Observational associations do not prove drug-induced inflammation, a harmful laser sequence or device superiority. Randomised findings still depend on the prespecified endpoint, selected population and follow-up.
Terminology and draft status
Terminology review is not complete. Selected actual WHO UMD queries and their limits are recorded in terminology.md; they do not approve the whole translation.
01SOURCE: association is not drug causation
The SOURCE retrospective study compared adults newly starting prostaglandin analogues (PGA), beta-blockers, alpha agonists or topical carbonic anhydrase inhibitors across 10 health systems. Patients with documented previous uveitis were excluded from the main analysis; prescribing selection and coded diagnoses can still confound comparisons. 1 2
At 3 months, reported uveitis incidence was 0.32% with PGA,1.95% with beta-blockers,1.63% with alpha agonists and 1.68% with carbonic anhydrase inhibitors. Adjusted odds were about 5–6 times higher for the other classes than PGA; this does not mean that PGA prevents uveitis or that other drops caused each event. 2
The source itself differs: abstract 67517 patients versus 64517 in results/Table 1. In the latter, class shares were approximately 62%,17.5%,16.7% and 3.7%; mean onset 47 days with PGA and about 38–43 days otherwise, and anatomical categories 84.7% anterior,10.2% posterior,2.6% intermediate. Original rounded 62/17/18/4 and 80/15/5 are not exact source proportions. 1 2
At suspected medication-related uveitis, assess infection, disease history and each drug with an ophthalmologist. Supervised withdrawal or substitution may be appropriate with pressure cover; do not automatically stop and rechallenge every drug, particularly where inflammation is serious. 1 2
| Parameter | Study and boundary |
|---|---|
| Design | Retrospective SOURCE database |
| Population | New adult drop users; prior uveitis excluded |
| Data | 10 health systems; 64517 main text/67517 abstract discrepancy |
| Comparison | PGA vs beta-blockers, alpha agonists, carbonic anhydrase inhibitors |
| Outcome | New uveitis at 3 months; 6 months secondary |
| Source / limit | Chauhan 2025; confounding and source discrepancies |
02ADHD medicines: separate angle closure and open-angle diagnoses
Darwich’s claims-based retrospective cohort with nested case-control analysis included 240257 new users of atomoxetine, amphetamines or methylphenidate, with 4 age-matched controls per case. Mean cohort age was 45 years; this is not a paediatric-only sample, and diagnostic codes do not replace angle examination. 1 3
Regular atomoxetine and amphetamine use was associated with angle-closure diagnoses, adjusted incidence-rate ratios 2.55 and 2.27, while methylphenidate was associated with open-angle glaucoma at 1.23. These relative associations neither quantify an individual’s absolute risk nor exclude angle closure with another agent or overturn a current product contraindication. 3
Original follow-up “up to 1 year” and US-use estimates 10% of children/4% of adults lack matched verification here. Discuss symptoms, anatomy, family history and other drugs; a pressure measurement alone cannot exclude an occludable angle, and medication changes should be coordinated with the prescriber. 1 3
| Parameter | Study and boundary |
|---|---|
| Design | Retrospective cohort with case-control analysis |
| Population | 240257 new ADHD-drug users |
| Data | PharMetrics Plus (IQVIA, USA) |
| Comparison | Atomoxetine, methylphenidate, amphetamines; 4 controls per case |
| Outcome / timing | Angle-closure/open-angle diagnoses; original 1 year limit unconfirmed |
| Source / limit | Darwich 2024; association, not causation or label change |
03GITS: read the primary endpoint before the subgroup
GITS randomised 242 patients with newly detected early or moderate untreated open-angle glaucoma at two Swedish centres to initial intensive medications plus 360°laser trabeculoplasty or stepwise monotherapy. The original 143 versus 155 eye counts need full-report reconciliation with the patient denominator and cannot replace randomised patient numbers. 1 4
At 5 years, median baseline pressure was 24 mmHg in both arms; follow-up pressure was 17 mmHg with monotherapy and 14 mmHg with intensive therapy, rather than 14 in both. Projected end-of-life VFI, the primary outcome, was 79.3% versus 87.1% (P=0.15); median annual losses 0.65 versus 0.25 percentage points (P=0.09). These comparisons were not statistically conclusive. 4
Progression events occurred in 21% versus 11% (P=0.03), with mostly mild adverse events in 25% versus 36%. Greater benefit above the median baseline around 24 mmHg came from post hoc analysis; it is a stratification hypothesis, not a universal 24 mmHg treatment threshold or proof of no benefit below it. Open-label treatment and projected lifetime outcomes limit interpretation. 4
| Parameter | Study and boundary |
|---|---|
| Design | Open-label two-centre randomised trial |
