Refining Visual Field Trend Progression Criteria in Glaucoma: Impact of Significance Thresholds and Test Frequency.

PURPOSE

To assess the stability of visual field (VF) progression events and to evaluate strategies to reduce false-positive (FP) progression.

DESIGN

Retrospective study.

SUBJECTS

Two thousand one hundred thirty-three eyes of 2133 patients with 12 consecutive VF tests from five National Health Service glaucoma clinics.

METHODS

A progression event was defined as a negative slope and P value < 0.05. False-positive progression events ("reversions") were defined as events where the P value fell below 0.05 at ≥ 1 visit and subsequently rose above 0.05 at a later time point. The impact of stricter probability thresholds (P < 0.02, 0.01 and 0.005) and event assessment at later visits on FP progression was evaluated.

MAIN OUTCOME MEASURES

Frequency of FP progression events.

RESULTS

Four hundred fifty-seven (21.4%) eyes had at least one reversion at the P < 0.05 threshold, for a median interval of 2 visits. The mean rate of change at the time of apparent progression was -1.05 dB/year (95% confidence interval -1.19 to -0.92 dB/year). The proportions of eyes with FP progression events were significantly lower at stricter probability thresholds (13.7%, 9.9%, and 7.7% at P < 0.02, < 0.01, and < 0.005, respectively; P < 0.001). Using stricter thresholds conferred significantly lower hazards of FP progression (hazard ratio 0.60 for P < 0.02, 0.42 for P < 0.01, and 0.38 for P < 0.005; all P < 0.001).

CONCLUSIONS

A substantial proportion of apparent progression at the established P < 0.05 threshold may be falsely positive. Using stricter probability thresholds and initiating event detection at later visits may improve progression specificity in clinical practice. FINANCIAL DISCLOSURE(S): Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.