Unexplained visual loss in retinal detachment repair: comparing gas, silicone oil and heavy silicone oil by multivariable regression.

PURPOSE: To measure the proportion of unexplained and all causes of visual loss following primary rhegmatogenous-retinal-detachment (RRD) repair, comparing gas tamponade (SF6 , C2 F6 , C3 F8 ), silicone oil (SO, 1000cs and 5000cs) and heavy silicone oil (Densiron).

METHODS: Retrospective, continuous, comparative study from 01/1/2017-31/5/2021. All primary RRDs were included after successful removal of SO and Densiron. Primary failures were excluded. Visual loss was defined as reduction of ≥0.30 logMAR units. Multivariable binary-logistic and linear regression models to compare tamponade, and all cases of unexplained visual loss and logMAR gain were performed. Covariates included age, ocular co-morbidities, pre-op vision, macula-status, high-myopia, giant-retinal-tear (GRT), perfluorocarbon-use, combined buckle/PPV, PVR-C, retinectomy, tamponade agent and post-operative lens status.

RESULTS: Of 1,012 primary RRDs, we found an incidence of unexplained visual loss in 15/1012 (1.5%, SF6 :1/341[0.3%], C2 F6 :4/338[1.2%], C3 F8 :2/239[0.8%], Densiron:0/33[0.0%], SO-1000cs:5/43[11.6%] and SO-5000cs:3/18[16.7%]), and visual loss of all causes in 57/1012 (5.6%, SF6 :13/341[3.8%], C2 F6 :14/338[4.1%], C3 F8 :15/239[6.3%], Densiron:2/33[6.1%], SO-1000cs:9/43[20.9%] and SO-5000cs:4/18[22.2%]). On multivariable binary-logistic regression, we report that macula-on RRD (Odds-Ratio[OR]5.7,95% Confidence-interval[CI]1.2-28.2, p=0.032), GRT (OR35.0,CI 2.0-617.3, p=0.015), combined buckle/PPV (OR37.7,CI 2.0-711.4, p=0.015), SO1000cs (OR86.6,CI 5.6-1,348.0), p=0.001) and 5000cs (OR37.2,CI 1.3-1,101.5, p=0.036) (Reference-tamponade:SF6 ) were associated with unexplained visual loss. Duration of oil tamponade was not linked to increase in unexplained visual loss (p=0.569).

CONCLUSIONS: Correlation between SO in detachment repairs and unexplained visual loss has been established, however incidence with HSO has not been compared to other agents. This study demonstrates that although SO was linked with risk-adjusted increased unexplained visual loss relative to gas tamponade, no such association was found for Densiron, on multivariable analysis.