Comparative Outcomes of Single-Stage Versus Two-Stage Silicone Oil Removal With Phacoemulsification and Intraocular Lens Implantation

Abstract

Purpose: To compare visual, refractive, and surgical outcomes between single-stage (concurrent) and 2-stage (sequential) silicone oil removal with phacoemulsification and intraocular lens implantation in patients previously managed for retinal detachment. Methods:This retrospective case series included patients who underwent silicone oil removal with cataract extraction between January 2018 and October 2024 at Bascom Palmer Eye Institute. Outcomes evaluated included best-corrected visual acuity (BCVA), spherical equivalent, and refractive accuracy based on median prediction error and mean absolute error. Results: A total of 272 eyes were analyzed: 26 in the sequential group (mean ± SD age, 48.7 ± 12.9 years; 54% male) and 246 in the concurrent group (mean ± SD age, 52.9 ± 12.2 years; 61% male). BCVA improved in both groups, with final median BCVAs of 0.53 logMAR in the sequential group and 0.71 logMAR in the concurrent group (P = .064). Spherical equivalent was similar between groups (P = .843), but the sequential group had better refractive predictability with a lower median prediction error compared with the concurrent group (median –0.26 diopters vs 0.21 diopters; P = .008). The mean absolute error was comparable between groups (P = .662). Subgroup analysis and multivariate regression revealed no significant predictors of outcomes. Cystoid macular edema occurred in 6 eyes in the sequential group (23%) and 7 eyes in the concurrent group (3%); glaucoma or elevated intraocular pressure occurred in 6 eyes in the sequential group (23%) and 16 eyes in the concurrent group (7%) (P = .013). Conclusions: Both approaches resulted in improved outcomes, with the sequential method providing superior refractive predictability with less hyperopic error and a trend toward better final visual acuity, suggesting a potential advantage for staged procedures in enhancing surgical precision. However, the sequential method was also associated with higher postoperative complication rates, underscoring the need to balance refractive precision with surgical safety when selecting between staged and combined procedures.

Keywords

phacoemulsification cataract surgery silicone oil removal visual acuity refractive outcomes prediction error

Introduction

Pars plana vitrectomy with silicone oil tamponade is a common method for managing complex retinal detachments, especially with proliferative vitreoretinopathy or giant retinal tears.¹ While effective for temporary stabilization, extended silicone oil retention (>3–6 months) increases the risk of cataracts, glaucoma, and corneal decompensation.^2,3 Due to rapid cataract development, combined cataract extraction and silicone oil removal is often needed. However, silicone oil interferes with axial length and intraocular lens (IOL) power measurements.^4,5

The optimal timing of phacoemulsification relative to silicone oil removal remains under debate. Single-stage surgery reduces anesthesia exposure and accelerates recovery but may increase refractive error.⁶ Two-stage surgery offers more accurate biometry and better outcomes but requires multiple procedures.⁷

This study compares visual outcomes and complications between single-stage and 2-stage silicone oil removal with cataract extraction and IOL implantation.

Methods

This retrospective case series comprised patients who underwent silicone oil removal with cataract extraction between January 2018 and October 2024 at Bascom Palmer Eye Institute. Included were patients who had undergone silicone oil removal with phacoemulsification and IOL implantation after retinal detachment repair and who had at least 2 months of follow-up. Exclusion criteria included prior silicone oil exchange, corneal pathology affecting refraction, prior corneal surgery, recurrent detachment after silicone oil removal, or incomplete data.

Patients were grouped by surgical timing: single-stage (concurrent) versus 2-stage (sequential) silicone oil removal with cataract extraction and IOL implantation. Collected variables included demographic characteristics, comorbidities, surgical details, best-corrected visual acuity (BCVA), complications, and refractive outcomes (spherical equivalent and refractive predictability based on median prediction error and mean absolute error). Biometry was performed using IOLMaster 700, with IOL power calculations using the Barrett Universal II formula.

Ethical approval was granted by the University of Miami Institutional Review Board. The study adhered to the principles of the Declaration of Helsinki and the 1996 Health Insurance Portability and Accountability Act guidelines. Data were analyzed using descriptive statistics, t tests, Wilcoxon tests, and multivariate regression. P values less than .05 were considered statistically significant.

Results

Sequential Group

Demographic Characteristics

The sequential group included 26 patients (mean ± SD age, 48.7 ± 12.9 years), of whom 14 (54%) were male. Racial distribution was as follows: 58% of patients (n = 15) reported as White, 23% (n = 6) as Black or African American, 12% (n = 3) as Asian, and 8% (n = 2) as unspecified. The left eye was affected in 17 cases (65%).

Systemic comorbidities included diabetes mellitus in 8 patients (31%) and hypertension in 7 patients (27%). Mental health disorders were present in 5 patients (19%). Ocular comorbidities included glaucoma in 5 patients (19%) and nuclear cataract in 15 patients (58%). Maculopathy was identified in 4 patients (15%), including diabetic macular edema (DME) in 3 (12%) and cystoid macular edema (CME) in 1 (4%).

