A recent Ophthalmology digest,
as sent to subscribers

Study Type: Systematic review and meta-analysis (33 studies, 5,235 eyes)

Key Question: Is the incidence of de novo glaucoma and ocular hypertension (OHT) lower after DMEK than DSAEK in glaucoma-naïve adult eyes?

Key Findings:

• Pooled incidence, DMEK vs DSAEK: glaucoma 3% vs 11%; OHT 9% vs 34%; medication initiation/escalation 6% vs 20%. Glaucoma surgery did not differ (1.0% vs 2.5%, p=0.14).
• Direction persisted after adjusting for follow-up (DSAEK 68 vs DMEK 35 months): OHT rate ratio 0.31 (0.14–0.71); glaucoma 0.45 (0.19–1.07, non-significant). Higher steroid exposure independently predicted OHT.
• Glaucoma defined by optic nerve/visual field criteria: 5.0% vs 8.1% (p=0.38). Head-to-head OR 0.51 (0.22–1.21). GRADE certainty low to very low.

Clinical Relevance: OHT is common after endothelial keratoplasty, so structured IOP surveillance and steroid-tapering protocols are warranted regardless of technique, and DMEK's glaucoma advantage should not be assumed.

Limitations: Heterogeneous glaucoma definitions, steroid confounding, and randomised evidence restricted to ultrathin/microthin DSAEK limit attribution of effect to technique.

Study Type: Multicentre retrospective cohort study (8 tertiary centres, South Korea; n=126 with biallelic EYS variants).

Key Question: What are the genetic spectrum and quantitative rates of structural and functional progression in EYS-associated retinal degeneration?

Key Findings:

• EZ width declined by 124.1 μm/year (95% CI 106.4–141.8; ~4.6% relative loss/year); √EZ area fell by 0.105 mm/year (95% CI 0.089–0.120); BCVA worsened by 0.043 logMAR/year (95% CI 0.025–0.060). Decline was linear over the observed interval.
• 66 unique variants identified (20 novel); structural variants in 8.7%, including a deep-intronic cryptic-exon variant (c.1766+2635G>A).
• No difference in progression between biallelic loss-of-function and ≥1 missense allele genotypes (P=0.923).

Clinical Relevance: Provides natural history benchmarks usable as reference endpoints for EYS gene therapy trials, and supports including structural and deep-intronic variant detection in genetic testing pathways.

Limitations: Retrospective design with variable follow-up (median 5.2 years, IQR 1.2–9.0), and serial OCT was analysable in only 84 of 126 patients; the single-ethnicity cohort may limit generalisability to UK populations.

Study Type: Retrospective comparative cross-sectional and clinical cohort study (n=140; 54 decompensated esophoria [DE], 86 sagging eye syndrome [SES]).

Key Question: Can the discrepancy between Hess screen test (HST) and prism cover test (PCT) esodeviation distinguish DE from SES in acquired esotropia?

Key Findings:

• In DE, HST esotropia (24.6 ± 11.8Δ) significantly exceeded distance PCT (18.4 ± 9.9Δ; p<0.0001). In SES, the two were near-identical (12.2Δ vs 11.9Δ; p=0.61).
• An HST-PCT discrepancy ≥10Δ gave a 91% posterior probability of DE.
• After DE surgery, residual discrepancy ≥10Δ was associated with 80% reoperation versus 14% with smaller discrepancy (p=0.008).

Clinical Relevance: HST is widely available in UK orthoptic clinics, so this offers a low-cost bedside method to differentiate central vergence disorder from connective tissue degeneration, potentially guiding surgical planning and counselling.

Limitations: Retrospective single-centre-style design with unclear diagnostic reference standard, and no external validation; the postoperative reoperation subgroup is likely small.

Study Type: Systematic review and meta-analysis (PROSPERO-registered; 21 unique studies, 9 nonpregnancy cohorts in the prevalence analysis, 455 tested participants).

Key Question: What is the procedure-specific prevalence of biochemical vitamin A deficiency after bariatric surgery, and what are the ocular manifestations, latency to presentation, and response to repletion?

Key Findings:

• Prevalence by procedure: RYGB 18.0%, BPD 59.0%, BPD-DS 37.8%, JIB 42.5%. Post-RYGB pregnancy: 63.3% (95% CI 43.9-80.1). The pooled 26.1% (95% CI 18.5-35.4) across procedures is a secondary estimate only.
• Night blindness prevalence was 60.1% in active cohorts. Median latency to ocular presentation was 5 years post-RYGB versus 14-27 years after BPD-DS, JIB, or revisional procedures (19 patients).
• 13 of 14 treatment-response records (93%) showed ocular improvement or resolution.

Clinical Relevance: Late presentation, potentially decades after surgery, supports a low threshold for serum retinol testing in post-bariatric patients with nyctalopia, ocular surface disease, or unexplained visual decline.

