Postoperative Outcomes of 1-Month Silicone Oil Tamponade in Rhegmatogenous Retinal Detachment: A Multicenter Study Open Access

Rhegmatogenous retinal detachment (RRD) is a retinal disease that seriously impairs visual function and retinal structure and needs operation treatment usually, which include scleral buckling and vitrectomy [1]. Vitrectomy combined with silicone oil (SiO) tamponade is a prevalent treatment method for RRD due to its high reattachment and low recurrence rates. SiO serves as an effective substitute for the vitreous body and offers good biohistocompatibility, high transparency, and stable volume [2]. It also supports and presses against retina through its surface tension, thereby enhancing the reattachment rate of RRD. However, prolonged SiO filling may increase the risks of ocular complications, such as SiO emulsification, increased intraocular pressure (IOP), cataracts, keratopathy, secondary glaucoma, etc. Therefore, SiO should not remain in the eye for an extended period and must be removed once the retina is stably reattached. However, the optimal timing for silicone oil removal (SOR) remains uncertain, and there is a scarcity of studies on the ideal duration of SiO filling in the treatment of RRD [3].

The macula is the most sensitive and critical area for human vision, and minor abnormalities in the retinal structure can lead to substantial vision impairment [4, 5]. It has been suggested that there is a significant correlation between central macular thickness (CMT) and central vision [3]. Studies have reported that the inner retinal and ganglion cell layers become thinner after SiO tamponade, which could be potentially associated with visual impairment [6‒8]. Optical coherence tomography (OCT), with its high resolution, can clearly delineate the structure of each retinal layer [9]. However, there are few studies investigating the effect of tamponade duration on retinal thickness, and there is no study investigating the differential effect on retinal structure resulting from short-term (1 month) versus long-term (3 months or longer) SiO tamponade. The study aimed to assess the postoperative outcomes of 1-month SiO tamponade in RRD and the correlation between different SiO tamponade duration and final visual acuity, IOP, macular microstructure, and postoperative complications in patients who underwent successful RRD surgery.

A retrospective analysis of medical records was conducted for 263 patients who underwent pars plana vitrectomy (PPV) for RRD at Zhujiang Hospital, Southern Medical University, Guangzhou Red Cross Hospital, and The Second People’s Hospital of Foshan from January 2019 to December 2023. Inclusion criteria were as follows: (1) patients diagnosed with RRD through fundus examination using a slit-lamp and ocular B-ultrasound, with no significant abnormalities in unaffected eyes; (2) affected eyes required PPV combined with SiO tamponade; (3) the age range of patients was from 18 to 80 years old; (4) the affected eye presents RRD for the first time, while the unaffected eye has no history of illness, and (5) the IOP was within normal range in both eyes preoperatively, and the anterior chamber angles were normal. Patients were excluded from the analysis if they: (1) were with non-RRD, such as traction retinal detachment, exudative retinal detachment, or secondary retinal detachment; (2) were with proliferative vitreoretinopathy (PVR) grade C3 or higher; (3) were with preexisting severe ocular disorders, such as traumatic retinal detachment, macular hole, age-related macular degeneration, macular epiretinal membrane, choroidal detachment, retinal arterial obstruction, retinal vein obstruction, diabetic retinopathy, amblyopia, and strabismus; (4) were with a history of intraocular surgery or ocular trauma; (5) developed detachment during SiO tamponade; (6) were with unavailable or low-quality, unreliable postoperative OCT images, or (7) were without regular follow-up visits.

Patients were divided into three groups based on the duration of SiO tamponade: (1) group 1 – patients who underwent SOR at 1 month after RRD surgery; (2) group 2 – patients who underwent SOR at 3 months after RRD surgery, and (3) group 3 – patients who underwent SOR at 6 months after RRD surgery. The timing for SOR was determined by the surgeons’ discretion, considering factors such as the condition of retinal reattachment, the status of the contralateral eye, and patient preferences.