Blurred or reduced vision was reported by 18 patients (69%), eye pain by 4 patients (15%), double vision by 2 patients (8%), and itching or tearing by 2 patients (8%). The preoperative BCVA in the sequential group was a mean ± SD 0.98 ± 0.50 logMAR. For the 15 patients (58%) with available refraction data, the mean ± SD refractive error was 2.43 ± 3.01 diopters. Axial length was normal in 15 eyes (58%), myopic in 10 eyes (38%), and hyperopic in 1 eye (4%) (Table 1).

Table 1.

Baseline (Preoperative) Characteristics of the Patients, by Surgical Group.

	Surgical Group^a
Characteristic	Sequential	Concurrent
Total patients, n (%)	26 (9.6)	246 (90.4)
Total eyes, n (OD/OS)	26 (9/17)	246 (131/115)
Mean age, (y) ± SD	48.7 ± 12.9	52.9 ± 12.2
Sex, n, male/female	14/12	150/96
Ethnicity, n (%)
White	15 (57.7)	167 (67.9)
Black	6 (23.1)	68 (27.6)
Asian	3 (11.5)	4 (1.6)
Unspecified	2 (7.7)	7 (2.9)
Medical history, n (%)
Type 2 diabetes	8 (30.8)	98 (40)
Hypertension	7 (26.9)	148 (60.2)
Mental disorders	5 (19.2)	–
Glaucoma	5 (19.2)	49 (20)
Nuclear cataract	15 (57.7)	–
Maculopathy	4 (15.4)	40 (16.3)
Sickle cell retinopathy	–	1 (0.4)
Mean preoperative BCVA, logMAR, ± SD	0.98 ± 0.50	1.39 ± 0.70
Mean preoperative refractive prediction error, diopters, ± SD	2.43 ± 3.01	1.84 ± 3.71
Axial length, n (%)^b
Normal	15 (57.7)	138 (56.1)
Myopic	10 (38.5)	102 (41.5)
Hyperopic	1 (3.8)	6 (2.4)

Abbreviation: BCVA, best-corrected visual acuity.

Patients underwent either single-stage (concurrent) or 2-stage (sequential) silicone oil removal with cataract extraction and intraocular lens implantation.

Normal = 22–25 mm; myopic = >25 mm; hyperopic = <22 mm.

Surgical Characteristics

Silicone oil was inserted during various retinal detachment repair surgeries, including pars plana vitrectomy (PPV) alone in 15 eyes (58%) and combined PPV with scleral buckle in 11 eyes (42%) (Table 2). In all cases, 1000-centistoke silicone oil was used.

Table 2.

Surgical Characteristics of the Patients, by Surgical Group.

	Surgical Group^a
Characteristic	Sequential(n = 26)	Concurrent(n = 246)
Surgery type for silicone oil insertion, n (%)
PPV alone	15 (57.7)	115 (46.7)
PPV/scleral buckle	11 (42.3)	131 (53.3)
Silicone oil viscosity, n (%)
1000 cSt	26 (100)	240 (97.6)
5000 cSt	–	6 (2.4)
Endolaser, n (%)	4 (15.4)	53 (21.5)
Intravitreal anti-VEGF injection, n (%)	3 (11.5)	47 (19.1)
Membrane peeling, n (%)	1 (3.8)	39 (15.9)
Sub-Tenon kenalog injection, n (%)	1 (3.8)	46 (18.7)
AC washout, n (%)	2 (7.7)	2 (0.8)
Gas tamponade, n (%)
C₃F₈	3 (11.5)	25 (10.2)
SF₆	–	17 (6.9)
Phacoemulsification surgeon subspecialty, n (%)
Cornea	22 (84.6)	3 (1.2)
Retina	3 (11.5)	243 (98.8)
Glaucoma	1 (3.85)	–
IOL types implanted, n (%)
TECNIS Eyhance	12 (46.2)	50 (20.3)
TECNIS Simplicity	5 (19.2)	84 (34.1)
TECNIS PCB00	4 (15.4)	48 (19.5)
AcrySof lenses (various models)	3 (11.5)	45 (18.3)
Clareon UVA	1 (3.8)	5 (2)
Sensar AR40E	1 (3.8)	9 (3.7)
Lens design, n (%)
One-piece	24 (92.3)	192 (78)
Three-piece	2 (7.7)	54 (22)
Toric	3 (11.5)	–
Mean IOL power, diopters, ± SD	19.2 ± 4.08	17.95 ± 4.93
Lens placement, n (%)
In-the-bag	24 (92.3)	216 (87.8)
Sulcus	2 (7.7)	29 (11.8)
Scleral fixated	–	1 (0.4)
Mean surgery duration (min) ± SD	First: 28.6 ± 19.04 Second: 26.96 ± 19.51	67.8 ± 37.4
Mean target refraction, diopters, ± SD	–0.56 ± 0.62	–0.41 ± 0.37

Abbreviations: AC, anterior chamber; anti-VEGF, anti–vascular endothelial growth factor (aflibercept); cst, centistoke; C₃F₈, perfluoropropane gas; IOL, intraocular lens; PPV, pars plana vitrectomy; SF₆, sulfur hexafluoride gas.

