Methods: Light perception was tested in the preoperative assessment. The following parameters were evaluated at the pre- and postoperative visits: morning Goldmann applanation tonometry; anterior segment biomicroscopy, gonioscopy, and eye pain intensity. Postoperative visits occurred on days 1, 3, 7, 14, and 30, once a month up to 6 months, and every 6 months from 6 months to 2 years. The criteria for IOP reduction success were 15 mmHg and a 20% reduction compared to the baseline preoperative IOP. Biomicroscopy variables included the presence or absence of hyperemia, discharge, conjunctival erosion and retraction, aqueous humor leakage (Seidel test), corneal edema, hyphema, AC reaction and depth, and device dislocation. A shallow AC was considered a Spaeth classification grade of I or II and a flat AC was considered grade III. A filtering bleb height above the LDI profile was considered elevated; otherwise, it was classified as flat. Gonioscopy was performed preoperatively and postoperatively. Ocular pain intensity was assessed before and after surgery using a verbal Numerical Rating Scale ranging from zero to 10, with zero representing the absence of pain and 10 representing the worst eye pain ever felt before surgery.
Anterior segment optical coherence tomography (AS-OCT) (Visante®, Carl Zeiss Meditec®, Inc., Dublin, CA, USA) was performed postoperatively at the highest resolution (enhanced high-resolution corneal mode).
Reoperations to unblock the internal ostium were not considered failures; thereafter, the day of the second surgery was considered postoperative day 0.
LDI A human LDI was designed by one of the authors (S.J.) after a preclinical trial in rabbits with a similar device. It is a 150-µm-thick foil made of polymethylmethacrylate (PMMA), composed of three functional areas (Fig. 1): 1) the tip, which has a triangular shape and is inserted into the AC angle; 2) the intermediate portion, which has a rectangular shape (4.5 mm long × 4.0 mm wide) and is positioned inside a scleral tunnel; and 3) the posterior plate, which has a round shape (diameter = 12.5 mm) and is positioned between the sclera and conjunctiva. The LDI plate is positioned only 4.5 mm from the corneoscleral limbus and no allograft coverage is necessary. Its hypothetical mechanism of action is similar to those of popular GDIs and shunts the aqueous humor from the AC to the subconjunctival plate. The difference between their designs is the absence of a tube in the LDI. Thus, aqueous humor drainage is supposed to occur in the space between the inner walls of the scleral tunnel and LDI and/or thought is cleft.
LDI Surgery All surgeries were performed under local peribulbar anesthesia by the same surgeon (S.J.).
Statistical Analysis The sample size was calculated based on IOP standard deviation values from the first seven patients who were operated on (an internal pilot study) while considering 15 mmHg as the minimal clinically relevant difference to be detected.
Histogram inspection and Shapiro-Wilk tests were used to assess the assumption of the normality of variables. Wilcoxon's rank-sum test was used to compare the median pain intensity at each postoperative visit to that at the preoperative visit. Conjunctival hyperemia, AC cells, corneal edema, and hyphema variables were analyzed using McNemar's test. The comparison between the pre- and postoperative mean IOP was performed using Student's two-tailed paired t-test. Results were analyzed using SPSS® (Statistical Package for the Social Sciences) for Windows version 18.0 (IBM, Chicago, IL, USA) and the alpha level (type I error) was set at 0.05.
