Preoperative OCT-Derived Nasal Perifoveal Retinal Nerve Fiber Layer Thickness Predictively Correlates Long-Term Visual Acuity Post Oil Removal Surgery

Abstract

Purpose: With the widespread application of silicone oil in vitreoretinal surgery, the purpose of this study was to determine the risk factors of long-term vision loss 12 months post oil removal in retina-detached eyes treated with vitrectomy and silicone oil tamponade.

Methods: Of the 592 patients approached, eligible eyes completed the investigation up to 12 months post-oil-removal. Eligible eyes underwent pars-plana vitrectomy following oil tamponade. Oil removal was performed after 3 to 28 months in different individuals, under the condition that the retina has reattached as well as the hemorrhage and inflammation has dissolved. Postoperative best-corrected visual acuity (BCVA), age, sex, and interval between tamponade and removal were recorded, and retinal thickness was determined using optical coherence tomography (OCT).

Results: Fifty eyes of 50 participants aged 31 to 83 years were enrolled. BCVA (LogMAR) 12 months post-oil-removal improved in 25 of 40 (62.5%) patients, varying from 0.05 (20/22) to 1.0 (20/200) (mean ± SD = 0.55 ± 0.32). Pre-oil-removal nasal perifoveal retinal nerve fiber layer thickness varied from 16 to 83 µm (38.40 ± 18.50), and was significantly linked with post-oil-removal BCVA (0.5%, 95% confidence interval 0.0%–1.0%; P = 0.046).

Conclusions: This study demonstrates the risk factors and prognosis of visual function after long-term regeneration post vitrectomy, oil tamponade, and oil removal, thereby underscoring the need for a complete, dynamic examination of retinal structure via OCT measurement. Related studies should be conducted on a larger scale to facilitate the stratification of late-period vision damage in retina-detached eyes.

Translational Relevance: This study developed OCT-based clinical markers for the postoperative visual prognosis of eyes affected by retinal detachment.

Introduction

Pars plana vitrectomy (PPV) is routinely conducted to treat complicated retinal detachment and proliferative vitreoretinopathy (PVR). In the meantime, silicone oil is used for retinal tamponade after mobilization of stiff retinal folds.^1,2 However, it is recommended that silicone oil be removed as early as possible after an adequate tamponade duration due to the risk of long-term complications, including corneal endothelial decompensation, progressive cataract, and secondary macrophagocytic open-angle glaucoma.^3,4 Along with retinal toxicity, these contribute to postoperative vision loss, usually in the form of central dark scotomas, although the mechanisms remain unclear.^5–7 

Consistent with our previous findings, subclinical macular and inner retinal atrophy have been reported in eyes undergoing vitrectomy with silicone oil tamponade.⁸ After removal of silicone oil, the retina thickens with a long-term, substantial gain of visual acuity. Thus, whether visual outcomes fluctuate with structural modifications is of crucial interest. 

The study aimed to evaluate the best-corrected visual acuity (BCVA) 12 months post oil removal, while exploring potentially associated intrinsic risk factors. Clarifying the physical-functional regularity of the retina will provide indications for the removal of silicone oil, additional monitoring procedures that should be performed to detect preclinical retinal degeneration, and a situation in which improvement in visual acuity tends to be impeded after oil removal. 

Patients and Methods

A total of 592 patients who underwent PPV consecutively with silicone oil tamponade and removal at Shanghai General Hospital between 2021 and 2023 were approached. Of these patients, 107 eyes were properly indicated. Finally, the investigation was carried out for up to 12 months post oil removal in 50 eyes included in this observational study. All procedures adhered to the tenets of the Declaration of Helsinki, and the experimental protocol was approved by the Institutional Review Board of Shanghai General Hospital. 

In general, silicone oil is used in eyes with retinal detachment combined with grade C or D PVR, and in eyes with giant retinal tears extending more than 90 degrees. Highly myopic eyes (spherical equivalent of refraction ≤ −8 diopters [D]) and patients with proliferative diabetic retinopathy, macular hole, or macular detachment were excluded. None of the 50 patients with PVR after rhegmatogenous retinal detachment underwent unsuccessful surgeries, including buckling procedures, before PPV with silicone oil instillation. During the entire follow-up period, patients who underwent cataract surgery or YAG laser capsulotomy for posterior capsule opacification were excluded, and considering the impact of high intraocular pressure (IOP) on optic nerve damage and vision loss, patients with IOP > 25 mm Hg, or with IOP of 21 to 25 mm Hg lasting > 3 weeks were also excluded. 

