Effectiveness and cost-effectiveness of MicroShunt implantation versus standard trabeculectomy for open-angle glaucoma (a SIGHT study): study protocol of a multicentre randomised controlled trial

The aim of this study is to evaluate whether MicroShunt implantation is non-inferior regarding effectiveness, where effectiveness is assessed using the IOP after 12 months of follow-up, as compared to trabeculectomy in open-angle glaucoma patients. In addition, IOP-lowering medication use, treatment failure, visual acuity, visual field progression, additional interventions, adverse events, patient-reported outcome measures (PROMs), and cost-effectiveness are evaluated.

A multicentre, non-inferiority, single-blinded, randomised controlled clinical trial will be performed at outpatient ophthalmology clinics in the Netherlands. The follow-up duration will be one year. The study design is presented in Fig. 1.

Hospitals with glaucoma surgeons that regularly perform trabeculectomy can participate in this trial. Five academic centres (Amsterdam University Medical Center; Erasmus Medical Center; Leiden University Medical Center; Maastricht University Medical Center+; University Medical Center Groningen), and four non-academic centres (Bravis Hospital, Bergen op Zoom; Deventer Hospital; The Rotterdam Eye Hospital; Zuyderland Medical Center, Heerlen) will participate. Surgeons will be trained for the MicroShunt procedures before participating in the study to minimize a learning curve. Before the start of the trial, a masterclass will be organised to reach a consensus on the details of the study methods and the surgical technique.

Caucasian patients between 18 and 80 years old with uncontrolled open-angle glaucoma on maximum tolerated medical therapy and/or progression on the visual field with an indication for primary glaucoma surgery will be suitable for inclusion. The exclusion criteria are listed in Fig. 2.

Patients meeting the eligibility criteria will be counselled about the study by their ophthalmologist. Subsequently, a local researcher will provide them with verbal and written information. Patients who agree to participate in the study must provide written informed consent before randomisation is performed.

The certified electronic data capture tool Castor EDC (Castor Electronic Data Capture, version 2022.2.2.2, Ciwit BV, Amsterdam, the Netherlands) will be used for randomisation, sending out questionnaires, and data collection [18]. The computer will assign each patient a unique code, consisting of a centre-specific prefix and a serial number. The subject code will then be allocated to either MicroShunt implantation (intervention) or trabeculectomy (control) in a 1:1 ratio through block randomisation, stratified according to centres, and using variable block sizes of 2, 4, and 6. Study participants will be blinded to the intervention for one year. Due to the nature of the intervention, it is impossible to blind the treating physician.

The intervention consists of MicroShunt implantation augmented with MMC application. It comes in a package including the surgical instruments required to implant the device. A detailed description of the procedure can be found elsewhere [19, 20]. To summarize, a fornix-based conjunctival/Tenon’s dissection will be made, and a deep sub-Tenon’s pocket will be formed. After application of MMC, the MicroShunt will be implanted into the anterior chamber through a 25-gauge scleral needle track, and the Tenon’s and conjunctiva will be sutured in a watertight fashion.

The intervention will be compared to the usual care, which is a fornix-based trabeculectomy augmented with MMC application, according to the surgeon’s preferred technique. Both procedures may be performed under local anaesthesia or general anaesthesia conformable to local conditions in the participating centre. After surgery, topical anti-inflammatory therapy will be commenced and gradually tapered off according to bleb formation. IOP-lowering medication may be (re)introduced according to the discretion of the treating ophthalmologist. In the event of an impending bleb failure, further surgical interventions may be performed.

The primary outcome will be the level of IOP one year after surgery. IOP will be measured by Goldmann tonometry in mmHg.

Secondary outcomes will include:

An overview of the study measurements per visit is presented in Table 1. Enrolled patients complete follow-up visits at one day, one week, four weeks, three months, six months, and one year after surgery. Time windows are based on the proposed time windows of the guidelines on design and reporting of the World Glaucoma Association (WGA) [21].

The sample size is based on the primary outcome, IOP, a numerical response variable, to detect non-inferiority of an experimental group compared to a control group at 12 months of follow-up. A couple of studies have been published on the efficacy of the experimental group. Based on observed means of postoperative IOP from these studies [11, 26, 27], a mean IOP for the experimental group of 12 mmHg was chosen. The non-inferiority margin (control minus intervention) of − 2.5 mmHg is chosen based on previous trials investigating glaucoma surgery [26, 28, 29]. Assuming this non-inferiority margin of − 2.5 mmHg, a within-group standard deviation (SD) of 4.4, and an expected effect (control minus intervention) of 0.4 [5], 49 patients per group will be required to show non-inferiority with 90% power and a one-sided significance level alpha of 0.025. Accounting for 20% loss-to-follow up, we will need to include 62 patients per group, i.e., 124 patients in total.

As for secondary outcome parameters, this number of patients should be sufficient to detect relevant differences in CDVA and postoperative intervention and complication rates. A 0.13 logMAR difference between the two groups (within group SD = 0.2) can be detected with 90% power and a two-sided alpha of 0.05. In the primary tube versus trabeculectomy study [5], the postoperative intervention and complication rate in the control group was 63 and 41%, respectively. With the planned sample size for our study (124 patients in total with 20% dropout), we can detect a difference of 32% (63% in the control group versus up to 31% in the intervention group) for postoperative interventions and 28% (41% in the control group versus 13% in the intervention group) for postoperative complications, with 90% power and a one-sided significance level alpha of 0.025.

