There is no proved treatment for non-arteritic anterior ischaemic optic neuropathy (NAION). Topical brimonidine tartrate has been reported to have a neuroprotective benefit for retinal ganglion cells following experimental elevation of intraocular pressure and optic nerve injury in the rat, which is blocked with coadministration of the α-2 antagonist, rauwalscine.1–3 Increased retinal ganglion cell survival has also been shown to occur following oral administration of brimonidine in monkeys with experimental glaucoma.4 These results were the basis of the recently aborted clinical trial of topical brimonidine purite for acute NAION and our retrospective study of 31 patients with NAION, who were evaluated within 3 weeks of the onset of visual loss and had follow up for a minimum of 8 weeks. During 2001–2, we treated all (14) patients with brimonidine tartrate within 14 days (mean 5.5, SD 5.52) of the onset of visual loss. Five patients were treated after 1 day of symptoms. The drops were taken four times a day in 11, three times a day in one, and twice a day in two patients. All (17) untreated patients were evaluated the year before and were matched to the treated group for age, sex, cardiovascular risk factors, previous aspirin use, and previous first eye NAION.
Snellen visual acuity and colour vision, using the Ishihara colour plates, were documented and expressed as a decimal equivalent (for acuity: 20/60 = 0.33 and light perception = 0.001; for colour vision: the number of correctly identified plates/the total number of plates). The visual fields (Humphrey or tangent perimetry) were analysed and defects were graded according to the following scale: 0 = normal, 1 = arcuate nerve fibre bundle defects, 2 = relative central (<6 degrees), caecocentral or altitudinal defects, 3 = altitudinal defect plus additional loss, 4 = no light perception. A third examiner, who was unaware of the dates of the visual fields and the patients’ treatment status, also evaluated all visual fields and determined, in each patient, whether the field was better or worse than or equivalent to the other field. The intraocular pressure was normal in all except two patients. The pressure was 23 mm Hg in one patient in the treated group and 24 mm Hg in one patient in the untreated group.
Statistical analysis of the data involving comparisons of the treated and untreated groups at baseline and 8–12 weeks was performed using the two tailed t test.5 The Wilcoxon signed rank test was used to compare the individual vision performance from baseline to the 8–12 week examination.6 For visual acuity and colour vision, a positive rank indicated improvement and a negative rank indicated a worse visual outcome. For the visual field grade, the scale was reversed. Spearman correlation analysis was performed on the time to start therapy and whether worsening in any visual parameter occurred.7
The mean baseline acuity (0.56, SD 0.30) and visual field (1.9, SD 0.73) for the treated group was similar to the acuity (0.40, SD 0.41; p=0.22) and field (1.9, SD 0.75; p=0.96) for controls. The mean baseline colour vision (0.74, SD 0.4) for the treated group was higher than the colour vision (0.45, SD 0.44) for controls, but the difference was not significant (p=0.07). At the 8–12 week examination, the mean visual acuity was 0.29 (SD 0.30) for treated and 0.49 (SD 0.39; p=0.12) for untreated patients. The mean visual field grade was 2.2 (SD 0.81) for treated and 1.62 (SD 0.70; p=0.04) for untreated patients. The mean colour vision was 0.42 (SD 0.41) for treated and 0.55 (SD 0.46; p=0.43) for untreated patients.
For the masked examiner’s evaluation, the mean baseline visual field (2.0, SD 0.91) was similar to the field (1.93, SD 0.96; p=0.85) for controls. At the 8–12 week examination, the mean visual field grade was 2.15 (SD 0.99) for treated and 1.87 (SD 0.92; p=0.43) for untreated patients. This examiner further found that the outcome visual fields for the treated group were improved in two patients, worse in six patients (50%), and unchanged in four patients. The outcome visual fields for the control group were improved in five patients, worse in two patients (13%), and unchanged in eight patients.
The Wilcoxon signed rank analysis demonstrated that for visual acuity, two patients in the control group and 10 patients in the treated group had negative values or a worse outcome at 8–12 weeks (p=0.007). For colour vision, one patient in the control group and eight patients in the treated group had negative values or a worse outcome (p=0.013). For visual fields, one patient in the control group and four patients in the treated group had positive values or a worse outcome at 8–12 weeks (p=0.046).
The average time to start the drops was 3.5 days from the onset of visual loss in those patients who worsened. There was no correlation with a worse outcome and time to initiate therapy.
For all parameters of vision testing, there was a trend for worse visual performance at 8–12 weeks in the group treated with topical brimonidine. Although there was no significant difference for the colour vision outcome, this might reflect that the baseline colour vision value was better for the treated group. The outcome visual field grade was significantly worse in the treated group. The masked examiner’s visual field evaluations demonstrated that more treated patients worsened than in the untreated group. When the baseline and outcome of all visual parameters for each individual were compared, the treated group had a significantly worse outcome at 8–12 weeks.
Our results are not the first description of worse outcome in patients treated with α-2 agonists for central nervous system ischaemic disease. Studies in animal models and clinical studies in humans suggest that certain classes of drugs, including α-2 receptor agonists, may impede recovery following stroke. Clonidine administration caused recurrence of the neurological deficit in animals who had initially recovered. In a retrospective clinical study, the level of motor recovery of stroke patients was worse in those treated with α-2 agonists than in patients not receiving these agents.8
Although in experimental optic nerve injury in animal models, brimonidine appears to offer neuroprotection, our results demonstrate that brimonidine tartrate, applied topically up to four times daily, does not appear to be a beneficial treatment for acute NAION. It is possible earlier treatment might have been more effective, although patients who worsened received treatment sooner than those who did not worsen. Increased dosing frequency or using a different preparation of brimonidine might be more effective. Additionally, the number of subjects in the study was small and a negative trend could appear more profound.
References
- 1.Woldemussie E, Ruiz G, Wijono M, et al. Neuroprotection of retinal ganglion cells by brimonidine in rats with laser-induced chronic ocular hypertension. Invest Ophthalmol Vis Sci 2001;42:2849–55. [PubMed] [Google Scholar]
- 2.Wheeler LA, Woldemussie E. Alpha-2 adrenergic receptor agonists are neuroprotective in experimental models of glaucoma. Eur J Ophthalmol 2001;11(suppl 2):S30–5. [DOI] [PubMed] [Google Scholar]
- 3.Yoles E, Wheeler LA, Schwartz M. Alpha2-adrenoreceptor agonists are neuroprotective in a model of optic nerve degeneration. Invest Ophthalmol Vis Sci 1999;40:65–73. [PubMed] [Google Scholar]
- 4.Wheeler LA, Gil DW, Woldemussie E. Role of alpha-2 adrenergic receptors in neuroprotection and glaucoma. Surv Ophthalmol 2001;45(suppl 3):S290–4. [DOI] [PubMed] [Google Scholar]
- 5.Fisher LD, Van Belle G. One- and two-sample inference. In: Biostatistics, a methodology for the health sciences. Ch 5. New York: John Wiley, 1993.146–7.
- 6.Fisher LD, Van Belle G. Nonparametric, distribution-free and permutation models: robust procedures. In: Biostatistics, a methodology for the health sciences. Ch 8. New York: John Wiley, 1993:310–14.
- 7.Fisher LD, Van Belle G. Association and prediction: linear models with one predictor variable. In: Biostatistics, a methodology for the health sciences. Ch 9. New York: John Wiley, 1993:385–6.
- 8.Goldstein LB. Potential effects of common drugs on stroke recovery. Arch Neurol 1998;55:454–6. [DOI] [PubMed] [Google Scholar]