All patients included in the study underwent PPV using the 3-port technique and silicone oil with a viscosity of 5000 centistokes. Oil tamponade was carried out 3 to 28 (mean ± SD = 7.52 ± 4.87, median = 6) months before oil removal surgery. Of the 50 patients, 8 (16%) were phakic, 1 (2%) was aphakic, 1 (2%) had mild cataract, and 40 (80%) were pseudophakic after PPV with oil tamponade. 

The criteria to remove silicone oil were a completely attached retina determined via both an ophthalmoscope and spectral domain optical coherence tomography (SD-OCT), and the duration of silicone oil tamponade was ≥ 3 months. Partially due to the latency caused by coronavirus disease (COVID) regulation, the duration of silicone oil tamponade was ≥ 6 months in 35 of 50 (70%) patients. Silicone oil was removed via pars plana sclerotomy in all patients without combining ultrasonic emulsification of the cataract, implantation of an intraocular lens, or other transpupillary approaches. 

Data acquisition was performed by the investigators, independent of the surgeon. The time for preoperative evaluation was 2 weeks before oil removal surgery, and, for postoperative evaluation, 12 months after oil removal surgery (Table 1). 

Table 1.

 

BCVA and Whole Retinal Thickness Change Pre- and PostOperation

 

Ocular Examination and Imaging

The patients were regularly examined using an intraocular tonometer, BCVA test, and SD-OCT at each visit. SD-OCT scans were obtained on a perifoveal scale of 6 mm × 6 mm of the retina using Heidelberg (Heidelberg, Germany) instrumentation. Each tomographic image was segmented into seven layers according to retinal anatomy: retinal nerve fiber layer (RNFL), ganglion cell layer, inner plexiform layer, inner nuclear layer, outer plexiform layer, outer nuclear layer, and outer segmental layer (from the external limiting membrane to the retinal pigmental epithelium layer). The thickness of the full layer and the aforementioned single layers were measured for each b-scan. Finally, using the algorithm attached to the software, all measurements were integrated and visualized under the Early Treatment Diabetic Retinopathy Study (ETDRS) division. 

Statistical Analysis

The results were collected, tabulated (see Supplementary Table S1), and statistically analyzed using Statistical Package for the Social Sciences (SPSS) version 23 (IBM Corp., Armonk, NY, USA). A paired, 2-tailed Student t-test was used to statistically analyze changes in visual acuity and retinal thickness. Multiple linear regression analysis was used to determine the influence of each baseline characteristic and retinal thickness alteration on BCVA outcomes. A single variable was discarded if the proportion of the lost data exceeded 40%. Any P value < 0.05 was considered to be statistically significant. 

Results

Improved BCVA and Retinal Thickness Post Silicone Oil Removal

During the follow-up period, none of the 8 patients with phakic lenses developed cataract, and none of the 50 patients had a recurrence of retinal re-detachment. 

Mean BCVA (LogMAR) was 0.37 ± 0.27 before silicone oil removal, and it significantly improved to 0.55 ± 0.32 at 12 months postoperatively (P < 0.0001, paired t-test; see Table 1). BCVA improved after surgery in 25 of 40 (62.5%) patients, but worsened or remained unchanged in 4 of 40 (10%), and 11 of 40 (27.5%) eyes, respectively. Among the 4 patients who exhibited a decrease of visual acuity, 3 had an oil tamponade duration of ≥ 6 months. 

Whole retinal thickness increased significantly post silicone oil removal in the nine subdivisions (P < 0.01, paired t-test; see Table 1). Mean gain of whole retinal thickness reached 9.92 ± 3.99 µm at the fovea. 

Reduced Nasal Perifoveal RNFL Thickness Identified as a Risk Factor for Low BCVA After Surgery

Multiple linear regression analysis was applied to determine the independent influence of age, sex, comorbidities, refraction, IOP, oil tamponade duration, thickness of each layer, and each ETDRS quadrant on the final BCVA outcome (12 months post oil removal). Among several promising candidates, age (P = 0.055), refraction (P = 0.72), and oil tamponade duration (P = 0.36) were insignificant, whereas RNFL thickness in the nasal perifoveal quadrant (0.5%, 95% confidence interval = 0.0%–1.0%; P = 0.046) was substantially associated with negative visual function outcomes (Table 2). 

Table 2.