| Population | 242 new early/moderate OAG patients, including PEX in original |
| Comparison | Stepwise monotherapy vs intensive drugs+360°LTP; 143/155 eye counts pending |
| Follow-up | 5 years |
| Outcome | Projected end-of-life VFI primary; progression events secondary |
| Source / limit | Bengtsson 2024; primaryP=0.15, high-IOP subgroup post hoc |
04After SLT: distinguish three active regimens from no treatment
Miranda’s prospective randomised trial included 134 eyes of 134 newly diagnosed POAG patients after primary SLT. It compared prednisolone acetate 1%, prednisolone acetate 0.12% and ketorolac 0.5%, not a placebo or untreated group; average follow-up was 287 days, so 12-month estimates are survival estimates rather than complete 1 year data for every eye. 1 5
The primary abstract reports 12-month success 83.7%,63.9% and 67.0% in that order (P=0.003), not 67.0% for prednisolone 1%. Overall final pressure was 14.39 mmHg from 19.57 mmHg; multivariable failure-related HR for the 1% group was 0.37 (95% CI0.15–0.91). Success definitions, censoring, dose schedule and steroid-response monitoring need the full report before applying the original one-week protocol. 1 5
| Parameter | Study and boundary |
|---|---|
| Design | Prospective randomised active-comparator trial |
| Population | 134 eyes/134 new POAG patients after primary SLT |
| Comparison | Prednisolone acetate 1%/0.12%; ketorolac 0.5%; original 360°/1 week details pending |
| Follow-up | 12 month survival estimate; mean 287 days |
| Outcome | Success 83.7/63.9/67.0%; original≥20%/no surgery-laser-drugs definition pending |
| Source / limit | Miranda 2025; no untreated arm or universal steroid rule |
05LTP before MIGS: identify confounding by treatment history
Mitchell used IRIS registry data 2013–2018 for angle-based MIGS with or without LTP in the preceding two years, with and without phacoemulsification. Propensity matching yielded 954 standalone eyes and 7522 combined eyes from 164965 procedures; medication data were unavailable. 1 6
Prior LTP was associated with further glaucoma surgery over 36 months; the abstract reports adjusted HR 1.53 (CI1.15–2.04). This is a relative reoperation hazard, not a 53 percentage-point risk increase. Severity, prior response and procedure selection can confound results; it does not prove that LTP damages later MIGS or justify universally skipping laser. 6
The original reports similar pressure reduction and higher reoperation risk with secondary glaucoma or very high baseline pressure; these details require matching full-report definitions. Reoperation is influenced by treatment thresholds and access, and is different from visual preservation or a complete medication-adjusted efficacy endpoint. 1 6
| Parameter | Study and boundary |
|---|---|
| Design | Retrospective propensity-matched cohort |
| Data | IRIS registry 2013–2018 |
| Comparison | Angle-MIGS±phaco, with/without LTP in previous 2 years |
| Follow-up | 36 months |
| Outcome | Subsequent glaucoma surgery; adjustedHR 1.53, not absolute 53% increase |
| Source / limit | Mitchell 2024; no medication data, residual confounding |
06Filtering devices: compare endpoints and baseline differences
Armstrong’s 6-centre retrospective nonrandomised series compared 577 eyes of 521 patients: 235 SIBS microshunts (PreserFlo),201 gelatin 45µm microstents (XEN) and 141 trabeculectomies with MMC, followed 12 months. SIBS eyes started with lower decision pressure; medication burden also differed, so adjustment cannot guarantee exchangeable groups. 1 7
For the main 17 mmHg criterion with at least 20% pressure reduction and limits on clinical hypotony, complete success was 68.8% with SIBS,46.2% with XEN and 58.0% with trabeculectomy; qualified success allowed medication and was 89.7%,70.1% and 83.6%. Original 65.3/40.1/52.1% at 6–14 mmHg and medians 12/14/13 mmHg need the matching secondary table and must not be substituted for the primary definition. 1 7
Complications were reported in 33.6%,42.8% and 56% respectively; definitions and intervention burden matter, not just a composite percentage. These findings support discussion, not universal device superiority; the original smaller Mainz comparison and claimed no-revision criterion remain unreconciled, and prospective independent comparisons are still needed. 1 7
| Parameter | Study and boundary |
|---|---|
| Design | Retrospective nonrandomised 6-centre series |
| Population | 577 eyes/521 patients, primary standalone filtration with MMC |
| Comparison | PreserFlo 235, XEN 201, trabeculectomy 141 |
| Follow-up | 12 months |
| Outcome | Primary 17 mmHg; secondary 14/21 mmHg; complete/qualified success, complications |
| Source / limit | Armstrong 2025; different baselines, not causal device ranking |
07Normal-tension glaucoma: a small uncontrolled series
Pillunat prospectively followed 42 White European POAG eyes after primary standalone PreserFlo: 29 high-pressure and 13 normal-pressure eyes, for one year. The median medicated mean 24 hour pressure in the normal-pressure group fell from 14.3 to 9.5 mmHg; 12.8 was the high-pressure group’s postoperative value, not the normal-pressure baseline. 1 8
