Diagnostic Accuracy of Artificial Intelligence-Based Automated Diabetic Retinopathy Screening in Real-World Settings: A Systematic Review and Meta-Analysis.

PURPOSE: To evaluate the diagnostic accuracy of artificial intelligence (AI)-based automated diabetic retinopathy (DR) screening in real-world settings. DESIGN: Systematic review and meta-analysis METHODS: We conducted a systematic review of relevant literature from January 2012 to August 2022 using databases including PubMed, Scopus and Web of Science. The quality of studies was evaluated using Quality Assessment for Diagnostic Accuracy Studies 2 (QUADAS-2) checklist. We calculated pooled accuracy, sensitivity, specificity, and diagnostic odds ratio (DOR) as summary measures. The study protocol was registered with the International Prospective Register of Systematic Reviews (PROSPERO - CRD42022367034). RESULTS: We included 34 studies which utilized AI algorithms for diagnosing DR based on real-world fundus images. Quality assessment of these studies indicated a low risk of bias and low applicability concern. Among gradable images, the overall pooled accuracy, sensitivity, specificity, and DOR were 81%, 94% (95% CI: 92.0-96.0), 89% (95% CI: 85.0-92.0) and 128 (95% CI: 80-204) respectively. Sub-group analysis showed that, when acceptable quality imaging could be obtained, non-mydriatic fundus images had a better DOR of 143 (95% CI: 82-251) and studies using 2 field images had a better DOR of 161 (95% CI 74-347). Our meta-regression analysis revealed a statistically significant association between DOR and variables such as the income status, and the type of fundus camera. CONCLUSION: Our findings indicate that AI algorithms have acceptable performance in screening for DR using fundus images compared to human graders. Implementing a fundus camera with AI-based software has the potential to assist ophthalmologists in reducing their workload and improving the accuracy of DR diagnosis.

Closure rate and recovery of subfoveal microstructures following conventional internal limiting membrane peeling versus per fluoro octane-assisted inverted flap for large macular holes - A randomized controlled trial (InFlap Study).

PURPOSE: To analyze the outcomes following conventional internal limiting membrane (ILM) peeling versus perfluoro octane-assisted inverted flap technique for large macular holes (MH). METHODS: A consecutive 99 eyes of 99 patients were enrolled {45 - conventional group and 54 - inverted flap (InFlap) group}. The primary outcome was a difference in closure rate. Secondary outcomes were differences in best-corrected visual acuity (BCVA), restoration of external limiting membrane (ELM) and ellipsoid zone (EZ) between groups at 3 (primary endpoint), 6 and 12 (secondary endpoints) months. Additionally, the effect of different gas tamponades on closure rates, ILM flap disintegration in InFlap group, and subfoveal thickness (SFT) between groups in closed. RESULTS: At 3 months, there was no difference in the closure rate and BCVA between groups. At six months, closure rate was significantly better in the InFlap group. However, this difference was not maintained at 12 months. There was no difference in BCVA between groups at any visit. The ELM recovery was significantly higher in the conventional group at three months; however, there was no difference in ELM/EZ recovery between groups at other visits. The closure rate in the InFlap group was the same irrespective of gas tamponade. The ILM flap was identifiable in one-third of patients at 12 months. In closed MH, SFT was significantly more in InFlap group. CONCLUSION: The closure rate and visual outcomes remained similar in both groups in the immediate and long term. Conventional ILM peeling technique seems to have early ELM recovery when compared to inverted flap technique.

Incidence, outcomes, and risk of surgical failure after development of early hypotony following Aurolab aqueous drainage implant surgery.

Purpose: To report the incidence, outcomes, and risk of surgical failure after early postoperative hypotony following Aurolab Aqueous Drainage Implant (AADI) surgery for adult and pediatric refractory glaucoma. Methods: Medical records of patients who underwent AADI between January 2013 and March 2017 with a minimum of 2-years follow-up were retrospectively reviewed. Early postoperative hypotony was defined as IOP ≤5 mmHg within the first 3 months after AADI. Surgical failure of AADI was defined as IOP >21 mmHg or reduced <20% below baseline on two consecutive follow-up visits after 3 months, IOP ≤5 mmHg on two consecutive follow-up visits after 3 months, reoperation for glaucoma or a complication, or loss of light perception vision. Results: Early postoperative hypotony was seen in 15/213 eyes (7%) in the adult group and in 6/101 eyes (6%) in the pediatric group. The onset of hypotony was significantly earlier in the pediatric group (median = 39 days post AADI, IQR = 20-58 days) compared with adult eyes (median = 51 days post AADI, IQR = 30-72 days) (P = 0.02). Eyes with early postoperative hypotony did not have an increased risk of cumulative surgical failure as compared with eyes without hypotony in both adult (33.3% vs. 23.7%; P = 0.48) and pediatric (33.3% vs. 13.7%; P = 0.16) refractory glaucoma. All eyes recovered from hypotony, though one adult eye developed retinal detachment and one pediatric eye developed corneal decompensation and lost vision. Conclusion: Early postoperative hypotony was an infrequent complication post AADI and occurred earlier in pediatric eyes. Early postoperative hypotony did not increase risk of surgical failure up to 2 years.

