Glucagon-like peptide-1 receptor agonists (GLP-1 RAs) are important therapeutic options for type 2 diabetes and obesity; however, concerns about ophthalmic safety persist. This study examined associations between GLP-1 RAs and ocular adverse events (AEs). Global observational pharmacovigilance study. We searched the US FAERS database (via OpenVigil 2.1) and WHO's VigiBase (via VigiAccess) for optic nerve and retinal AEs associated with semaglutide and tirzepatide, covering the period from their respective approval dates-December 2017 for semaglutide and May 2022 for tirzepatide-through September 2024. In FAERS, all other drugs were compared, while in VigiBase, metformin, empagliflozin, dulaglutide, and insulin served as controls. Disproportionality metrics included reporting odds ratios (RORs) with 95% confidence intervals. Semaglutide and tirzepatide accounted for 76 444 cases (0.59%) in FAERS (n = 12 936 341) and 118 639 cases (0.34%) in VigiBase (n > 35 000 000). Semaglutide showed significantly higher odds of ischemic optic neuropathy (ION) (FAERS: ROR = 11.12, 95% CI = 8.15-15.16; VigiBase: ROR = 68.58, 95% CI = 16.75-280.67), diabetic retinopathy (DR) (FAERS: ROR = 17.28, 95% CI = 13.62-21.91; VigiBase: ROR = 7.81, 95% CI = 5.60-10.90), as well as retinal/vitreous detachment, retinal/vitreous hemorrhage, and retinal tear (FAERS: ROR = 2.44-5.89, 95% CI = 1.70-8.97, all P < .001, IC₀₂₅ = 0.49, compared to all other drugs. VigiBase: ROR = 5.49-20.91, 95% CI = 2.71-90.11, all P ≤ .0001, IC₀₂₅ ≥ 0.53, compared to metformin). Unique to VigiBase were macular edema (ROR = 3.87, 95% CI = 1.89-7.92), macular hole (ROR = 20.90, 95% CI = 2.65-165.01), and papilledema (ROR = 6.97, 95% CI = 2.53-19.17) (all P ≤ .004, IC₀₂₅ ≥ 0.27, compared to metformin). Sensitivity analyses using empagliflozin and dulaglutide revealed significant associations with ION and DR, while vitreous detachment and hemorrhage were significant when compared to dulaglutide. Additionally, when insulin was used as a comparator, semaglutide showed a higher ROR for ION (ROR = 9.84, 95% CI = 4.25-22.81, P < .0001, IC₀₂₅ = 0.42). However, tirzepatide was only significantly associated with DR in FAERS. Given the widespread use of semaglutide, its association with ocular AEs highlight the need for global pharmacovigilance and post-marketing surveillance.

Subthreshold lasers operate below the threshold of visible tissue damage, thereby preventing ophthalmoscopically visible thermal damage to the chorio-retinal layers. They could represent a safe and effective alternative and/or adjunctive procedure to conventional lasers in treating diabetic macula edema (DME), central serous chorioretinopathy (CSCR), and branch retina vein occlusion (BRVO). This review focuses on the use of subthreshold micropulse laser (SMPL), its settings, and clinical applications. Despite their widespread use, a standardized protocol for sub-threshold laser settings has not been established yet, and thus, there is uncertainty in selecting effective and safe parameters for any specific situation. We conducted a comprehensive overview of the existing indications for subthreshold laser therapy and their settings for different retinal diseases. The debate revolves around which parameters could guarantee the safety of the procedure for each case, depending on the duty cycle, the laser wavelength, the spot duration, and the power, with laser power titration on one side or choosing a fixed lowered power value on the other side. SMPL therapy for DME, CSCR, and BRVO-associated macular edema has shown significant effectiveness in reducing the macular thickness, facilitating subretinal fluid absorptions, increasing the best corrected visual acuity (BCVA) and reducing the number of intravitreal injections (IVI) required annually. We presented a broad list of the laser parameters reported in the literature, organized into different tables divided based on the specific pathology, with the aim of providing a useful tool for future studies.

Purpose: To compare the performance and cost-effectiveness of DeepSeek-R1 with OpenAI o1 in diagnosing and managing ophthalmology clinical cases. Study Design: Cross-sectional evaluation. Methods: A total of 300 clinical cases spanning 10 different ophthalmology subspecialties were collected from StatPearls. Each case presented a multiple-choice question regarding the diagnosis or management of the clinical case. DeepSeek-R1 was accessed through its public chat-based interface, while OpenAI o1 was queried via an Application Program Interface (API) with a standardized temperature setting of 0.3. Both models were prompted using the Plan-and-Solve+ (PS+) prompt engineering method, instructing them to answer multiple choice questions for each case. Performance was calculated as the proportion of correctly answered multiple choice questions. McNemar’s test was employed to compare the two models’ performance on paired data. Inter-model agreement for correct diagnoses was evaluated via Cohen’s kappa. A token-based cost analysis was performed to estimate the comparative expenditures of running each model at scale, accounting for both input prompts and model-generated output. Results: DeepSeek-R1 and OpenAI o1 both achieved an identical overall performance of 82.0% (n=246/300; 95% CI: 77.3-85.9). Subspeciality-specific analysis revealed numerical variation in performance, though none of these comparisons reached statistical significance (p>0.05). Agreement in performance between the models was moderate overall (κ=0.503, p<0.001), with substantial agreement in Refractive Management/Intervention (κ=0.698, p<0.001) and moderate agreement in Retina/Vitreous (κ=0.561, p<0.001) and Ocular Pathology/Oncology (κ=0.495, p<0.01) cases. Cost analysis indicated an approximately 15-fold reduction in per-query, token-related expenses when using DeepSeek-R1 compared with OpenAI o1 for the same workload. Conclusions: DeepSeek-R1 demonstrates robust diagnostic reasoning and management decision-making capabilities, performing comparably to OpenAI o1 across a range of ophthalmic subspecialty cases, while also offering a substantial reduction in usage costs. These findings highlight the feasibility of utilizing open-weight, reinforcement learning-augmented LLMs as an accessible, cost-effective alternative to proprietary models.