Five-year visual field outcomes of the HORIZON trial

Abstract

Purpose : To compare the rate of progression (RoP) of visual field (VF) damage between the two arms of the HORIZON prospective randomized multicentre trial

Methods : A total of 556 participants (one eye per participant) were randomised 1:2 to receive cataract surgery alone (CS, n = 187) or in combination with a Hydrus Microstent (Ivantis, Inc. Irvine, CA; CS-HMS, n = 369). Patients followed for 5 years had planned VF tests at baseline and 6, 12, 24, 36, 48 and 60 months after surgery using a Humphrey Field Analyzer (HFA, Zeiss Meditec, Dublin, CA, USA), 24-2 pattern, SITA-Standard strategy. Additional tests were performed if a worsening in mean deviation of 2.5 dB or more was observed. A total of 3701 VF from 554 (99.6%) patients were available for analysis. Of these, 121 VFs were excluded due to poor reliability (false positive rate > 15%). To eliminate any effects of cataract on VF, pre-surgery baseline tests were excluded from the analysis (561 reliable VFs). We analysed eyes with at least 3 VFs available over at least one year (165 CS and 352 CS-HMS, 2966 VFs). A secondary more inclusive analysis was conducted on all eyes with at least 2 VFs (177 CS and 361 CS-HMS, 3008 VFs). The difference in RoP between the two groups was estimated using a hierarchical mixed effect Bayesian model, modelling point-wise sensitivity values grouped by location, Garway-Heath clusters, eye (all random effects) and arm of the trial (fixed effect). A statistically significant difference in RoP was tested via an interaction term between the effect of the arm assignment and the time from surgery. Significance was reported using a Bayesian p-value. Estimates of the effect are reported, along with 95% Bayesian Credible Intervals.

Results : There was no significant difference in the estimated baseline (intercept of the model) between the two arms, either with the main or secondary analysis. The RoP was -0.26 [-0.36, -0.16] dB/year (47%) slower in the CS-HMS arm (p = 0.0138) in the main analysis and -0.23 [-0.40, -0.05] dB/year (43%, p = 0.0284) in the secondary analysis. Numerical results are reported in Table 1 and Figure 1.

Conclusions : CS-HMS significantly reduced the speed of VF loss compared to CS alone in glaucoma patients by a clinically significant amount.

This abstract was presented at the 2022 ARVO Annual Meeting, held in Denver, CO, May 1-4, 2022, and virtually.

 

Table 1

 

Figure 1