Outcomes of the Aurolab aqueous drainage implant and trabeculectomy with mitomycin C in iridocorneal endothelial syndrome.

PURPOSE: To compare the surgical outcomes of the Aurolab aqueous drainage implant (AADI) and trabeculectomy with mitomycin C (MMC) in patients with glaucoma secondary to iridocorneal endothelial (ICE) syndrome. MATERIALS AND METHODS: This retrospective comparative case series included 41 eyes of 41 patients with ICE syndrome and glaucoma who underwent either a trabeculectomy with MMC (n = 20) or AADI surgery (n = 21) with a minimum of 2 years follow-up. Outcome measures included intraocular pressure (IOP), the use of glaucoma medications, visual acuity, additional surgical interventions, and surgical complications. Surgical failure was defined as IOP > 21 mmHg or reduced < 20% from baseline, IOP ≤ 5 mmHg, reoperation for glaucoma or a complication, or loss of light perception vision. RESULTS: The cumulative probability of failure at 2 years was 50% in the trabeculectomy group (95%CI = 31-83%) and 24% in the AADI group (95%CI = 11-48%) (p = 0.09). The IOP was consistently lower in the AADI group compared with the trabeculectomy group at 6 months and thereafter. Surgical complications occurred in 13 eyes (65%) in the trabeculectomy group and 12 eyes (57%) in the AADI group (p = 0.71). Reoperations for glaucoma or complications were performed in 12 eyes (60%) in the trabeculectomy group and 5 patients (24%) in the tube group (p = 0.06). Cox proportional hazards showed that AADI had a 53% lower risk of failure at 2 years (p = 0.18; HR = 0.47; 95%CI = 0.16-1.40). CONCLUSION: AADI surgery achieved lower mean IOPs than trabeculectomy with MMC in managing glaucoma secondary to ICE syndrome. A trend toward lower rates of surgical failure and reoperations for glaucoma and complications was observed following AADI placement compared with trabeculectomy with MMC in eyes with ICE syndrome.

Intermediate-term outcomes of the Aurolab aqueous drainage implant in neovascular glaucoma.

PURPOSE: To assess the outcomes of the non-valved Aurolab aqueous drainage implant (AADI) in neovascular glaucoma (NVG). METHODS: Data of consecutive patients with NVG who underwent AADI and had a minimum follow-up of 2 years were included. The primary outcome measure was the cumulative rate of surgical failure defined as intraocular pressure (IOP) >21 mm Hg or reduced <20% below baseline, IOP ≤5 mm Hg, reoperation for glaucoma or a complication, or loss of light perception vision. RESULTS: We included 85 eyes of 85 patients with NVG, with a mean age of 61.2±9.3 years. The most common aetiologies were proliferative diabetic retinopathy (n=43) and central retinal vein occlusion (n=24). The mean IOP decreased from 36.8±12.5 mm Hg at baseline to 15.8±7.5 mm Hg at 2-year follow-up (p<0.001) and the number of IOP-lowering medications reduced from 3.4±0.8 to 1.5±1.1 (p<0.001). The cumulative rate of failure increased from 3.1% (95% CI 1.1% to 11.8%) at 1 year to 33.8% (95% CI 20.4% to 52.5%) at 2 years. Multivariable analysis showed that eyes with open angles had a lower risk of failure (HR 0.17, 95% CI 0.10 to 1.03, p=0.09). The logarithm of minimum angle of resolution visual acuity declined from 0.98±0.7 to 1.8±1.0 at 2 years (p<0.001). CONCLUSION: Approximately one-third of NVG eyes that received the AADI failed after 2 years of follow-up similar to other series. Early AADI implantation at the open angle stage of NVG may yield better results.

