We appreciate the editor's selection of our paper for the Glaucoma Dialogue, and we thank the reviewers for their insightful comments. This post-publication peer review provides a unique opportunity to explore the data from the Tube Versus Trabeculectomy (TVT) Study, and the conclusions derived from these data. The reviewer comments may be grouped into specific areas of interest.
The TVT Study is a multi-center randomized clinical trial, designed to compare the safety and efficacy of tube shunt surgery to trabeculectomy with mitomycin C (MMC) in patients with previous cataract and/or glaucoma surgery. The design and methods of the study were described in detail in a prior publication.1 Reviewer 6 questioned whether the study was originally designed in a predetermined 'stratified sampling' design or whether 'poststratification' was performed. When the TVT Study was being designed, the investigators were concerned that the type of previous ocular surgery might be an important risk factor affecting the rate of surgical failure. Therefore, patients were stratified at randomization into four groups based upon the type of qualifying surgery. This process was successful in balancing the treatment groups with respect to type of prior ocular surgery, and removed this as a potential confounder of the results. The study was never intended to be four studies in one as suggested by reviewer 6, and it was not powered to try to discern treatment differences between individual strata. However, it is noteworthy that stratum was not a significant risk factor for failure,2 and the difference in failure rates was similar among strata (non-significant treatment by stratum interaction).2,3
Both surgical procedures under investigation in the TVT Study were standardized, but allowed the surgeon some latitude to perform the operations in a manner with which he or she felt comfortable. Patients randomized to the tube group all had placement of a 350-mm2 Baerveldt glaucoma implant. Although several retrospective studies have failed to detect a difference in surgical results when comparing various implant types,4-8 reviewers 2 and 4 are correct in noting that the TVT Study cannot be generalized to other glaucoma drainage implants. All patients randomized to the trabeculectomy group underwent a trabeculectomy superiorly with a standard dosage of MMC of 0.4 mg/ml for four minutes. Reviewer 4 has appropriately indicated that modifications in trabeculectomy technique have developed since the TVT Study was initiated. In particular, there has been a trend toward use of a fornix-based conjunctival flap with diffuse application of MMC at a lower dosage.9 These modifications may result in a lower risk of bleb leaks and bleb-related infections.
Reviewer 2 questioned why the study did not provide information on early postoperative management specific to trabeculectomies. A prior paper reviewing the results of the TVT Study during the first year of follow-up included data on postoperative interventions, including laser suture lysis, bleb needling, and subconjunctival 5-fluorouracil (5-FU) injections.10It seemed redundant to publish this information again in an interim paper discussing the 3-year results of the study, but it is readily available to readers.
Reviewers 1 and 5 were surprised by the high rate of postoperative complications in the TVT Study. Half of the patients in the study experienced complications during the first three years of follow-up, including 39% in the tube group and 60% in the trabeculectomy group.3 The rate of perioperative complications in the first month after trabeculectomy was 50% in the Collaborative Initial Glaucoma Treatment Study (CIGTS).11 Most complications in both of these trials were transient and self-limited, such as shallowing of the anterior chamber and choroidal effusions. It is not unexpected that prospective studies like CIGTS and TVT have reported higher complication rates than retrospective case series. Complications may be overlooked unless attention is specifically directed toward their detection. Moreover, even when surgical complications are observed, they may not be documented in the medical record (especially if they are believed to be insignificant). We share the enthusiasm of reviewer 1 to find a glaucoma procedure that optimizes surgical success and minimizes complications.
Complications were more common after trabeculectomy with MMC than tube shunt surgery in the TVT Study. However, all complications are not equal in severity, and the rate of serious complications was similar with both surgical procedures. We had considered trying to 'weight' complications, but reviewer 3 has clearly identified the difficulty in attempting to assign relative values to each complication. Furthermore, individual complications have a broad range of severity. For example, diplopia may represent an infrequent symptom elicited in extreme gaze or a disabling problem that is always present. Similarly, a suprachoriodal hemorrhage may be visually devastating or limited and without sequelae. We instead chose to define serious complications as those that were associated with a reoperation to manage the complication and/or loss of two or more lines of Snellen visual acuity.3,10
Reviewer 6 noted that the frequency of wound leaks was higher in the TVT Study than has been reported in other clinical trials. The rates of wound leaks were 11% in the TVT Study,3,10 5% in CIGTS,11and 6.5% in the Advanced Glaucoma Interventions Study.12 However, there are fundamental differences in the study populations, use of antifibrotic agents, and methods for assessing for wound leaks between these studies. CIGTS and AGIS enrolled patients without prior ocular surgery, and MMC was not used in any initial trabeculectomies in AGIS12 and in only 5% of cases in CIGTS.11 Patients in the TVT Study had previous cataract and/or glaucoma surgery, and all trabeculectomies were performed with adjunctive MMC. The TVT Study protocol required Seidel testing at every follow-up visit, and this method for meticulously detecting wound leaks was not used in AGIS and CIGTS. The Fluorouracil Filtering Surgery Study (FFSS) may be a more appropriate study to compare the rate of wound leaks with the TVT Study, because both studies enrolled similar patient groups (i.e., prior cataract extraction or failed filtering surgery) and employed the same protocol of compulsive Seidel testing at each visit. The rate of early postoperative wound leaks was 20% in the standard treatment group and 32% in the 5-FU group in FFSS.13 Wound leaks were associated with an increased risk of surgical failure in both the FFSS and TVT Study.10,13
Reviewer 5 suggested the benefit of an economical analysis from the TVT Study. A paper is under preparation that provides a cost analysis of tube shunt surgery and trabeculectomy with MMC in the study.
Each reviewer has highlighted important findings of the TVT Study. Tube shunt surgery had a higher success rate than trabeculectomy with MMC after three years of follow-up applying prospectively defined success and failure criteria.3 Both surgical procedures were associated with similar IOP reduction and use of supplemental medical therapy at three years. The rate of serious postoperative complications was not significantly different after tube shunt surgery and trabeculectomy with MMC. The TVT Study does not demonstrate clear superiority of one glaucoma operation over the other. There are other factors that must be considered when selecting a surgical procedure in similar patients with medically uncontrolled glaucoma, such as the surgeon's skill and experience with both operations, the patient's willingness to undergo repeat glaucoma surgery, and the surgeon's planned surgical approach should failure occur. Reviewer 2 suggests that the reader is left 'none the wiser' because the TVT Study does not show which glaucoma procedure is superior. However, we agree with reviewer 3 who has suggested that the value of the TVT Study does not come from showing superiority of one operation over the other, but rather in providing comparative data on the outcomes of both surgical procedures that will help surgeons in selecting the most appropriate glaucoma operation for their individual patients.