Patients underwent either single-stage (concurrent) or 2-stage (sequential) silicone oil removal with cataract extraction and intraocular lens implantation.

Silicone oil removal was performed by a retina specialist in all eyes, while cataract surgery was performed by cornea specialists in 22 eyes (85%), by retina specialists in 3 eyes (12%), and by a glaucoma specialist in 1 eye (4%). Additional procedures included endolaser (4 eyes [15%]), intravitreal antivascular endothelial growth factor (anti-VEGF) injection (3 eyes [12%]), anterior chamber washout (2 eyes [8%]), membrane peeling (1 eye [4%]), sub-Tenon kenalog injection (1 eye [4%]), and gas tamponade (3 eyes [12%]). The mean ± SD interval between surgeries was 7.3 ± 4.6 months.

In cataract surgery with IOL implantation, the models used included TECNIS Eyhance in 12 eyes (46%), TECNIS Simplicity in 5 eyes (19%), TECNIS PCB00 in 4 eyes (15%), AcrySof lenses (SN60WF, SN6AT3, MA60AC) in 3 eyes (12%), Clareon UVA in 1 eye (4%), and Sensar AR40E in 1 eye (4%). One-piece lenses were implanted in 24 eyes (92%), and 3-piece lenses in 2 eyes (8%). Toric lenses were used in 3 eyes (12%), including 2 TECNIS Eyhance Toric II and 1 AcrySof SN6AT3. The mean ± SD duration of the first surgery was 28.6 ± 19.04 minutes, slightly longer than the second surgery (mean ± SD 26.96 ± 19.51 minutes).

IOL powers ranged from 10.0 to 26.5 diopters (mean ± SD 19.2 ± 4.08 diopters), and most lenses (n = 18, 69%) were between 17.0 diopters and 23.0 diopters. In-the-bag placement was achieved in 24 eyes (92%), with sulcus placement performed in 2 eyes (8%). No anterior chamber, scleral-fixated, or iris-fixated lenses were used. The mean ± SD target refraction calculated preoperatively was −0.56 ± 0.62 diopters (Table 2).

The mean ± SD follow-up in the sequential group was 30.4 ± 20.9 months. The number of surgeries performed per eye by last follow-up was a mean ± SD 2.58 ± 0.70; a total of 13 eyes (50%) underwent 2 surgeries, 12 eyes (46%) underwent 3 surgeries, and 1 eye (3.8%) underwent 5 surgeries by last follow-up (Table 3).

Table 3.

Visual, Refractive, and Surgical Outcomes in Patients, by Surgical Group.

	Surgical Group^a
Outcomes	Sequential(n = 26)	Concurrent(n = 246)	P Value
Mean follow-up duration, months, ± SD	30.4 ± 20.9	15.90 ± 17.68	–
Mean total surgeries until final follow-up, n, ± SD	2.58 ± 0.70	1.18 ± 0.49	–
Final BCVA, logMAR, median (IQR)	0.53 (0.30–0.98)	0.71 (0.40–1.24)	.064
Final BCVA 20/40 or better, n (%)	7 (26.9)	41 (16.7)	.229
Final BCVA between 20/40 and 20/200, n (%)	12 (46.2)	109 (44.3)	.860
Final BCVA 20/200 or worse, n (%)	7 (26.9)	96 (39.0)	.204
Spherical equivalent, diopters, median	−0.13	−0.25	.843
Mild myopia (0 to −1.00 diopters), n (%)	7 (33.3)	70 (40)	.448
Moderate-to-high myopia (≤−1.00 diopters), n (%)	4 (19.1)	29 (16.6)	.755
Hyperopia (>0 diopters), n (%)	5 (23.8)	55 (31.4)	.486
Emmetropia (0 diopters), n (%)	5 (23.8)	21 (12)	.041^b
Refractive predictability
Prediction error, diopters, median	−0.26	0.21	.008^b
Absolute error, diopters, mean	0.46	0.41	.662
Eyes within ±0.5 diopters, n (%)	14 (53.8)	147 (59.8)	.708
Eyes within ±1.0 diopters, n (%)	16 (61.5)	191 (77.6)	.111
Postoperative complications, n (%)			.013^b
Cystoid macular edema	6 (23.1)	7 (2.8)
Secondary glaucoma	6 (23.1)	16 (6.5)
Posterior capsular opacification	6 (23.1)	1 (0.4)
Persistent hypotony	–	1 (0.4)

Abbreviations: BCVA, best-corrected visual acuity; IQR, interquartile range.

Patients underwent either single-stage (concurrent) or 2-stage (sequential) silicone oil removal with cataract extraction and intraocular lens implantation.

Statistically significant difference between groups (P < .05).

