We agree with Gedde and associates that the 3-year results of the Tube Versus Trabeculectomy Study show that greater medication use in the tube group at years 1 and 2 but not at year 3. A reduction in medication use in the trabeculectomy group during the first 2 years after surgery is a benefit to patients. We also appreciate their clarification regarding the important complication of diplopia, which was seen in 5% of patients having tube shunts at 3 years, and that new motility disturbances were seen in 9.9% of these patients in the first year.

Regarding complications as reported in the 3-year outcomes article, similar rates of early postoperative complications were reported for both groups. Regarding late postoperative complications, we note that the tube group had more diplopia (5% vs 0% in the trabeculectomy group), erosions (5% vs 0%), macular edema (5% vs 1%), and corneal edema (9% vs 6%). However, the trabeculectomy group had more dysesthesia (8% vs 1% in the tube group), hypotony maculopathy (4% vs 1%), and more leaks (5% vs 0%). Treating physicians may consider which complications they would rather treat, and this may differ among individual physicians. Finally, we recognize the limitations involved in designing a study to address these complex issues. We agree with the authors that gauging the severity of complications provides pragmatic challenges and commend them on finding a way to address these obstacles by defining serious complications as those requiring reoperation. We would have designed the study the same way. However, we think it is important to be mindful of problems, such as the weighting of complication severity, that are not easily quantified and compared, which is one of the perils of glaucoma surgical outcome analysis.