Patients underwent comprehensive ophthalmologic examinations before surgery and at each follow-up visit, including the day before SOR and 1 month following SOR. The examinations included best-corrected visual acuity (BCVA) in logarithm of the minimum angle of resolution (logMAR), IOP measured with a noncontact tonometer, and an evaluation of the anterior segment and fundus using a slit-lamp microscope. A full ophthalmologic examination with the Spectralis HRA-OCT (Heidelberg Engineering, Heidelberg, Germany) was systematically carried out 1 month after SOR. Measurements were taken for CMT, inner retinal layer thickness (IRLT) (distance between the inner limiting membrane and the external limiting membrane), outer retinal layer thickness (ORLT) (distance between the external limiting membrane and the retinal pigment epithelium), total thickness of ganglion cell layer and inner plexiform layer (GCL-IPL), and submacular fluid using the built-in Spectralis software. In addition, patients completed optometry, fundus photography, B-scan ultrasonography, and biomicroscopy with a three-mirror contact lens.

All procedures were performed by surgeons with the level of deputy chief physician or higher. All patients underwent 3-port PPV and complete fluid-air exchange. The surgeon utilized a 23G trocar for scleral punctures at the 2, 7, and 10 o’clock positions, precisely placed 3.5 mm from the corneal limbus. After complete vitrectomy of both the central and peripheral vitreous, retinal reattachment was ultimately achieved through fluid-air exchange drainage of subretinal fluid via the retinal tear or by performing air-fluid exchange following the use of perfluorocarbon liquids for cases meeting predefined surgical indications, including severe PVR, giant retinal tears, or refractory subretinal fluid accumulation. All retinal tears and breaks were treated with endo-photocoagulation. SiO was injected to tamponade the vitreous cavity for cases with surgical indications including large retinal tears, extensive detachment, or chronic detachment requiring prolonged intraocular tamponade. SiO extraction was performed once retinal reattachment was stable. All patients were treated with levofloxacin eye drops, polyvinyl alcohol eye drops, human epidermal growth factor eye drops, tobramycin, and dexamethasone eye drops after operation.

The main outcome measures included BCVA, IOP, and central retinal layer thickness and macular microstructure which encompasses CMT, IRLT, ORLT, GCL-IPL, and submacular fluid, 1 month after SOR. Secondary outcomes included ocular complications, such as SiO emulsification, increased IOP, cataracts, keratopathy, secondary glaucoma.

Data are presented as reported as median with interquartile range (IQR), for BCVA, IOP, central retinal layer thickness, age, and duration of RRD, while presented as number for sex. Group data were compared using the chi-square (χ²) test, linear regression analysis alongside the Kruskal-Wallis analysis. Statistical analysis was conducted using SPSS 25 software (EPICOS, New York). Statistical significance was set out as p < 0.05.

A total of 263 eyes of 263 patients met the inclusion criteria for the study, with 55, 176, and 32 eyes in group 1, group 2, and group 3, respectively.

The baseline characteristics and postoperative data for each group are presented in Table 1. Out of 263 patients, 101 (38.4%) were female, and 162 (61.6%) were male. Group 1 included 34 (61.8%) men. Group 2 included 107 (60.8%), and group 3 included 21 (65.6%) (p = 0.87).

The median age was 56 years (IQR: 49.0–60.0) in group 1, 56 years (IQR: 50.0–62.0) in group 2, and 55 years (IQR: 48.0–63.0) in group 3 (p = 0.67). There were no significant differences among three groups in terms of the duration of RRD (p = 0.24), preoperative BCVA (p = 0.20), preoperative macula status (p = 0.28), preoperative lens status (p = 0.14), number of quadrants of RRD (p = 0.85), PVR grades (p = 0.51), and preoperative IOP (p = 0.29), whereas a statistically significant difference (p = 0.001) was observed in the number of combined cataract surgeries among the three groups. Groups 2 and 3 demonstrated significantly higher rates of combined cataract procedures compared to group 1, attributable to the greater prevalence of complicated cataracts in these cohorts.

The results of the linear regression analysis between 1-month postoperative BCVA and baseline variables are presented in Table 2. Preoperative BCVA (B = 0.12, p < 0.001), preoperative macular status (B = −0.35, p < 0.001), and PVR grades (B = 0.16, p < 0.001) were significantly associated with 1-month postoperative BCVA. In contrast, duration of retinal detachment (B = 0.00, p = 0.56), age (B = 0.002, p = 0.51), sex (B = 0.12, p = 0.14), and number of quadrants of RRD (B = 0.09, p = 0.17) showed no statistically significant associations. Patients with better preoperative BCVA, intact macular status, and lower grades of PVR demonstrated superior postoperative BCVA outcomes.