Final BCVA was a mean ± SD 0.69 ± 0.65 logMAR, demonstrating significant improvement from baseline (median change −0.31 logMAR; P = .003, by Wilcoxon test). At final follow-up, BCVA was 20/40 or better in 7 eyes (27%), between 20/40 and 20/200 in 12 eyes (46%), and 20/200 or worse in 7 eyes (27%).

Final spherical equivalent values were available for 21 eyes (81%). The mean ± SD spherical equivalent at final follow-up was −0.18 ± 1.34 diopters, showing a significant improvement from baseline (P = .014). Mild myopia (defined as 0 to −1.00 diopters) was seen in 7 eyes (33%), moderate-to-high myopia (defined as ≤ −1.00 diopters) in 4 eyes (19%), hyperopia (defined as >0 diopters) in 5 eyes (24%), and emmetropia (defined as 0 diopters) in 5 eyes (24%).

Refractive predictability was evaluated using a 1-sample t test based on prediction error. The prediction error was a mean ± SD −0.26 ± 1.32 diopters (n = 26), not significantly different from zero (95% CI, −0.79 to +0.28 diopters; P = .33), indicating that there was no systematic over- or undercorrection of the data. The absolute error was a mean ± SD 0.88 ± 1.00 diopters.

CME, secondary glaucoma, or elevated intraocular pressure, and posterior capsular opacification each occurred in 6 eyes (23%) (Table 3).

Predictive Factors

To assess the impact of axial length, patients were grouped as myopic, emmetropic, or hyperopic. One-way analysis of variance (ANOVA) showed no significant differences in prediction error (F(2,23) = 0.1869, P = .831) or mean absolute error (F(2,23) = 0.4031; P = .672) between these groups.

No significant correlations were found between IOL power and prediction error (r = −0.1962, P = .337) or between IOL power and mean absolute error (r = 0.1825, P = .372). Similarly, no significant differences were observed in prediction error (F(4,21) = 0.93, P = .465) or mean absolute error (F(4,21) = 0.42, P = .793) across the different IOL types.

Multivariate regression analyses were performed to identify predictors of refractive accuracy (based on mean absolute error), final BCVA, and final spherical equivalent. Independent variables included preoperative BCVA, preoperative spherical equivalent, axial length category, IOL power, IOL type, lens placement, and surgeon subspecialty. No model showed a statistically significant correlation with refractive predictability (R² = 0.863, adjusted R² = 0.522, P = .192), final BCVA (R² = 0.660, adjusted R² = 0.049, P = .493), or final spherical equivalent (R² = 0.595, adjusted R² = −0.216, p = 0.672). Logistic regression analysis to identify predictors of postoperative complications also showed no significant predictors (P = .820).

Concurrent Group

Demographics Characteristics

The concurrent group included 246 patients (mean ± SD age, 52.9 ± 12.2 years), of whom 150 (61%) were male. Racial distribution was as follows: 167 (68%) reported as White, 68 (28%) as Black or African American, 4 (2%) as Asian, and 7 (3%) as unspecified. The right eye was affected in 131 cases (53%).

Systemic comorbidities included hypertension in 148 patients (60%) and type 2 diabetes mellitus in 98 (40%). Ocular comorbidities included glaucoma in 49 patients (20%), DME in 32 (13%), and CME in 8 (3%); 1 patient (0.4%) had sickle cell retinopathy.

Blurred or reduced vision was the most common preoperative symptom, reported by 88 patients (36%), followed by floaters, glare, or flashes in 32 (13%), eye pain or discomfort in 27 (11%), itching or tearing in 19 (8%), curtain-like shadows in 3 (1%), dry eye symptoms in 3 (1%), headaches in 2 (0.8%), and diplopia in 1 (0.4%).

The preoperative BCVA was a mean ± SD 1.39 ± 0.70 logMAR. Refraction data were available for 107 patients (44%) in the concurrent group. The refractive error was a mean ± SD 1.84 ± 3.71 diopters. Axial length was normal (range, 22–25 mm) in 138 eyes (56%), myopic (>25 mm) in 102 eyes (42%), and hyperopic (<22 mm) in 6 eyes (2%) (Table 1).

Surgical Characteristics

Silicone oil was inserted during combined PPV/scleral buckle surgery in 131 eyes (53%) and inserted during PPV alone in 115 eyes (47%). Surgeries were performed by retina specialists in 243 cases (99%) and cornea specialists in 3 cases (1%).

Silicone oil was used at 1000 centistoke in 240 eyes (98%) and at 5000 centistoke in 6 eyes (2%). Additional intraoperative procedures included laser retinopexy in 53 eyes (22%), intravitreal anti-VEGF injection in 47 eyes (19%), membrane peeling in 39 eyes (16%), sub-Tenon kenalog injection in 46 eyes (19%), anterior chamber washout in 2 eyes (0.8%), and gas tamponade with C₃F₈ in 25 eyes (10%) or with SF₆ in 17 eyes (7%). Operative time ranged from 27 minutes to 257 minutes (mean ± SD 67.8 ± 37.4 minutes; median 57 minutes).