Correlation of pattern of visual field loss by perimetry and anterior chamber angle parameters by anterior segment optical coherence tomography in primary angle-closure glaucoma.

Purpose: To correlate and analyze the pattern of the visual field (VF) defects by perimetry and anterior chamber angle parameters by AS-OCT in primary angle-closure glaucoma (PACG) across varied severity levels on presentation to a tertiary eye care center. Methods: This was a cross-sectional study, which included 323 eyes of clinically diagnosed cases of PACG. Glaucoma severity was categorized according to mean deviation (MD) as mild (-6.00 dB or more), moderate (-6.01 to -12.00 dB), and severe (-12.01 to -30.00 dB). AS-OCT measured the nasal (N) and temporal (T) angle opening distance at 500 µm (AOD 500) and 750 µm (AOD 750), anterior chamber angle (ACA), lens vault (LV), and anterior chamber width (ACW). The VF severity was then correlated with the AS-OCT parameters using statistical analysis. Results: The mean age ± standard deviation (SD) of the patients included in the study was 56.03 ± 8.6 years, with a 1:1.2 gender ratio. The number of eyes with mild, moderate, and severe VFs were 140 (43.3%), 88 (27.24%), and 95 (29.41%), respectively. There was no statistically significant correlation in the mean anterior chamber angle parameters (AOD 500, AOD 750, ACA 500, ACA 750, LV, ACW, and axial length (AL)) among the groups. However, the correlation between AOD 500 and LV thickness was found to be significant (P = 0.0000) with a negative Spearman's rank correlation coefficient (r = -0.3329). Conclusion: The ACA parameters obtained by AS-OCT along the horizontal axis after elimination of pupillary block by laser peripheral iridotomy do not correlate and cannot be used to assess the disease severity of PACG.

Comparison of uveal parameters between acute primary angle-closure eyes and fellow eyes in South Indian population.

Purpose: : To analyze the ocular biometric parameters of eyes with acute primary angle closure (APAC) as compared to fellow eyes. Methods: : A cross-sectional study was conducted on 27 patients presenting with recent onset APAC to a tertiary eye institute in India. Anterior and posterior ocular biometric parameters were measured simultaneously by anterior segment optical coherence tomography (AS-OCT), A-scan, ultrasound biomicroscopy (UBM), and B-mode ultrasonogram (USG). The parameters measured were anterior chamber depth (ACD), anterior chamber angle (ACA), angle opening distance (AOD500, AOD750), lens vault (LV), axial length (AL), ciliary body thickness maximum (CBTmax) and at the point of scleral spur (CBT0), anterior placement of the ciliary body (APCB), and retinochoroidal thickness (RCS). Results: Mean age ± SD of patients with APAC was 55.66 ± 7.2 years with female preponderance (21:6 patients). Mean presenting IOP ± SD of the affected eye and fellow eye were 54.74 ± 11.67 mm Hg and 18.7 ± 11.67 mm Hg, respectively. Eyes with APAC had statistically significant narrower anterior ocular biometric parameters, higher LV, decreased ciliary body thickness, more APCB, and longer AL than the fellow eyes. CBTmax is the only variable that had significance (ß = -0.421,95% CI: -0.806 to - 0.035, P = 0.034) in the univariate analysis with RCS thickness in APAC eyes. Further, there was a correlation between CBT0 and APCB with CBTmax both in univariate (ß = 0.894, P < 0.0001 and ß = -0.351, P = 0.039) and multivariable analysis (ß = 0.911, P < 0.0001 and ß = -0.416, P = 0.016). Conclusion: Compared to the fellow eyes, APAC eyes had different ocular biometric parameters. In addition to known biometric parameters associated with pupillary block (narrower anterior biometric parameters-ACA, ACD, and AOD), our study found multiple nonpupillary block factors such as higher lens vault and thinner and more anteriorly placed ciliary body to be associated with APAC.

Intraocular Pressure Control Following Phacoemulsification in Eyes With Pre-existing Aurolab Aqueous Drainage Implant.

