This 3-year, prospective, observational post-marketing surveillance (PMS) study was conducted in daily clinical settings in Japan between July 2012 and December 2017 as a mandatory step for the approval of EX-PRESS in Japan and complied with the Japanese ministerial ordinance on standards for PMS studies, known as Good Post-Marketing Study Practice (GPSP). A GPSP study is conducted to detect or verify information on the incidence of adverse reactions and the quality, efficacy, and safety of a product without specifying the conditions of patients to be included. Under this ordinance, ethical approval of the participating medical institutions and patient consent were not required and were left to the discretion of the participating institutions. In practice, approximately half of the participating institutions obtained approval from their respective institutional review boards, while the remaining institutions adopted an opt-out approach. Importantly, informed consent for the surgical procedure was obtained from all patients, ensuring appropriate ethical consideration from a clinical perspective. The study protocol was submitted to the Pharmaceuticals and Medical Devices Agency (PMDA) on May 9, 2012, before the study began, and an outline of the study is included in the PMDA's re-examination report [16].

This continuous surveillance included all patients with glaucoma with insufficient IOP reduction despite medication, laser therapy, or prior glaucoma surgery who then underwent EX-PRESS filtration surgery at participating medical institutions. In principle, patients with the following conditions, as specified in the product’s package insert, were excluded from the study: uveitis, ocular infections, severe dry eye, severe blepharitis, angle-closure glaucoma, history of metal allergy, and any other ophthalmic or systemic condition deemed unsuitable for inclusion by the investigator. However, in clinical practice, certain cases were included at the investigator’s discretion, depending on individual case circumstances. The enrollment period was from July 2012 to December 2013. Each patient was followed up until the end of the observation period or when no further information could be obtained, with a standard observation period of 3 years for each patient. Postoperative management of glaucoma medication and the need for additional surgical procedures were left to the discretion of the treating surgeon.

The standard procedure for EX-PRESS filtration surgery is the same as that for trabeculectomy up to the point of fenestration, where a 25G needle is used to create a track parallel to the iris plane into the anterior chamber. Then, the EX-PRESS device, preloaded in the delivery system, is inserted and placed in the anterior chamber. After gently withdrawing the delivery system, the surgeon confirms that the device is in place and that the aqueous humor is flowing out of the eye. Finally, the surgeon sutures the sclera and conjunctiva in accordance with the trabeculectomy procedure. In actual surgical procedures, EX-PRESS filtration surgery was performed using either a fornix-based or limbus-based conjunctival incision, with the former being more common (608 eyes, 73.8%) than the latter (216 eyes, 26.2%). The scleral flap was most frequently rectangular in shape (722 eyes, 87.6%), while triangular, trapezoidal, double-flap, and other configurations were less commonly used. Scleral flap suturing was performed in all 821 eyes, with the number of sutures ranging from 0 to 11 (mean ± SD: 4.0 ± 1.4; median: 4.0). Most eyes received three to four sutures (49.6%), followed by five to six (29.9%), two or fewer (16.9%), and seven or more (3.3%).

Observations were made before EX-PRESS filtration surgery (baseline), during surgery, and on postoperative day 1 (POD1), day 2–3 (POD3), week 1 (POW1), month 1 (POM1), month 3 (POM3), month 6 (POM6), month 12 (POM12), month 18 (POM18), month 24 (POM24), month 36 (POM36), or when the patient became lost to follow-up.

The key variables collected were patient demographics, IOP, glaucoma medications, corrected visual acuity, endothelial cell count, and EX-PRESS implantation site. Adverse events (AEs) and malfunctions were reported in detail and assessed by the Safety Evaluation Committee as necessary.

In the survival analysis, success was defined as either definition 1 “IOP ≥ 5 mmHg and ≤ 21 mmHg without the use of glaucoma medications to lower IOP, with a IOP reduction of at least 20% from baseline, and without the need for glaucoma surgery” or definition 2 "meeting the criteria for definition 1, but with the use of glaucoma medications to lower IOP”. An event was defined as not meeting each of the success criteria.

Although this product was approved as a new medical device in Japan, it was unclear how much the transition from trabeculectomy to EX-PRESS filtration surgery would progress. As such, the target sample size was set at 600 patients to enable evaluation of efficacy and safety, as well as detection of less common AEs during the 3-year study period.

For the efficacy analysis, one eye per patient was used as the study eye. If the patient had undergone EX-PRESS filtration surgery in both eyes, study eyes for efficacy analysis was selected in the following order of priority; the eye with (1) available IOP data within 1 month before surgery, (2) higher average IOP of the three measurements from the preoperative examination, (3) higher maximum IOP in the past month from the preoperative examination, and (4) right eye.

For each variable, the following descriptive statistics were calculated; number of patients and eyes, mean ± SD, median, and maximum. Changes of IOP and endothelial cell density from baseline to each observation point were evaluated using the paired t test. Survival analysis for success 1 and success 2 was performed using the Kaplan–Meier method and the success rate at 3 years postoperatively was calculated. Beyond the initial analysis plan, further analyses were performed to assess changes in medication count using the paired t test, and changes over time in IOP, medication count, and endothelial cell count using the Wilcoxon signed rank test and mixed models for repeated measures. In this report, a p value of < 0.05 was used for determining statistical significance, but no adjustment was made for multiple comparisons. Statistical analysis was performed using SAS® 9.4 (SAS Institute, Cary, NC, USA).