At 1 year,92.5% of all eyes were medication-free; this overall denominator must not become an exact 92% estimate for the 13 normal-pressure eyes. Peak pressure and fluctuations decreased without statistically clear between-group differences; non-significance in a small series does not establish equivalent efficacy or safety. 8
Original targets 6–14/6–18 mmHg and an additional 30% normal-pressure reduction require their precise protocol definitions. Without randomisation or a trabeculectomy comparator, this is an early pressure-profile study, not proof of superior visual outcomes, no hypotony risk or a universal replacement for trabeculectomy. 1 8
| Parameter | Study and boundary |
|---|---|
| Design | Prospective uncontrolled single-centre case series |
| Population | 42 eyes: 29 high-pressure/13 normal-pressure POAG |
| Intervention | Primary standalone PreserFlo |
| Follow-up | 12 months |
| Outcome | 24 hour mean/peak/fluctuation; original 6–14/6–18 and≥30% definitions pending |
| Source / limit | Pillunat 2025; small normal-pressure sample, no trabeculectomy comparator |
What are the main learning points?
Observational associations do not prove drug-induced inflammation, a harmful laser sequence or device superiority. Randomised findings still depend on the prespecified endpoint, selected population and follow-up.
When is assessment urgent?
Severe eye pain, redness, rapid visual loss, halos with headache or nausea, marked inflammation or suspected infection or neurological disease needs urgent assessment; do not stop glaucoma or ADHD medication on your own.
References and access status
- 1Immutable original export; seven-study2025lecture; Thomas Klink/DOG selection and approximately2,000-paper pool not independently verified
Complete export read; original attribution and any recording/transcript or appointment not independently verified.
- 2Chauhan et al.,2025 SOURCE
Parent read the complete primary abstract14November2026; SOURCE also selected results/Table1/Table2, ADHD indexed abstract with direct-page failure. Full reports, protocols, figures, secondary definitions, current local labels and source discrepancies remain unresolved.
- 3Darwich et al.,2024 ADHD medications
Parent read the complete primary abstract14November2026; SOURCE also selected results/Table1/Table2, ADHD indexed abstract with direct-page failure. Full reports, protocols, figures, secondary definitions, current local labels and source discrepancies remain unresolved.
- 4Bengtsson et al.,2024 GITS five years
Parent read the complete primary abstract14November2026; SOURCE also selected results/Table1/Table2, ADHD indexed abstract with direct-page failure. Full reports, protocols, figures, secondary definitions, current local labels and source discrepancies remain unresolved.
- 5Miranda et al.,2025 anti-inflammatory treatment after SLT
Parent read the complete primary abstract14November2026; SOURCE also selected results/Table1/Table2, ADHD indexed abstract with direct-page failure. Full reports, protocols, figures, secondary definitions, current local labels and source discrepancies remain unresolved.
- 6Mitchell et al.,2024 IRIS angle-MIGS after LTP
Parent read the complete primary abstract14November2026; SOURCE also selected results/Table1/Table2, ADHD indexed abstract with direct-page failure. Full reports, protocols, figures, secondary definitions, current local labels and source discrepancies remain unresolved.
- 7Armstrong et al.,2025 filtration comparison
Parent read the complete primary abstract14November2026; SOURCE also selected results/Table1/Table2, ADHD indexed abstract with direct-page failure. Full reports, protocols, figures, secondary definitions, current local labels and source discrepancies remain unresolved.
- 8Pillunat et al.,2025 PreserFlo HPG/NPG
Parent read the complete primary abstract14November2026; SOURCE also selected results/Table1/Table2, ADHD indexed abstract with direct-page failure. Full reports, protocols, figures, secondary definitions, current local labels and source discrepancies remain unresolved.
For patients
Studies help your clinician choose between drops, laser and surgery, but averages and relative risks do not predict your individual response. Treatment changes require a plan that preserves pressure control and considers other illnesses.
Examination and selected tests guide individual treatment and follow-up; this educational text cannot replace your own assessment.
Severe eye pain, redness, rapid visual loss, halos with headache or nausea, marked inflammation or suspected infection or neurological disease needs urgent assessment; do not stop glaucoma or ADHD medication on your own.
Severe eye pain, redness, rapid visual loss, halos with headache or nausea, marked inflammation or suspected infection or neurological disease needs urgent assessment; do not stop glaucoma or ADHD medication on your own.