 

Multivariable Regression Results of BCVA 12 Months Post Oil Removal

 

Of all patients investigated, the final outcome of BCVA was significantly better in those with nasal perifoveal RNFL thickness ≥ 38 µm at the pre-oil-removal timepoint, compared to those with < 38 µm (Fig. 1A). The average nasal perifoveal RNFL thickness measured pre oil removal was 38 µm. In the group of patients who achieved this median, 82% exhibited final BCVA outcomes of ≥ 0.3, whereas the mean BCVA of the entire group was 0.72. On the contrary, the prevalence of low visual acuity (BCVA < 0.3) was 30% among patients with nasal perifoveal RNFL < 38 µm at the final timepoint, and their mean BCVA was 0.46, which was significantly different compared to the other group (Fig. 1B). 

Figure 1.

 

(A) BCVA difference between two groups with pre-operative nasal perifoveal RNFL thickness ≥ or < 38 µm. *P < 0.05. (B) Composition of eyes of different thickness group in classified BCVA outcome post operation. BCVA, best-corrected visual acuity; RNFL, retinal nerve fiber layer.

Linear regression analysis revealed an approximately weak linear correlation (R² = 0.12) between postoperative BCVA and nasal parafoveal RNFL thickness (Fig. 2). 

Figure 2.

 

Fitted curve (dotted line) and equation of linear regression revealing correlation between preoperative nasal perifoveal RNFL thickness and postoperative BCVA. RNFL, retinal nerve fiber layer; BCVA, best-corrected visual acuity.

To elucidate the predictive effect of nasal perifoveal RNFL thickness on BCVA post silicone oil removal, we used two patients distributed into different groups according to their postoperative nasal perifoveal RNFL thickness as examples to deliver a more intuitive impression (Supplementary Fig. S1, Supplementary Table S1). 

Discussion

In the present study, the efficacy of vitrectomy combined with silicone oil tamponade was confirmed, as most patients showed improvements in visual function and retinal structure. Notably, patients with different essential pathological conditions demonstrated improvements of varying extents. Furthermore, we determined pre-oil-removal nasal perifoveal RNFL thickness as a potential structural risk factor for late-period visual acuity after silicone oil removal and verified its prognostic potential in real clinical cases. 

Risk Factor of Poor Visual Outcome

We found that pre-oil-removal RNFL thickness in the nasal perifoveal quadrant was significantly correlated with BCVA outcome 12 months post oil removal, suggesting a profound impact of retinal structure on visual function. Although RNFL thickness has been valued as being closely related to visual function for dacades,^9,10 this finding appears interesting because the nasal perifoveal RNFL has seldom been endowed with such importance. Conversely, in one of the previous investigations of patients with normal-tension glaucoma, a significant BCVA decrease was identified in glaucomatous eyes, which demonstrated RNFL defects in the temporal quadrants rather than in those with RNFL defects in other areas.¹¹ 

From an anatomic view, the RNFL is enriched in horizontal neuron fibers from retinal ganglion cells, and within nasal perifoveal quadrants, most of the papillomacular and papillofoveal bundle of fibers dominate the crucial centric vision from the macula.¹² In recent studies, RNFL optical texture analysis was applied in patients with glaucoma and ocular hypertension (OHT) to obtain the pattern of RNFL defects, in which the papillomacular bundle appeared to be one of the most involved areas. According to a cross-sectional study enrolling 204 eyes with early glaucoma, RNFL defects in papillomacular and papillofoveal bundles were demonstrated in 71.6% and 17.2% of the eyes, respectively.¹³ In another study, approximately 40% of 600 eyes with OHT exhibited RNFL defects in papillomacular bundles and 10.8% in papillofoveal bundles.¹⁴ These evidences partially explain the relevance of nasal perifoveal RNFL thinning and the loss of visual acuity, calling for more exploration of the function and structural characteristics of the nasal perifoveal axonal fiber bundle, under both physiological and pathological conditions. 

