Mean deviation fluctuation in eyes with stable Humphrey 24-2 visual fields.
AIM: To describe the expected fluctuation in the mean deviation (MD) scores on a large, long-term series of stable visual field reports for particular grades of defect in order to give clinicians an aid to the correct diagnosis of glaucomatous progression. METHOD: Visual field reports of subjects with five reliable consecutive Humphrey 24-2 visual fields, recorded over a period of at least 3 years, were scored using the Advanced Glaucoma Intervention Study (AGIS) system. The AGIS scores of the first and last visual fields were required to be identical. RESULTS: A total of 202 eyes from 202 patients were used in the study, with a total of 1010 visual fields being used in the analysis. Visual fields with no defect (AGIS score 0) had a 99% confidence interval (CI) of 0.3 dB, mild defects 0.4 dB, moderate defects 0.8 dB, severe 1 dB, and 1.3 dB for end-stage defects when considering variation of MD scores. Using a one-way ANOVA incorporating all stages showed very little fluctuation throughout the series (P=0.96). The correlation between the CI and grade of field defect showed a good positive correlation (r=0.7, P=0.0003) indicating an increase in CI as field defects worsen. CONCLUSION: When considering a series of reliable visual fields in a stable eye, one should expect only very little fluctuation in the MD, indicating that an increase in the MD beyond that of the 99% CI described may suggest progression. The reliability indices of the visual field test should be regarded as a primary consideration when assessing visual fields.