Speculum-Free Intravitreal Injection Using Cotton-Tipped Applicator Retraction: Pain, Procedure Time, Patient Satisfaction, and Safety

Overview

This randomized controlled trial compares two techniques for eyelid retraction during intravitreal injection (IVI): an adjustable solid-blade eyelid speculum (Group A) and single-hand retraction of the upper eyelid with one cotton-tipped applicator (Group B), in patients with neovascular age-related macular degeneration (including polypoidal choroidal vasculopathy), diabetic macular edema, or retinal vein occlusion treated at Walailak University Hospital, Nakhon Si Thammarat, Thailand. The study evaluates four outcomes: (1) pain on a 0-10 numbered pain scale with a faces pictogram, 1-5 minutes after injection; (2) procedure duration from the first contact of the retraction device with the eyelid to its removal; (3) patient satisfaction on a 5-item Likert questionnaire; and (4) safety, including subconjunctival hemorrhage, corneal abrasion, lens touch, vitreous hemorrhage, endophthalmitis, retinal detachment, and intraocular pressure elevation. In both groups, the cap of the insulin syringe is used as an injection-site marker instead of a caliper. Randomization was stratified by diagnosis and prior injection history using permuted blocks of 4 and 6. All 129 participants were randomized and injected on September 9, 2026.

Background: Intravitreal injection is one of the most frequently performed ophthalmic procedures. Eyelid specula are widely used to keep the eye open but may cause discomfort. Cotton-tipped applicator retraction is an alternative that may reduce pain, and when a single applicator is held in one hand, the other hand remains free for injection-site marking and injection, so that one surgeon can complete the procedure without an assistant. Randomized data on procedure duration and satisfaction are limited. Study Design: Single-center, prospective, parallel-group randomized controlled trial with 1:1 allocation to an eyelid speculum (Group A) or cotton-tipped applicator retraction (Group B). The principal investigator generated stratified permuted-block sequences (block sizes 4 and 6) with Study Randomizer software; there were six strata: diagnosis (nAMD / DME / RVO) x prior injection history (injection-naive / previously injected). A research assistant placed the assignments in sequentially numbered, opaque, sealed envelopes, which another research assistant opened immediately before each procedure. Interventions: One vitreoretinal surgeon performed all injections without an assistant. Injections were given in the superotemporal quadrant in right eyes and the superonasal quadrant in left eyes (surgeon preference, same in both groups). All participants received tetracaine 0.5% and povidone-iodine 5% drops in the waiting area (contact time of at least 5 minutes), povidone-iodine 10% periocular skin preparation, a second instillation of tetracaine, and a fenestrated cloth drape. Anti-VEGF agents (bevacizumab, aflibercept, or faricimab) were prepared by the hospital pharmacy in 1-mL insulin syringes with a fixed 31G needle (Terumo). Group A: adjustable solid-blade eyelid speculum with a screw lock (PMS) retracting the upper and lower eyelids. Group B: one sterile cotton-tipped applicator held in the left hand to retract the upper eyelid; conversion to the speculum if retraction failed after 2 attempts. If the first attempt failed because of forceful eyelid squeezing or a marked Bell phenomenon, the participant was asked to close the eye, the applicator was used to retract the lower eyelid, and the injection was given in the inferotemporal or inferonasal quadrant. The injection site was marked with the povidone-iodine-dipped cap of the insulin syringe. Levofloxacin 0.5% was instilled after injection. Outcome Measures: All four outcomes are primary. Pain (0-10 numbered pain scale with faces), procedure duration (seconds), and satisfaction (5-item Likert questionnaire, total 5-25) were assessed on Day 0, 1-5 minutes after injection. Safety outcomes were assessed on Day 0, Day 1 (telephone), and Week 4 (clinic visit). Pain, procedure duration, and satisfaction are tested at a Bonferroni-adjusted alpha of 0.0167; safety is summarized descriptively (statistical analysis plan version 1.0, locked September 21, 2026). Masking: Open label. Participants, the surgeon, and the research assistants who timed the procedure and assessed the outcomes were aware of group assignment. Sample Size: Based on Chen et al. (2025) (pooled SD 1.71 for pain; Cohen's d = 0.57), with alpha = 0.025 (Bonferroni correction for 2 outcomes) and 80% power, 60 patients per group are required for pain (the earlier version of this record reported 52 per group because of a calculation error) and 46 per group for procedure duration. The enrollment target was 60 per group (120 total). All eligible patients who attended the injection session and gave consent were randomized (actual enrollment, 129).