Marker for Oil Removal Timing and Prognostic Effect of Post-Removal Visual Benefit

Based on our findings, 38 µm is a reliable touchstone that provides surgeons with a reference for oil removal decision making and timing. As previously reported, emulsified silicone oil droplets have been identified within the retina and optic nerve in both pathologic specimens of enucleated eyes and, more recently, in the eyes of living patients using SD-OCT imaging.^15–17 Moreover, intraocular structural detriments can be caused by the toxicity of silicone oil incorporated into the posterior segment structures via direct contact or indirect pressure.¹⁸ To different degrees, the number of patients experienced loss of thickness in the whole retina or in certain layers after vitreous oil tamponade due to the physical and chemical interferences of silicone oil. Therefore, we suggest oil removal procedure to be carried out as early as possible, in oil-endotamponaded eyes with stably flatted retina, especially that the nasal perifoveal RNFL thickness reduces to approximately 38 µm. On the other hand, the necessity of performing PPV with oil tamponade in vulnerable patients who do not fully meet the criteria should be further evaluated under a balance of long-term benefit and harm. 

From the point of prognosis, patients with nasal perifoveal RNFL thinner than 38 µm are inferred as candidates who obtain less visual function benefit from the silicone oil removal surgery. This underscores the necessity of predictive interventions and daily health care for late-onset retinal degeneration and visual impairment among vulnerable patients as long as possible. Neuroprotective medicines may be provided to support the regeneration of the lesioned nerve fiber layer when necessary. Precautions for poorer visual improvement should be taken as early as possible, along with clinical advice on rehabilitation for low vision. 

Indications for Post-Surgical Surveillance

As is internationally recognized, the World Health Organization criteria defines that patients with visual acuity < 0.3 are classified as “disabled” low vision or blindness population who usually encounter problems in daily life, study, and work.¹⁹ Our study verified immensely increased tendency of low vision and blindness in post-oil-removal patients with primary nasal perifoveal RNFL thickness < 38 µm, implying the critical importance of OCT-based evaluation of RNFL thickness during the follow-up after PPV and oil tamponade, with the aim of fast judgment of postoperative recovery. 

Indeed, the accuracy and credibility of our linear regression equations were unsatisfactory, indicating that more factors should be added to construct a precise predictive model of BCVA outcomes. From a clinical view, several factors linked to RNFL atrophy-like defects should be carefully observed during clinical practice and surveillance, including the existence of constant high IOP,²⁰ optic disc pit,²¹ optic nerve hemorrhage,²² and parapapillary choroidal microvasculature dropout,²³ as well as many systemic comorbidities, such as arterial hypertension, diabetic mellitus,²⁴ multiple sclerosis,²⁵ and neurodegenerative diseases.²⁶ Besides, complete ocular examinations and systemic medical assessment are regularly required over preoperative conditions as well as postoperative recovery. Given the universal and accumulative harm of high IOP on retinal remodeling,²⁷ it is particularly advised to track retinal structural changes during and after oil tamponade while carefully maintaining a normal and stable IOP level in the affected eyes. Patients should be reminded to seek internal medical consultation and prevent harmful interference on the retina to ensure a successful surgical outcome and long-term improved quality of life. 

Study Limitations

The results of our study are multifactorial and based on the retrospective, comparative, and observational characteristic of this investigation. First, except for age and sex, the patients were not randomly distributed between various controlled parameters, including the criteria of PPV and oil tamponade, and the intrinsic ocular nature, such as axial length and baseline BCVA, thus limiting the generalizability of the results. Second, as the patients included were treated by the same surgeon with the aim of controlling for confounding surgical factors, a possible deviation in efficacy inevitably exists between individuals and the overall populations of the surgeon. In addition, some of the technically missing data led to pseudo-significant and pseudo-nonsignificant differences between the study groups, which were most obvious for OCT measurements in the inferior perifoveal quadrant. Finally, this study was unable to accurately reflect the real-world prevalence of poor visual acuity in oil-tamponaded eyes with potential structural damage owing to the limited population. Despite these limitations, we achieved the main goal of examining the risk factors and their relevance to BCVA reduction to better assist in determining the post-silicone oil removal prognosis. 

Acknowledgments

Supported by the National Key R & D Program of China (2022YFC2502800 to X.S.), Science and Technology Commission of Shanghai Municipality Innovation Action Plan Medical Innovation Research Special Project (23Y11901300 to J.S.), Science and Technology Commission of Shanghai Municipality (20Z11900400 to X.S. and 21ZR1451500 to T.L.), National Natural Science Foundation of China (82101168 to T.L.), and Shanghai Collaborative Innovation Center for Translational Medicine (CCTS-202202 to X.S.). The funding organization had no role in the design or conduct of this research. 

Disclosure: S. Wang, None; Z. Wang, None; H. Jia, None; H. Wang, None; T. Li, None; J. Sun, None; X. Sun, None