The most frequently implanted IOLs were 1-piece TECNIS Simplicity (ZCB00) in 84 eyes (34%), TECNIS Eyhance (DIB00) in 50 eyes (20%), TECNIS PCB00 in 48 eyes (20%), and AcrySof lenses (SN60WF, SN6AT3, MA60AC) in 45 eyes (18%). Sensar AR40E was used in 9 eyes (4%), and Clareon UVA and CT LUCIA in 5 eyes each (2%).

Overall, 192 eyes (78%) received 1-piece IOLs and 54 (22%) received 3-piece lenses. IOL power ranged from 2 diopters to 29 diopters (mean ± SD 17.95 ± 4.93 diopters). In-the-bag placement was achieved in 216 eyes (88%), sulcus placement in 29 eyes (12%), and scleral fixation in 1 eye (0.4%). No anterior chamber or iris-fixated lenses were used. The preoperative target refraction was a mean ± SD −0.41 ± 0.37 diopters (Table 2).

Follow-up in the concurrent group was a mean ± SD 15.90 ± 17.68 months, and the number of surgeries performed per eye was a mean ± SD 1.18 ± 0.49. Final BCVA improved significantly from baseline, to a mean ± SD 0.92 ± 0.72 logMAR (median change −0.42; P <.0001, by Wilcoxon test). Final BCVA was 20/40 or better in 41 eyes (17%), between 20/40 and 20/200 in 109 eyes (44%), and 20/200 or worse in 96 eyes (39%) (Table 3).

Final spherical equivalent, available for 175 eyes (71%) in the concurrent group, was a mean ± SD 0.54 ± 9.71 diopters. The spherical equivalent shifted significantly postoperatively (median –2.38 diopters; P < .0001). Mild myopia was present in 70 eyes (40%), moderate-to-high myopia in 29 eyes (17%), hyperopia in 55 eyes (31%), and emmetropia in 21 eyes (12%) (Table 3).

Refractive predictability was assessed using a 1-sample t test. The mean ± SD prediction error was 0.72 ± 8.17 diopters (n = 246), not significantly different from zero (95% CI, −0.31 to 1.75 diopters; t(243) = 1.37, P = .1723; η² = .008), indicating no systematic over- or undercorrection of the data.

Secondary glaucoma was the most common postoperative complication, seen in 16 eyes (6%), followed by CME in 7 eyes (3%), and single cases of persistent hypotony and posterior capsular opacification, each occurring in 0.4% of eyes (Table 3).

Predictive Factors

Patients were categorized by preoperative axial length into myopic, emmetropic, and hyperopic groups. One-way ANOVA showed no significant differences in prediction error (F(2,221) = 0.29, P = .7515) or mean absolute error (F(2,223) = 0.46, P = .6334) of the predictive predictability between these groups.

Correlation analysis revealed no significant association between IOL power and prediction error (r = 0.089, P = .178) or between IOL power and mean absolute error (r = 0.076, P = .249). IOL type also showed no significant effect on prediction error (F(2,241) = 0.50, P = .610) or mean absolute error (F(2,243) = 0.04, P = .960). No significant differences in refractive outcomes were found among the different IOL types.

Multivariate regression analysis to identify predictors of refractive predictability based on the mean absolute error showed no significant associations (R² = 0.057, P = .822). In assessing final spherical equivalent as a predictor in the overall multivariate regression model, a modest significant association was observed (R² = 0.128, P = .028); however, no individual variables were identified as independent predictors. The model assessing final BCVA showed a moderately significant association (R² = 0.324, P < .001). These findings suggest that a combination of factors—such as preoperative BCVA, spherical equivalent, axial length, IOL characteristics, and surgeon subspecialty—may collectively influence visual outcomes, though no single factor was independently predictive. Logistic regression found no significant associations between preoperative clinical or demographic variables and postoperative complications (all P > .5; Nagelkerke R² < 0.05).

Overall, biometric, surgical, and lens-related factors did not significantly influence refractive predictability or complication rates, although a combination of preoperative BCVA, spherical equivalent, axial length, IOL parameters, and surgeon’s subspecialty was moderately predictive of the final BCVA.

Comparison of Outcomes Between Sequential and Concurrent Groups

Both groups demonstrated significant postoperative improvement in BCVA from baseline (P = .003 for the sequential group and P < .0001 for the concurrent group). There was a trend toward better BCVA at last follow-up in the sequential group (median 0.53, IQR 0.30–0.98) compared with the concurrent group (median 0.71, IQR 0.40–1.24; P = .064).