PURPOSE: The aim was to investigate intraocular pressure (IOP) control after phacoemulsification in adult glaucomatous eyes with a functioning nonvalved Aurolab Aqueous Drainage Implant (AADI) compared with eyes that did not have cataract extraction post-AADI. METHODS: In this retrospective study, we reviewed records of 47 patients (47 eyes) who had a clear corneal phacoemulsification after AADI placement with a minimum of 2 years of follow up. The control group included 89 patients (89 eyes) who had a functional AADI at 1 year, minimum of 3 years of follow up post-AADI implantation, and no cataract extraction. The main outcome measure was failure (IOP >21 mm Hg or increased by >20% from prephacoemulsification level requiring at least 1 additional glaucoma medication, IOP ≤5 mm Hg, reoperation for glaucoma, or loss of light perception vision). RESULTS: The median interval between AADI and phacoemulsification was 11.5 months (range: 4 to 68 mo), and the mean follow-up time after phacoemulsification was 35.6±6.4 months. The cumulative probability of failure was 14% (95% confidence interval=6%-31%) in the phaco group and 6% (95% confidence interval=3%-13%) in the control group at 2 years (P=0.11). Mean IOP was reduced from 16.5±4.5 mm Hg preoperatively to 15.4±4.7 mm Hg at 2 years after phacoemulsification (P=0.10). Mean LogMAR visual acuity improved from 1.1±0.6 preoperatively to 0.6±0.7 at 2 years after phacoemulsification (P<0.001). CONCLUSIONS: In eyes with a pre-existing AADI, phacoemulsification resulted in visual improvement without a significant rise in IOP or increased risk of AADI failure after 2 years follow up.

Screening first-degree relatives of glaucoma patients reveals barriers to participation.

PURPOSE: To report the results of a glaucoma screening campaign targeting first-degree relatives of glaucoma patients in South India. METHODS: 1598 glaucoma patients were contacted via letter or letter and phone call and asked to bring their siblings and children to a glaucoma screening. Participants underwent standardised eye examinations and completed questionnaires that assessed barriers to participation and awareness of glaucoma risk. Two-proportion z-tests were used to compare categorical data. Costs associated with the screening were recorded. RESULTS: 206 probands (12.9%) attended the screening along with 50 siblings and children. Probands were nearly twice as likely to attend if they had been contacted via both letter and phone call rather than letter only. Over half of probands reported that their relatives could not participate because they did not live in the region, and one-fifth reported that their relatives had other commitments. Fifty-eight per cent of the siblings and children who attended did not know that they were at increased risk for glaucoma due to their family history, and 32.0% did not know that the relative who had invited them to the screening had glaucoma. Thirteen siblings and children (26.0% of those who attended) were found to have findings concerning for glaucoma. The average cost per first-degree relative who was screened was INR2422 (£26). CONCLUSION: Participation in this glaucoma screening campaign was poor. The major barrier to participation was distance from the screening site and associated indirect costs. Better strategies for bringing first-degree relatives in for examinations are needed.

Community Health Workers for Prevention of Corneal Ulcers in South India: A Cluster-Randomized Trial.

PURPOSE: To determine whether a community health worker (CHW) program increases referrals to local eye care providers and ultimately reduces the incidence of corneal ulcers. DESIGN: Cluster-randomized trial performed from 2014 to 2017 in rural South India. METHODS: This was a community-based study that included all inhabitants of 42 rural South Indian communities. CHWs were trained to diagnose corneal abrasions and assist participants in seeking care at a local vision center. Given the nature of the intervention, the trial was not masked. The main outcome measure was incident corneal ulcer, defined as an active infiltrate or evidence of a new opacity, as assessed by means of penlight examination during an annual door-to-door census. RESULTS: Twenty-one study clusters were randomized to the CHW intervention and 21 to no intervention. Vision centers diagnosed 195 corneal abrasions from the intervention clusters during the 2-year study (rate, 223 per 100,000 person-years; 95% CI, 28-1743) and 62 from the control clusters (rate, 62 per 100,000 person-years; 95% CI, 8-496; incidence rate ratio, 3.57; 95% CI, 2.01-6.35; P < .001). The estimated incidence of corneal ulceration during the study period was 60 per 100,000 person-years (95% CI, 25-141) in the intervention group and 32 per 100,000 person-years (95% CI, 13-80) in the control group (incidence rate ratio, 1.86; 95% CI, 0.5-6.4; P = .32). CONCLUSIONS: A CHW program resulted in 3.5 times more referrals to local eye care providers for corneal abrasions, but no difference could be detected in the incidence of corneal ulceration. CHW programs provide a mechanism for increasing referrals to eye hospitals. TRIAL REGISTRATION: ClinicalTrials.gov ID: NCT02284698.