Limitations: Substantial heterogeneity, few cohorts per procedure subgroup, and selective reporting of ocular outcomes limit certainty; the treatment-response data rest on only 14 records.

Study Type: Multicentre retrospective cohort study (139 eyes, 126 patients) with an independent comparison cohort (55 eyes).

Key Question: Can unsupervised clustering of baseline OCT and functional variables identify structural phenotypes of chronic CSC, and do these add to photoreceptor staging for prognosis?

Key Findings:

• Three phenotypes emerged: pachychoroid-dominant (26%), transitional (57%) and degenerative (17%). The degenerative group gained 0.33 logMAR less than pachychoroid-dominant eyes after verteporfin PDT (95% CI 0.23–0.44) and had lower SRF resolution (57% vs 83%; HR 0.39).
• Phenotype added nothing to photoreceptor stage for visual recovery (marginal R² 0.575 to 0.590, P = .23) or SRF resolution (P = .45).
• Pachychoroid-dominant and transitional eyes were indistinguishable by photoreceptor staging but differed in SCT (583 vs 397 µm) and age (45.5 vs 52.0 years). The age gradient replicated in the independent cohort; the choroidal separation was attenuated.

Clinical Relevance: Visual prognosis after PDT is determined by photoreceptor integrity, supporting photoreceptor staging on OCT as the key prognostic assessment when counselling chronic CSC patients.

Limitations: Retrospective design and attenuated choroidal separation in the independent cohort limit generalisability; the 520-µm threshold applies only to pachychoroid-confirmed disease.

Study Type: Cross-sectional baseline analysis from an ongoing prospective observational study (87 eyes, 87 participants).

Key Question: Which structural imaging metrics best correlate with which functional outcomes in early atrophic AMD?

Key Findings:

• Lesion size metrics tracked localised retinal sensitivity loss: each doubling of targeted GA size reduced mean targeted mesopic and dark-adapted sensitivity by 3.55 dB and 3.65 dB (n=38, P<0.001); doubling of hypertransmission size reduced them by 1.65 dB and 1.97 dB (n=81, P<0.001; ρ up to -0.66).
• Central function depended on foveal proximity: each mm of greater distance from the fovea was associated with ~4–5 letters better BCVA, LLVA and MAT, and better reading acuity (-0.11 logRAD/mm; adjusted P<0.05). Lesion size showed no consistent association.

Clinical Relevance: Supports lesion-targeted fundus-controlled perimetry as a sensitive endpoint for early atrophic AMD, where BCVA may remain insensitive to progression, relevant to trial design and future monitoring.

Limitations: Cross-sectional design with univariable analyses only precludes causal or prognostic inference, and the GA subgroup is small (n=38).

Study Type: Retrospective multicentre cohort study (9 US academic tertiary centres, 2012–2024)

Key Question: How do presentation, management and outcomes of culture-positive endogenous endophthalmitis differ by causative organism?

Key Findings:

• 305 eyes (261 patients): *S. aureus* 37.7%, fungal 19.3%, gram-negative 15.4%, *Streptococcus* 13.4%. Blood cultures were positive in 66.2% of eyes versus 31.1% for vitreous and 8.9% for aqueous; blood–eye concordance was only 62% (24/39).
• Fungal cases presented with better VA (logMAR 1.30 vs 1.78–2.14 for bacteria, p<0.001); final VA differences were not significant (p=0.06). Retinal detachment was commonest with *S. aureus* and *Streptococcus*. Enucleation was highest with gram-negatives (27.7%), driven by *Serratia* (85.7%, p=0.05).
• Presenting VA was the most consistent predictor of final VA.

Clinical Relevance: Blood cultures and systemic source-finding are central to management, and presenting VA and organism (particularly *Serratia*) inform prognosis and counselling, though intraocular culture yield is poor.

Limitations: Retrospective design with only culture-positive cases at tertiary centres introduces selection bias, and the small subgroups limit organism-specific inference.

Study Type: Retrospective, multicentre observational case series (25 patients, 29 eyes; six tertiary centres, 28 years).

Key Question: What are the clinicoradiological features and treatment outcomes of biopsy-proven orbital and periorbital epithelioid hemangioma?

Key Findings:

• Among 20 eyes with orbital involvement on imaging, lesions were predominantly superior (70%), with lacrimal gland involvement in 70% and extraocular muscle involvement in 50% (superior rectus-levator complex most common, 35%).
• A "hypervascular" MRI/angiographic phenotype (heterogeneous enhancement, internal flow voids) occurred in 36% of eyes.
• Excisional biopsy was curative (p=0.0005); residual tumour was more common in infiltrative, ill-defined lesions (p=0.0196).

Clinical Relevance: A superior orbital lesion involving the lacrimal gland, especially with hypervascular features, should prompt consideration of EH, with complete excision where feasible being the preferred strategy.

Limitations: Small, retrospective, heterogeneously managed cohort, with imaging available in only 21 of 25 cases and no standardised follow-up reported.