when asked at the beginning of the ophthalmologic investigation. The visual acuities were normal in both patient groups and in controls. Two right eyes in the vigabatrin group, one aphacic and one microphthalmic, were excluded from the study. There were no significant abnormal ophthalmologic findings.
Visual fields.
Of the 32 patients treated with vigabatrin monotherapy, 3 (9%) had severely concentrically constricted visual fields and 10 (31%) had mildly constricted visual fields, whereas none of the carbamazepine monotherapy patients or normal controls presented with visual field defects (figure 1) (chi-square test, p = 0.0001). The visual field tests were verified in eight cases (two severely abnormal cases, five mildly abnormal cases, and one normal case) in another university hospital by a neuro-ophthalmologist who was able to confirm our results.
Figure 1. Goldmann fields of right eye in a patient in the carbamazepine group (A) and in the most severely affected patient in the vigabatrin group (B).
Twelve of the 13 patients with abnormal fields had used vigabatrin as their initial AED, and one used it as a second AED. The ages of the patients with abnormal fields varied between 22 and 72 years. The male/female ratio in abnormal vigabatrin cases was 6/7 and did not differ statistically from that in normal vigabatrin cases (chi-square test, p = 0.4029). Eight of the patients with abnormal visual fields had cryptogenic epilepsy, and 5 had remote symptomatic epilepsy (2 had parietal dysplasia, 2 had temporal contusions, and 1 had temporal infarction). There was no statistical difference in the proportion of patients with cryptogenic epilepsy between abnormal and normal vigabatrin cases (chi-square test, p = 0.4661).
After the visual field investigation, patients with abnormal fields were again asked about symptoms, but only one (a vigabatrin patient whose visual field is shown in figure 1) reported any: This patient reported that she had been bumping into people in shops and kicking her cat by accident lately but had thought these to be signs of clumsiness. Earlier during our vigabatrin studies we had noticed an increased frequency of photopsy13 in vigabatrin-treated patients. Three of the abnormal cases (23%) and three of the normal cases (16%) had complained of intermittent flashes of bright light or scintillations in the peripheral visual fields at some stage while taking vigabatrin (chi-square test, p = 0.6039). Three other patients with abnormal fields had experienced short-lasting blurred vision years earlier while taking vigabatrin. However, there was no difference in the presence or absence of earlier visual symptoms between vigabatrin patients with abnormal and those with normal fields (chi-square test, p = 0.0606).
The table shows the extent of temporal, nasal, lower, and upper visual fields for both drug groups and for healthy control subjects in both eyes. In the right eyes, ANOVA showed a significant group effect for temporal extent (F = 28.1, df = 2,63, p = 0.0001), lower extent (F = 18.6, df = 2,63, p = 0.0001), nasal extent (F = 13.8, df = 2,63, p = 0.0001), and superior extent (F = 12.2, df = 2,63, p = 0.0001). In the left eyes, ANOVA showed a significant group effect for temporal extent (F = 29.9, df = 2,65, p = 0.0001), lower extent (F = 13.5, df = 2,65, p = 0.0001), nasal extent (F = 16.6, df = 2,65, p = 0.0001), and superior extent (F = 8.5, df = 2,65, p = 0.0005). Post hoc analyses for both eyes with the Scheffé F-test showed at a significance level of 0.01 that the temporal, lower, and nasal extents were significantly constricted in the vigabatrin group as compared with the carbamazepine group and with controls and that the superior extents were significantly constricted both in the vigabatrin group and the carbamazepine group compared with controls. ⇓
Extent of the visual fields in controls and in patients with vigabatrin and carbamazepine monotherapy
There was no statistically significant correlation between the different extents of the visual fields and the duration of vigabatrin therapy (figure 2), the vigabatrin dosage (e.g., temporal extent: right eye, R = 0.291, p = 0.1183; left eye, R = 0.203, p = 0.2645), or the cumulative amount of vigabatrin consumed during the treatment (e.g., temporal extent: right eye, R = 0.409, p = 0.0248; left eye, R = 0.178, p = 0.3300).
Figure 2. Correlation between the duration of vigabatrin therapy and temporal visual field extents in left and right eyes by linear regression.
Additional investigations.
All the patients with severe visual field defects and 60% of those with mild visual field defects were found to have reduced oscillatory potentials in electroretinograms. Patients with severe visual field defects also had reduced amplitude of a and b waves in cone photopic and rod scotopic electroretinograms. All MRI of orbits, visual tracts, and occipital lobes were normal and did not reveal any new data regarding the etiology of the visual field defects.
Discussion.
Altogether 40% of the epilepsy patients treated with vigabatrin monotherapy were found to have asymptomatic visual field defects, whereas none of the carbamazepine monotherapy patients or normal controls had any visual field defects. The extents of the visual fields were significantly constricted in vigabatrin group as compared with the visual fields of the patients in carbamazepine group or healthy controls.