Final spherical equivalent did not differ significantly between groups (median –0.13 in the sequential group vs – 0.25 in the concurrent group; P = .843). Refractive predictability differed significantly between the groups, with the sequential group exhibiting a deviation from the target toward mild myopia (median prediction error −0.26) and the concurrent group showing a deviation toward hyperopia (median prediction error 0.21) (95% CI, 0.1250–0.7500; P = .0075). This observation holds clinical significance, as a postoperative hyperopic outcome is typically less favorable than mild myopia in patients with a pseudophakic lens status, especially in terms of visual satisfaction and reduced dependence on corrective lenses. Despite the directional variance, the mean absolute error was comparable between groups (mean 0.46 vs 0.41; P = .662), reflecting similar refractive predictability. Target refraction within 0.5 diopters and within 1.0 diopters was achieved in 54% and 62% of eyes, respectively, in the sequential group, and in 60% and 78% of eyes, respectively, in the concurrent group (P = .708 and P = .111).

Postoperative complications were more frequent in the sequential group than in the concurrent group. A total of 23% of patients in the sequential group vs 3% of patients in the concurrent group developed CME, while 23% of patients in the sequential group versus 6% of patients in the concurrent group developed secondary glaucoma, and 23% of patients in the sequential group versus 0.4% of patients in the concurrent group developed posterior capsular opacification (overall P = .013) (Table 3).

Discussion

Silicone oil remains a valuable tamponade for managing complex retinal detachments, offering advantages over gas tamponade, such as permitting air travel, maintaining clear media in monocular patients, and supporting the inferior retina.^8–10 However, long-term use increases risks of elevated intraocular pressure, cataract formation, and emulsification, often necessitating silicone oil removal with cataract extraction and IOL implantation.^11,12

Achieving accurate refractive outcomes following combined surgery is challenging, with a tendency toward hyperopic shift.^13,14 Recent advances in optical biometry and newer-generation IOL formulas, such as Barrett Universal II and Kane, have improved refractive predictability in silicone oil–filled eyes compared to ultrasound biometry.^15,16

The timing of phacoemulsification relative to silicone oil removal remains controversial. Single-stage surgery simplifies care and accelerates recovery but may increase refractive error and complication risk. Two-stage surgery permits more accurate biometry and may yield better visual outcomes, though it requires an additional procedure. Despite better fundus visualization during combined surgery, surgically induced astigmatism can still compromise visual quality.

In our cohort, the development of cataracts played a significant role in determining the timing of surgery. Increasing lens opacity often compromised visualization of the posterior segment, making retinal assessment and postoperative monitoring more challenging. In the sequential cohort, the mean interval between retinal detachment repair and subsequent cataract surgery was 7.3 ± 4.6 months, reflecting the point at which visualization or visual function typically became compromised. The decision to proceed with silicone oil removal was individualized rather than time-based and was influenced by several clinical considerations, including tamponade duration, intraocular pressure elevation or secondary glaucoma, silicone oil emulsification, and, less frequently, the development of CME or epiretinal membrane. In most cases, surgery was performed once retinal stability was achieved and these secondary indications emerged. This highlights the multifactorial nature of timing decisions in eyes with silicone oil tamponade, where achieving an optimal balance between retinal safety, anterior segment clarity, and refractive precision remains critical.

Madanagopalan et al compared single-stage and 2-stage surgeries in 135 eyes. At 3 months, a myopic shift was seen in 92% of eyes, with the 2-stage group showing significantly less refractive error (−0.64 diopters vs −1.73 diopters; P = .002) and greater BCVA improvement. Our results similarly demonstrated a lower prediction error in the sequential group and a trend toward better BCVA, supporting the visual and refractive benefits of a staged approach.¹⁷

Krepler et al¹⁸ found comparable visual acuity and complication rates between single-stage and 2-stage surgery groups, but faster recovery in the combined group. Refractive precision was also better, likely due to more stable ocular measurements after oil removal, whereas biometry during silicone oil presence introduces variability due to altered ocular indices.¹⁸ However, in our study, the sequential group demonstrated higher complication rates for CME, secondary glaucoma, and posterior capsular opacification (each 23%) compared with the concurrent group (3%, 6%, and 0.4%, respectively; P = .013). These differences likely reflect that the follow-up duration was substantially longer in the sequential cohort than in the concurrent cohort (mean ± SD 30.4 ± 20.9 months vs 15.9 ± 17.7 months), which allowed for greater detection of late-onset events such as posterior capsular opacification and CME.

Nevertheless, refractive predictability favored the sequential approach, suggesting improved biometric accuracy after oil removal. These findings emphasize that while staged procedures may enhance refractive precision, they must be weighed against a higher observed complication rate and tailored to individual patient factors, such as ocular comorbidities, surgical risk, and tolerance for multiple procedures. Future prospective studies with standardized follow-up are warranted to clarify the safety–precision trade-off between approaches.

These refractory findings align with prior reports supporting an approach of delaying cataract surgery for improved refractive predictability. Still, the need for 2 surgeries may not be ideal for all patients. Factors such as cataract density, need for posterior visualization, and patient comorbidities must guide decision-making.

