Surgeon Experience as a Risk Factor for Short-Term Failure for Ab Interno Gelatin Microstent: A Canadian Multicenter Propensity-Matched Study

Abstract

To compare the efficacy and safety of early versus later ab interno gelatin microstent placement with mitomycin C. Canada-wide, multicenter, retrospective propensity score-matched cohort study. Two hundred seventy eyes (135 early cases and 135 later cases) with no prior incisional surgery. Surgeons' first 20 patients (early cases group), from 6 glaucoma surgeons across 4 Canadian sites, were matched 1:1 to patients with the closest propensity score from the later (21+) patients (later cases group). Primary outcome was hazard ratio (HR) of failure of the early versus later cases groups, with failure defined as IOP of less than 6 mmHg with more than 2 lines of vision loss or more than 17 mmHg with no medications (complete success) on 2 consecutive visits despite in-clinic maneuvers (including needling) more than 1 month after surgery. Secondary outcomes were HRs for failure, defined as IOP outside the range of 6 to 14 mmHg and 6 to 21 mmHg with and without allowing for medications (qualified success), interventions, complications, and reoperations. Hazard ratio of failure for early versus later cases groups was 1.38 (95% confidence interval [CI], 0.97-1.96) for the IOP range of 6 to 17 mmHg, 1.29 (95% CI, 0.90-1.84) for 6 to 14 mmHg, and 1.48 (95% CI, 1.03-2.13) for 6 to 21 mmHg without medication and 0.95 (95% CI, 0.55-1.64), 0.95 (95% CI, 0.61-1.48), and 0.95 (95% CI, 0.52-1.75) for the same IOP ranges allowing for medications. Needling rates were 43.0% (early cases group) and 41.5% (later cases group). Complication rates after 1 month occurred in 9.6% (early cases group) and 11.1% (later cases group; P= 0.69). Reoperation rates were 14.8% (early cases group) and 8.1% (later cases group; P= 0.08). There is some evidence for improved success in the later cases group. Similar needling rates, similar complication rates, and a slightly higher reoperation rate were found for the early cases group. The results suggest that this procedure can be adopted by existing surgeons with current training regimens, although they may see an improvement in their success outcomes over time.