Purpose: To test the hypothesis that baseline optical coherence tomography (OCT) measures predict visual field (VF) progression in a cohort of patients with suspected or established glaucoma and to compare their performance to semiquantitative optic disc measures.

Design: This was an observational cohort study.

Methods: The setting of this study was an academic institution. The study population included 171 eyes of 95 patients with good-quality baseline retinal nerve fiber layer (RNFL) and macular OCT images and disc photographs with >2 years of follow-up and ≥5 VFs. The observation procedures were baseline macular and RNFL OCT measures and cup-to-disc ratio and disc damage likelihood score. The main outcome measure was prediction of glaucomatous VF deterioration according to trend and event analyses.

Results: Median (interquartile range) baseline mean deviation and follow-up were -2.9 (-6.4 to -1.1) dB and 54 (44-65) months, respectively. Seventeen and 25 eyes progressed by final visit based on pointwise event analysis and trend analysis of visual field index (VFI), respectively. Thinner central corneal thickness (P = .005), female gender (P = .015), and thinner average peripapillary RNFL (P = .001) predicted VF progression on proportional hazard models. Thinner RNFL at baseline (P = .006) or thinner average ganglion cell-inner plexiform layer (P = .028) along with higher baseline VFI (P = .018 and .048, respectively) predicted VFI progression. Neither optic disc measures predicted VF progression in any of the explored models.

Conclusions: Baseline structural OCT measures predicted subsequent VF progression in contrast to semi-quantitative optic disc measures. OCT-based structural measures should be included in prognostic models of glaucomatous VF deterioration.