Zheng et al¹⁹ assessed simultaneous silicone oil removal and sulcus IOL implantation in 40 aphakic eyes, finding good visual improvement but greater refractive variability in eyes requiring capsulotomy. Similarly, our concurrent group showed higher prediction error, reinforcing the impact of biometry inaccuracies during oil presence. Lingam²⁰ emphasized early oil removal to reduce complications and noted optical biometry’s advantages. In our study, axial length was measured by available methods, and no significant refractive accuracy differences were found across axial length categories, suggesting satisfactory outcomes are possible with careful planning.

Shu et al outlined key contributors to refractive error: axial length inaccuracies, anterior chamber depth shifts, capsular contraction, and the effect of silicone oil.¹³ Targeting mild myopia and using optical biometry when possible are recommended.¹³ Postoperative air or gas tamponade may contribute to myopic shift by causing forward displacement of the IOL during phacovitrectomy, leading to refractive errors that can persist even after tamponade absorption.²¹ Our study also found no significant association between refractive outcomes and axial length, IOL power, or lens placement, underlining the multifactorial nature of prediction error.

The hyperopic shift in the concurrent group may reflect inaccurate estimates of IOL position and underpowered IOL selection during oil-filled biometry.²² After oil removal, posterior lens shift may exacerbate this effect. In contrast, the sequential approach allows for more accurate measurement and better refractive predictability.^23,24 Moreover, the sequential approach allows for biometry to be conducted after silicone oil removal, likely improving measurement accuracy and refractive predictability.¹⁷

Ji et al reported early corneal astigmatism following combined surgery that normalized within a month, with visual gains noted early. Our sequential group showed slightly slower visual improvement but better final BCVA and lower prediction error, suggesting long-term refractive benefit over short-term recovery.²⁵ In contrast, the sequential group in the current study showed a trend toward better final BCVA (median 0.53 vs 0.71 logMAR; P = 0.064) and less residual hyperopic error (−0.26 vs 0.21; P = 0.0075), favoring long-term refractive precision over rapid early gains.

In a study by Karimi et al, modest visual improvement and stable intraocular pressure were observed in patients after combined procedures. In comparison, our staged approach resulted in improved BCVA and a final spherical equivalent closer to the target, supporting its refractive and visual advantages.⁶

In this study, additional pre-, intra-, or postoperative interventions such as membrane peeling, laser, or intravitreal injections were performed only when clinically indicated. In particular, epiretinal membranes were addressed selectively when tractional changes or macular involvement were present and likely to compromise visual recovery. While such adjunctive procedures may contribute to variability in postoperative visual outcomes, their targeted use in clinically significant cases suggests a limited confounding effect on the overall comparative results between groups.

This study has limitations, including its retrospective, nonrandomized design, which may introduce selection bias. Provider-dependent surgical decisions and the smaller sample size in the sequential group limited subgroup analysis. Incomplete preoperative refractive data and the inclusion of eyes with preexisting maculopathy may have affected visual and refractive outcomes, despite efforts to account for new-onset pathology.

Despite these limitations, this study represents the largest comparative series evaluating refractive and visual outcomes between concurrent and sequential silicone oil removal with cataract surgery. The real-world, multicenter design strengthens the generalizability of the findings by reflecting the variability encountered in clinical practice.

In conclusion, both single-stage and 2-stage approaches to silicone oil removal with cataract extraction significantly improved visual and refractive outcomes in patients previously treated for retinal detachment. However, the sequential method showed superior refractive predictability with less hyperopic error, and a trend toward better visual acuity, but a higher incidence of postoperative complications, emphasizing the need to balance refractive precision with surgical safety. Future prospective studies with standardized follow-up are needed to validate these results in broader clinical settings.

Footnotes

Author Notes

A. Ahmed and S. Borges contributed equally as co-second authors.

During the preparation of this manuscript, the authors used GPT-4 (OpenAI) to assist with grammar checking, language refinement, and formatting suggestions. All content generated or edited with the assistance of this tool was thoroughly reviewed and revised by the authors to ensure accuracy, validity, and appropriateness. The authors take full responsibility for the integrity and originality of the final version of the manuscript.

ORCID iDs

Seyyedehfatemeh Ghalibafan

Landon J. Rohowetz

Nicolas A. Yannuzzi

Ethical Considerations

This retrospective study was approved by the Institutional Review Board of the University of Miami.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Macula Foundation Research Grant and Research to Prevent Blindness – Unrestricted Grant (GR004596-1).

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Statement of Informed Consent

The Institutional Review Board of the University of Miami waived the requirement for individual informed consent due to the use of deidentified data.

References

Schildroth

Wingelaar

Boeke

, et al. The Wisconsin silicone oil study (report #1): anatomical and functional outcomes of repair of retinal detachment with proliferative vitreoretinopathy. Eye (Lond). 2024;38(14):2731-2736.

Chen

Kearns

Zhou

, et al. Silicone oil in vitreoretinal surgery: indications, complications, new developments and alternative long-term tamponade agents. Acta Ophthalmol. 2021;99(3):240-250.

