Intended for healthcare professionals


Visual field defect associated with vigabatrin: observational cohort study

BMJ 1999; 319 doi: (Published 30 October 1999) Cite this as: BMJ 1999;319:1165
  1. L V Wilton, senior research fellow (admin{at},
  2. M D B Stephens, research associate,
  3. R D Mann, Director
  1. Drug Safety Research Unit, Southampton SO31 1AA
  1. Correspondence to: L V Wilton

    Since Eke et al's report of persistent visual field constriction associated with vigabatrin,1 several letters have been published in the BMJ, including that by Mackenzie and Klistorner,2 that draw attention to the occurrence of visual field changes in asymptomatic patients as well as those with symptoms.

    Subjects, methods, and results

    A prescription event monitoring study of vigabatrin conducted between 1991 and 1994 identified four cases of bilateral, persistent visual field defect for which there was objective evidence and no alternative cause given.3 The study cohort consisted of 10 033 living patients for whom questionnaires containing clinical information were returned (table) The mean duration of treatment for these four cases was 19.8 months; for the total cohort it was 14.2 months Subsequently, we were informed of another seven patients in this cohort in whom visual field defect had been reported after the end of the study's observation period. This suggested that further cases might be identified if the period of observation was extended.4

    Cases of visual field defect attributed to treatment with vigabatrin in prescription event monitoring study 1991-4 and long term follow up study 1998-9

    View this table:

    The 7228 patients who were still taking vigabatrin at the end of the study were followed up by sending a simple questionnaire to their general practitioner to ask if vigabatrin treatment was continuing and whether any serious adverse events or changes in vision had been reported since the questionnaire had been returned. If patients had been referred for visual problems, the ophthalmologist was asked to complete a questionnaire giving details of visual field testing before and during treatment with vigabatrin, to comment on the possible cause of the defect, and to send copies of test results.

    The results of both studies are shown in the table. Altogether 2068 of the returned questionnaires were not considered further because they did not contain clinical data (1735) or the patients had died (333) In the surviving 4741 patients, 89 reports were followed up with ophthalmologists.

    The ophthalmologists consider that 30 (83%) of the 36 cases for which there is objective evidence of visual field defect are probably or possibly related to vigabatrin. In four cases they did not know whether the defect was related to vigabatrin treatment, and in the remaining two cases the defect was considered not to be due to vigabatrin. The mean age at the time of diagnosis for the 19 men was 45.7 years (range 20-64), and for the 11 women it was 37.6 years (range 20-56). The mean duration of treatment to time of diagnosis in these 30 cases was 66.3 months (range 33-96).

    In this second study information was requested on events reported after the original questionnaire was returned. One of the original four cases is also included in the 30 cases of the current study because visual field tests were carried out after the original questionnaire had been returned. In these two studies, 33 cases of visual field defect considered by ophthalmologists to be probably or possibly related to vigabatrin were identified, giving an incidence risk of 7.0 per 1000 patients (33/4741). The minimal risk, based only on cases considered probably due to vigabatrin, is 3.4 per 1000 patients (16/4741).


    Although publicity bias may have influenced the number of patients referred for visual field tests since the original study,14 this substantial increase in the number of confirmed cases of visual field defect attributed to vigabatrin shows the clinical importance of this lesion in the long term use of vigabatrin. We have also shown that signals of long latency adverse drug reactions identified in a prescription event monitoring study can be successfully investigated in a long term follow up study.


    We thank all the general practitioners and hospital doctors who have taken part in these two studies. We also thank Jackie Barfoot for her help in coordinating the study.

    Contributors: LVW participated in the study design and the follow up questionnaire, was responsible for the execution and coordination of the study, analysis and interpretation of data, and writing the short report. MDBS identified and validated the original cases, participated in formulating the design of the follow up study and the questionnaire, and contributed to writing the paper. RDM participated in study design and interpretation of results and edited the short report. RDM will act as guarantor.


    • Funding None.

    • Competing interests None declared.


    1. 1.
    2. 2.
    3. 3.
    4. 4.
    View Abstract