Identification of risk factors for targeted diabetic retinopathy screening to urgently decrease the rate of blindness in people with diabetes in India.

PURPOSE: There is an exponential rise in the prevalence of diabetes mellitus (DM) in India. Ideally all people with DM should be periodically screening for diabetic retinopathy (DR) but is not practical with current infrastructure. An alternate strategy is to identify high-risk individuals with vision-threatening diabetic retinopathy (VTDR) for priority screening and treatment. METHODS: We reanalyzed four population-based studies, conducted in South India between 2001 and 2010, and reclassified individuals above 40 years into known and newly diagnosed diabetes. Multiple regression analysis was done to identify risk factors in people with known and new DM. RESULTS: The prevalence of DR in 44,599 subjects aged ≥40 years was 14.8% (18.4 and 4.7% in known and new DM, respectively), and the prevalence of VTDR was 5.1%. Higher risk factors of VTDR were older age >50 years, diabetes duration >5 years, and systolic blood pressure >140 mm Hg. Targeted screening of people with diabetes using high-risk criteria obtained from this study was able to detect 93.5% of all individuals with VTDR. CONCLUSION: In a limited resource country like India, a high-risk group-based targeted screening of individuals with DM could be prioritized while continuing the current opportunistic screening till India adopts universal screening of all people with DM.

Urgent and emergent glaucoma care during the COVID-19 pandemic: An analysis at a tertiary care hospital in South India.

PURPOSE: To describe the demographic profiles, clinical characteristics, and clinical outcomes of patients presenting with glaucoma emergencies during the COVID-19 lockdown in India. METHODS: This retrospective, cross-sectional, observational case series involved review of medical records of all patients presenting to the glaucoma service during the COVID-19 lockdown period and comparison with the previous year (March 23 to June 23, 2020 Vs 2019) in a tertiary center in India. RESULTS: We found a 78.9% reduction in overall outpatient visits (54,345 vs. 257,339; P < 0.001) and 80.9% reduction in the number of glaucoma outpatient visits (4,788 vs. 25,083; P < 0.001). Additionally, the proportion of true glaucoma emergency visits significantly increased by 62.4% in 2020 Vs 2019 (1,408/4,788 (29.4%) vs. 4,542/25,083 (18.1%); P < 0.001). Lens-induced glaucomas were the most common glaucoma surgical emergency (13.4%) in 2020. Moreover, comparison of procedures demonstrated a proportionate decrease in incisional glaucoma surgeries (70/115 (60.86%) vs. 806/939 (85.83%); P < 0.001) and an increase in the proportion of emergency cataract surgeries (129/475 (27.15%) vs. 170/2715 (6.26%); P < 0.001) and transscleral cyclophotocoagulation (45/115 (39.13%) vs. 133/939 (14.16%); P = 0.0001) during 2020 vs. 2019. CONCLUSION: Our study demonstrated a 62% increase in the proportion of visits that were true glaucoma emergencies. Additionally, the proportions of emergency cataract surgeries increased by 4.3 times and the proportion of transscleral cyclophotocoagulation increased by 2.8 times during the pandemic. More nonincisional procedures and less diagnostic testing were performed to minimize postoperative visits and virus transmission. Further understanding of the profile of emergencies may help in developing novel strategies to anticipate future challenges in managing glaucoma care during subsequent waves of the pandemic.

Impact of COVID-19 on follow-up and medication adherence in patients with glaucoma in a tertiary eye care centre in south India.

PURPOSE: COVID-19 pandemic has affected the healthcare system worldwide hindering the continuum of treatment of chronic disease patients. The objective of the study is to analyze the barriers encountered by the glaucoma patients for the follow-up visit and medication adherence during the pandemic. METHODS: This cross-sectional study included glaucoma patients who did not attend the scheduled appointment from April 1, 2020 to July 31, 2020 in a tertiary eye care center (88.21%). Eligible patients of age >18 years and advised antiglaucoma medication in Madurai Zone were interviewed with validated questionnaire through telephonic call. RESULTS: 363 patients answered the questionnaire through telephonic interview. 57.3% of the patients were found to be non-adherent to medication. The main barriers for glaucoma follow-up visit during the pandemic were lockdown restriction, transport problem, and financial difficulties. The top barriers for medication adherence were non availability of medication (54.81%), financial difficulties (30.29%), did not feel much improvement with eye drops (20.19%). On multiple regression analysis, longer distance to hospital, low socioeconomic status, more than one antiglaucoma medication use, lack of awareness of glaucoma, non-complaint before COVID-19 and stress due to the pandemic were found to be significant factors for medication non adherence. CONCLUSION: COVID-19 pandemic has emphasized the need for reformation in health care system for accessibility of medical care to patients in rural areas. Decentralization of health system to primary care level and utilization of teleophthalmology should be considered by health care planners in future.