Issa

Xia

Zarbin

Bhagat

Silicone oil removal: post-operative complications. Eye (Lond). 2020;34(3):537-543.

Grinbaum

Treister

Moisseiev

Predicted and actual refraction after intraocular lens implantation in eyes with silicone oil. J Cataract Refract Surg. 1996;22(6):726-729.

Chen

Zhao

, et al. Accuracy of new intraocular lens calculation formulae in eyes undergoing silicone oil removal/pars plana vitrectomy-cataract surgery. Retina. 2023;43(9):1579-1589.

Karimi

Nikkhah

Farzanbakhsh

Karjou

Safi

Outcomes of combined phacoemulsification/intraocular lens implantation and silicone oil removal. Int Ophthalmol. 2022;42(7):2267-2272.

Bellucci

Romano

Ramanzini

Tedesco

Gandolfi

Mora

Pars plana vitrectomy alone or combined with phacoemulsification to treat rhegmatogenous retinal detachment: a systematic review of the recent literature. J Clin Med. 2023;12(15):5021.

Wang

Snead

Alexander

Ian Wilson

Assessing bulk emulsification at the silicone oil–saline solution interface in a 3D model of the eye. Acta Ophthalmol. 2021;99(2):e209-e214.

Schwartz

Flynn HW

Wang

Kuriyan

Abariga

Lee

WH.

Tamponade in surgery for retinal detachment associated with proliferative vitreoretinopathy. Cochrane Database Syst Rev. 2020;5(5):CD006126.

10.

Iyer

SSR

Regan

Burnham

Chen

. Surgical management of diabetic tractional retinal detachments. Surv Ophthalmol. 2019;64(6):780-809.

11.

Feng

Zheng

, et al. Risk of silicone oil as vitreous tamponade in pars plana vitrectomy: a systematic review and meta-analysis. Retina. 2017;37(11):1989-2000.

12.

Kunikata

Abe

Nakazawa

Historical, current and future approaches to surgery for rhegmatogenous retinal detachment. Tohoku J Exp Med. 2019;248(3):159-168.

13.

Shu

Che

Shan

Zhao

JS.

Topical review: causes of refractive error after silicone-oil removal combined with cataract surgery. Optom Vis Sci. 2020;97(12):1099-1104.

14.

Al-Habboubi

Al-Zamil

Al-Habboubi

Khandekar

Visual outcomes and refractive status after combined silicone oil removal/cataract surgery with intraocular lens implantation. J Ophthalmic Vis Res. 2018;13(1):17-22.

15.

Zhang

Han

Jin

, et al. Correction of axial length measurement error by IOLMaster 700 could improve refractive prediction accuracy in silicone oil-filled eyes. Acta Ophthalmol. 2024;102(5):e718-e726.

16.

Kaiser

Jandewerth

Bucur

Kohnen

Lwowski

Axial length adjustment in eyes with silicone oil endotamponade reduces overestimation by a swept-source optical coherence tomography-based biometer. Clin Exp Ophthalmol. 2024;52(8):833-839.

17.

Madanagopalan

Susvar

Arthi

Refractive outcomes of a single-step and a two-step approach for silicone oil removal and cataract surgery. Indian J Ophthalmol. 2019;67(5):625-629.

18.

Krepler

Mozaffarieh

Biowski

Nepp

Wedrich

Cataract surgery and silicone oil removal: visual outcome and complications in a combined vs. two-step surgical approach. Retina. 2003;23(5):647-653.

19.

Zheng

Huang

Chen

Huang

Chen

Refractive outcomes following simultaneous silicone oil removal and ciliary sulcus intraocular lens implantation. Eur J Ophthalmol. 2025;35(3):1103-1108.

20.

Lingam

Commentary: refractive outcomes of a single-step and a two-step approach for silicone oil removal and cataract surgery. Indian J Ophthalmol. 2019;67(7):1246-1247.

21.

Shiraki

Wakabayashi

Sakaguchi

Nishida

Effect of gas tamponade on the intraocular lens position and refractive error after phacovitrectomy: a swept-source anterior segment OCT analysis. Ophthalmology. 2020;127(4):511-515.

22.

Kanclerz

Grzybowski

Accuracy of intraocular lens power calculation in eyes filled with silicone oil. Semin Ophthalmol. 2019;34(5):392-397.

23.

Lwowski

Miraka

Müller

Singh

Koch

Kohnen

Intraocular lens calculation using 8 formulas in silicone oil-filled eyes undergoing silicone oil removal and phacoemulsification after retinal detachment. Am J Ophthalmol. 2022;244:166-174.

24.

Song

Kim

Lee

SC.

Refractive status and visual acuity changes after oil removal in eyes following phacovitrectomy, intraocular lens implantation, and silicone oil tamponade. Can J Ophthalmol. 2010;45(6):616-620.

25.

Corneal astigmatism alteration after combined silicone oil removal and cataract surgery with intraocular lens implantation. J Ophthalmol. 2023;2023:6175272.