All of the abnormal visual fields were concentrically constricted, and none of the patients had local (e.g., hemianoptic, quadrantanoptic, or glaucomatous) defects. The investigation of the optic nerve heads and retinas did not show any explanation for the abnormal visual fields. Neither the history, clinical investigations, nor previous CT or MRI scans provided an explanation for the visual field defects.
Vigabatrin has been shown to cause microvacuolation in white matter tracts in rodents and dogs, but not in monkeys.14,15 The optic tract was the only consistent area of vigabatrin-induced white matter vacuolization in all sensitive species. In these animals, microvacuolation is associated with functional changes in afferent conduction in the visual tracts demonstrable by prolongation of the central conduction times of visual evoked potentials16 and increased T2-weighted signal intensity in quantitative MRIs.17 No myelin microvacuolation has been observed in human neuropathologic studies from autopsy and epilepsy surgery cases.18 We have also earlier shown in our vigabatrin monotherapy patients that longitudinal visual evoked potential studies do not show increased latencies,10,19 nor do qualitative or quantitative MRI studies show an increase in signal intensity.11 In the present study, the MRI of orbits, visual tracts, and occipital lobes did not show any new signal changes.
Retinal degeneration has been shown to occur in albino, but not in pigmented, rats treated with vigabatrin.14 In another study with albino rats, retinal degeneration was found to depend on the light exposure, and it was concluded that vigabatrin increases light exposure or sensitizes the already overly sensitive albino retina to light.15 Retinal degeneration has not been observed in dogs or monkeys.15 However, the abnormal electroretinograms in patients in the present study and in earlier case reports2,7-8 indicate retinal dysfunction as the pathophysiology underlying the visual field constriction. The precise nature of this retinal abnormality is not yet known. If changing the persistence of GABA in retinal cells does turn out to be critical, all AEDs using this mode of action should cause similar changes, and the effects of other new AEDs that mediate their action via interactions with the GABA system, such as gabapentin, tiagabine, and topiramate, should also be evaluated.
Kinetic perimetry with the Goldmann perimeter was chosen to test the full limit of peripheral vision,12 as the earlier case reports had described concentrically constricted fields in patients taking vigabatrin. In the present study also, all abnormal cases had concentrically constricted peripheral fields, without any impairment of the central field. To minimize variability on the part of the examiners, the same ophthalmologist calibrated the Goldmann perimeter and investigated all the patients. Perimetry relies on subjective patient responses, and the results depend on the patient’s psychomotor as well as visual status. In earlier investigations of vigabatrin visual field defects in drug-resistant epilepsy patients, the outcome may have been influenced by fatigue and psychomotor slowing caused by the high frequency of seizures, underlying severe brain pathologies, and AED polytherapy.20 In the present study, our patients were seizure free and had normal general intelligence, verified by neuropsychological testing.10,11
Although the patients did not complain of visual symptoms before testing, the “severe” visual field defects that we observed are of clinical relevance, as they might compromise the patient’s ability to drive a car, for example. Moreover, it is currently unclear whether the defect is progressive or whether it will subside when the vigabatrin treatment is discontinued. It is therefore important to register mild asymptomatic cases as well, as early as possible.
The current study shows that there is a causal relationship between vigabatrin treatment and a specific pattern of bilateral visual field constriction. Although we lack pretreatment perimetries, our patient population provides unique ways of eliminating many of the confounders of previous investigations reporting visual field defects associated with chronic vigabatrin therapy. We could not find any clear risk factors for the vigabatrin-induced visual field defects, but larger series are needed to evaluate the role of duration of treatment, cumulative vigabatrin dosage, age, gender, and other predictive factors or prodromal symptoms. This retinal toxicity presented as an all or none response might also be genetically based, in which case similar studies should be conducted in other populations with different gene pools.
- Received March 31, 1999.
- Accepted June 3, 1999.
References
Disputes & Debates: Rapid online correspondence
REQUIREMENTS
If you are uploading a letter concerning an article:
You must have updated your disclosures within six months: http://submit.neurology.org
Your co-authors must send a completed Publishing Agreement Form to Neurology Staff (not necessary for the lead/corresponding author as the form below will suffice) before you upload your comment.
If you are responding to a comment that was written about an article you originally authored:
You (and co-authors) do not need to fill out forms or check disclosures as author forms are still valid
and apply to letter.
Submission specifications:
- Submissions must be < 200 words with < 5 references. Reference 1 must be the article on which you are commenting.
- Submissions should not have more than 5 authors. (Exception: original author replies can include all original authors of the article)
- Submit only on articles published within 6 months of issue date.
- Do not be redundant. Read any comments already posted on the article prior to submission.
- Submitted comments are subject to editing and editor review prior to posting.