Comparison of surgical repositioning rates and outcomes for hydrophilic vs hydrophobic single-piece acrylic toric IOLs.

PURPOSE: To evaluate the comparative rotational stability and the repositioning rates and outcomes of a hydrophilic and hydrophobic acrylic monofocal toric intraocular lens (IOLs). SETTING: Single center, Aravind Eye Hospital in Madurai, India. DESIGN: Retrospective cohort study. METHODS: All patients had either an Auroflex hydrophilic acrylic toric IOL (n = 4603) or AcrySof hydrophobic acrylic toric IOL (n = 926) implanted between January 2015 and April 2019. Surgical repositioning was offered if the toric IOL was 15 degrees or greater misaligned or there was a 2-line or greater improvement in uncorrected distance visual acuity (UDVA) predicted with repositioning. RESULTS: A total of 5529 single-piece toric IOLs were implanted during the study period. There was no statistically significant difference between the hydrophilic and hydrophobic acrylic toric IOL groups in need for surgical repositioning (2.5% vs 1.9%; P = .34) or actual surgical repositioning rate (1.8% vs 1.5%; P = .59). A higher rate of surgical toric IOL repositioning was associated with younger age (62.7 vs 67.2 years; P < .0001), higher presurgical astigmatism (2.32 vs 1.99 diopters [D]; P < .0001), and with-the-rule astigmatism (4.7% vs 2.1% against-the-rule astigmatism; P = .002). In 95 eyes that underwent toric IOL repositioning, there was significant preoperative to postoperative improvement in UDVA (0.48 vs 0.12 logarithm of the minimum angle of resolution; P < .001), mean residual cylinder (1.79 vs 0.7 D; P < .001), and mean toric IOL misalignment (49.7 vs 5.4 degrees; P < .001). The repositioning outcomes were statistically similar with both toric IOL models. There were no major complications or recurrent toric IOL rotation after surgical repositioning. CONCLUSIONS: The hydrophilic acrylic toric IOL had comparable rates of postoperative misalignment and surgical repositioning with the hydrophobic acrylic toric IOL. Early postoperative surgical repositioning was both safe and effective with both toric IOLs.

Surgical Outcomes of Superotemporal Versus Inferonasal Placement of Aurolab Aqueous Drainage Implant in Refractory Pediatric Glaucoma.

PURPOSE: We sought to describe the outcomes of the Aurolab aqueous drainage implant (AADI) placed in the superotemporal (ST) versus the inferonasal (IN) quadrant in pediatric eyes with refractory glaucoma. DESIGN: Retrospective comparative interventional case series. METHODS: This was a retrospective study of patients ≤18 years of age who underwent AADI implantation and completed a minimum of 2-year follow-up. The choice of the quadrant depended upon the amount of scarring and conjunctival mobility. Cumulative success at 2 years was defined as intraocular pressure (IOP) ≤21 mm Hg or reduced by ≥20% below baseline on 2 consecutive follow-up visits after 3 months, IOP ≤5 mm Hg on 2 consecutive follow-up visits after 3 months, reoperation for glaucoma or a complication, or loss of light perception vision. RESULTS: A total of 144 patients (144 eyes) underwent AADI placement, including 48 eyes (33%) in the IN and 96 eyes (67%) in the ST quadrants. The IOP was significantly higher in the IN group (17.5 ± 7.4 mm Hg vs 13.7 ± 6.2 mm Hg, P = .005) with a greater number of medications (1.5 ± 1.0 vs 0.8 ± 0.9, P = .001) after 2 years of follow-up. Cumulative success rates at 2 years were 50.7% (95% confidence interval 35.4%-63.9%) in the IN group and 65.6% (95% confidence interval 56.5%-75.7%) in the ST group (P = .15). Complications occurred more frequently in the IN group, with significantly more tube exposure (12% vs 0%, P = .05). CONCLUSIONS: Placement of the AADI in the ST quadrant has better IOP-related outcomes and is a safer surgical option in pediatric eyes compared with the IN quadrant. It may be prudent to avoid AADI in the IN quadrant in children unless the ST location is not a viable option.