We thank Drs Minckler, Francis, Loewen, Mosaed, and Weinreb for their thoughtful comments on our manuscript, as well as for the opportunity to clarify some issues surrounding the data, particularly concerning the difference between survival rates and mean effects. We agree that it is too simplistic to have a single intraocular pressure (IOP) level serve as a cutoff for the use of Trabectome, and we appreciate the opportunity to offer our thoughts on the importance of considering the individual patient situation when attempting to incorporate study data into clinical practice.

In our study, at 24 months mean IOP was reduced 29% to 15.3 ± 4.6 mm Hg ( P < .001) and glaucoma medications were reduced 38% to 1.9 ± 1.3 ( P < .001), similar to previous studies. However, success rates were 62% (95% CI, 56%-68%) using criterion A and only 22% (95% CI, 16%-29%) using criterion B. The question, then, is why was the success rate using criterion B so low? The reason for this apparent discrepancy is that the magnitude of IOP and medication reduction was much greater in the Success group than in the Failure group. Although not included in the original manuscript, we compared the change in IOP and medications for these 2 groups. At 2 years, the mean IOP reduction for the Success group was 10.6 ± 8.1 mm Hg, and the mean reduction in medications was 1.7 ± 1.4. In contrast, the Failure group had a mean IOP reduction of only 0.7 ± 5.5 mm Hg and a mean reduction in medications of 1.0 ± 1.3 at the time point of failure. Over 80% of the Failure group for criterion B failed to meet the requirement of 20% IOP reduction from baseline.

This additional analysis suggests that the subset of patients who were successful by criterion B were extraordinarily successful. The mean change in IOP (and to a lesser extent the number of medications) is a reflection of this group and cannot be used to extrapolate to individual patients. However, survival analysis can provide some clues, and in our manuscript we reported that past argon laser trabeculoplasty and primary open-angle glaucoma were risk factors for failure, while pseudoexfoliation was a predictor of success. However, further work is clearly required to refine the factors that will allow us to identify optimal patients for this procedure.

Concerning the issue of patients who had previous trabeculectomy or tube shunt potentially biasing the results, we reanalyzed the success rates excluding these subjects and found very similar rates. The success rate using criterion A was 61.8% (95% CI, 55.5%-68.9%) at 2 years, compared with our originally reported analysis showing a success rate of 62% (95% CI, 56%-68%) at 2 years. For criterion B, the success rate at 2 years was 21.8% (95% CI, 16.3%-29.0%), compared with our original analysis, which showed a success rate of 22% (95% CI, 16%-29%).

The issue of comparing Trabectome and trabeculectomy is an important one. Although there are no prospective study results available, a retrospective cohort study by Jea and associates found that Trabectome reduced IOP from 28.1 ± 8.6 mm Hg at baseline to 15.9 ± 4.5 mm Hg at 24 months (similar to our study and previous work), while trabeculectomy had a significantly larger reduction of IOP, from 26.3 ± 10.9 mm Hg at baseline to 10.2 ± 4.1 mm Hg at 24 months ( P < .001). Success rates, defined as IOP ≤21 mm Hg and 20% or more reduction below baseline, were 22.4% for the Trabectome group and 76.1% for the trabeculectomy group ( P < .001) at 24 months. It should be noted that the Trabectome success rate was nearly identical to our success rate using the similar criterion B, and the trabeculectomy success rate was similar to the 3-year survival in the Tube Versus Trabeculectomy Study. These results again illustrate the importance of analyzing success rates in addition to mean IOP reduction.

Finally, we fully agree that the results of this study should not be the sole basis for a decision concerning whether or not to perform ab interno trabeculotomy using the Trabectome. In our conclusion, we stated that “this surgery is appropriate for patients requiring a target IOP of 21 mm Hg or above.” This should not imply that it is inappropriate for all patients with lower target pressures. The decision about whether or not to offer a treatment must include an analysis of the risks vs benefits. A low-risk procedure with lower success rates, such as Trabectome, can reasonably be offered prior to a higher-risk procedure with higher success rates, such as trabeculectomy, as long as the ability to perform future interventions is not impaired. Previous studies have reported that Trabectome does not appear to impair future filtering surgery. Thus, success rate alone should not determine whether or not to offer Trabectome as a treatment for glaucoma. Instead, this decision should be made on an individual basis within the doctor-patient relationship.