A single main analysis will be performed at the end of the trial when all follow-up visits have been completed.

The data analysis will be performed according to the intention-to-treat principle. Additionally, a per-protocol analysis will be performed, as is recommended for non-inferiority trials [30]. The per-protocol population includes all randomized patients who complete treatment without major protocol deviations. The details of the number of eligible patients for the trial, the number of consenting, and the number of randomised patients will be shown graphically as a CONSORT diagram [31], to describe the course of patients throughout the trial.

Baseline characteristics will be presented as means with standard deviation and 95% confidence intervals, as median and interquartile range (IQR), or as frequencies (with percentages), as appropriate. For the primary outcome, IOP, a linear mixed model analysis will be used since it accounts for the stratification variable (centre), baseline differences, uses all available data, corrects for correlation between repeated measures, and assumes missing data to be missing at random (MAR). A similar model will be used for the secondary outcomes that are also repeatedly measured. Different options will be considered for the covariance structure of the repeated measures as well as the random part of the model, and the one with the lowest Bayesian Information Criterion (BIC) will be chosen. To determine the success rates, Kaplan-Meier survival methods as well as Cox regression analysis will be used.

A trial-based cost-effectiveness analysis will be conducted from a societal and a healthcare perspective over a 12-month time horizon. A detailed and complete analysis of the costs of each patient included in the study up to 12 months from a healthcare and societal perspective will be performed. Resource use will be collected by means of the case report form (CRF) regarding both procedures, (pre- and postoperative) visits, and treatment of adverse events. Information on resource use outside of the hospital (e.g., productivity losses, homecare) will be collected by means of a short cost questionnaire. Valuation of resource use will be performed in accordance with the Dutch guideline for cost analysis [32] using either reference prices or local cost prices provided by the Maastricht University Medical Center+, whichever is deemed more appropriate.

QALYs will be calculated based on EuroQol’s EQ-5D-5L using the Dutch tariffs [24, 33]. Additionally, the HUI-3 will be used because it is the only generic preference-based HRQL questionnaire that includes questions about vision [25].

ICERs will be calculated for the subtypes described earlier in this paper and are calculated by dividing the cost difference between the groups by the difference in effects. To estimate uncertainty in the point estimates of the ICERs, bootstrap cost-effect pairs will be plotted on cost-effectiveness planes. The probability that the MicroShunt is cost-effective will be estimated using cost-effectiveness acceptability curves for a range of threshold values. Additional sensitivity analyses will be performed to investigate the impact of varying input variables.

Alongside the cost-effectiveness analysis, a budget impact analysis (BIA) will be performed to evaluate the impact of implementation of the MicroShunt on the Dutch healthcare budget compared to standard trabeculectomy. The BIA addresses the financial stream of consequences related to the implementation of the MicroShunt to assess affordability and will be performed in accordance with the Dutch guidelines for economic evaluations and the International Society for Pharmacoeconomics and Outcomes Research (ISPOR) guidelines [33, 34]. Costs will be calculated using similar methods to those described in the cost-effectiveness analysis section. Prices will be corrected to the price levels of each specific budget period (i.e., one calendar year), and costs will not be discounted. A simple decision analytic model will be built using Microsoft Excel (Microsoft Corporation, Redmond, Washington, United States, version 2016 or higher). The analysis will use the perspective of the budget holder, complemented with additional perspectives such as healthcare providers and insurers. A time horizon of five years will be used to account for the fact that implementation will take place gradually. It will be assumed that the difference in costs will be related entirely to the substitution of trabeculectomy by MicroShunt implantation. Various scenarios will be compared to investigate various levels of implementation (i.e., 25, 50, 75, and 100% of eligible patients). Furthermore, scenarios will be modelled in which the timeline of implementation in 100% of the hospitals varies from direct implementation to implementation in five years. Sensitivity analyses will be performed to test the robustness of the analysis.

Since this study investigates an approved intervention within its indication of use, a Data Safety Monitoring Board will not be installed, and an interim analysis will not be conducted.

Personal data will be handled confidentially. Data will be collected in the online CRF by a member of the study team of each study centre. The dataset is coded, and the local research team will have access to the key. All data will be stored and analysed using the unique study code. After completion of the study, the data will be archived for 15 years. Monitoring will be performed by the Clinical Trial Center Maastricht (CTCM), an academic research organisation affiliated with the sponsor, in order to protect patient rights and the accuracy of the reported trial data. Site monitoring visits will take place to assess the adherence to the study protocol and the performance of the participating sites. Qualified and independent monitors will have access to the data and source documents.

Adverse events (AE) are defined as any undesirable experience occurring to a subject during the study. All AEs related to the study eye, predefined by the WGA [21], reported spontaneously by the subject or observed by the investigator or his staff will be recorded unless the event is of negligible impact and, in addition, has no connection to the study anywise (e.g., the common cold). All serious adverse events (SAE) will be reported by the local investigator to the principal investigator. The principal investigator will report a SAE to the ethics committee of Maastricht University Medical Centre+ and Maastricht University (azM/UM) within a period of maximum 15 days after the first knowledge of the serious adverse event. If SAEs occur that are unrelated to the intervention, these SAEs will be reported in line listing in the annual progress report.