In group 1, BCVA improved from 1.6 (IQR: 1.0–3.3) logMAR pre-PPV to 0.7 (IQR: 0.3–1.0) logMAR after SOR (p < 0.001). In group 2, BCVA improved from 2.1 (IQR: 0.9–3.3) logMAR pre-PPV to 0.7 (IQR: 0.4–1.0) logMAR (p < 0.001). In the group 3, BCVA improved from 2.5 (IQR: 1.5–3.5) logMAR pre-PPV to 0.8 (IQR: 0.3–1.2) logMAR (p < 0.001). Figure 1 indicates BCVA in each group improved after SOR when compared to pre-PPV levels. However, there were no significant differences in BCVA (p = 0.66) among these three groups (Table 1).

In group 1, IOP increased from 13.0 mm Hg (IQR: 11.0–15.0) to 16.0 mm Hg (IQR: 13.0–17.8) (p < 0.001). In group 2, IOP increased from 12.0 mm Hg (IQR: 10.2–14.3) to 14.0 mm Hg (IQR: 12.0–16.5) (p < 0.001). In group 3, IOP increased from 12.0 mm Hg (IQR: 10.2–14.0) to 17.0 mm Hg (IQR: 12.0–19.0) (p < 0.001). A statistically significant increase in IOP values was observed after SOR among the three groups. Figure 2 shows IOP changes. IOP significantly differed among groups (p = 0.015) after SOR (Table 1). Pairwise comparison revealed a statistical difference in IOP after SOR between group 2 and group 3 (p = 0.041), while no significant difference was observed among the other groups.

Figure 3 demonstrates the layered architecture of the retina. And Table 3 shows the retinal layer thickness and macular microstructure after SOR for the three groups. One month after SOR, the median CMT was 182.5 μm (IQR: 156.0–214.5) in group 1, 170.0 μm (IQR: 140.3–211.5) in group 2, and 152.0 μm (IQR: 92.3–195.3) in group 3, showing significant difference among the three groups (p = 0.03). Pairwise comparison showed a statistically significant difference in CMT between the group 1 and group 3 (p = 0.02), while there was no statistically significant difference among the other groups. The mean GCL-IPL was 80.5 μm (IQR: 70.0–92.3) in group 1, 73.0 μm (IQR: 65.0–81.3) in group 2, and 65.0 μm (IQR: 56.3–79.0) in group 3, showing significant difference among the three groups (p = 0.006). Pairwise comparison showed statistically significant difference in GCL-IPL between group 1 and group 3 (p = 0.004), while there was no statistically significant difference among the other groups. Besides, the mean IRLT was 97.0 μm (IQR: 77.0–140.3) in group 1, 105.0 μm (IQR: 78.8–131.3) in group 2, and 89.5 μm (IQR: 57.5–126.8) in group 3 (p = 0.31). One month after SOR, IRLT did not differ significantly between the groups. The mean ORLT was 76.5 μm (IQR: 50.8–91.8) in group 1, 70.0 μm (IQR: 53.0–80.0) in group 2, and 64.5 μm (IQR: 43.8–76.5) in group 3 (p = 0.19). One month after SOR, ORLT did not differ significantly among three groups. Figure 4 shows the postoperative retinal layer thickness. Subretinal fluid was observed in 9 patients (group 1), 14 patients (group 2), and 2 patients (group 3), and the remaining patients exhibited no subretinal fluid accumulation, with no statistically significant difference across groups (p = 0.49).

One month after SOR, 2 of 55 patients in group 1 developed ocular hypertension and chronic angle-closure glaucoma, respectively. Two of 176 patients in group 2 were diagnosed with ocular hypertension, and 7 exhibited secondary glaucoma, respectively. Two out of 32 patients in group 3 developed secondary glaucoma. No patients had recurrent retinal detachment 1 month after SOR.

We conducted a multicenter retrospective analysis of 263 patients who underwent PPV combined with SiO tamponade for RRD. Our findings indicated that prolonged SiO tamponade is correlated with increased IOP and the thinning of the CMT and GCL-IPL over time, with favorable postoperative outcomes for 1-month SiO tamponade. Given the potential risks of extended tamponade, it is advisable to remove SiO as soon as anatomically feasible.

There are few studies investigating the comparative effects of short-term (1 month) versus long-term (3 months or more) SiO tamponade on retinal structure. Our retrospective study revealed that the prolonged SiO tamponade correlated with the thinning of the CMT and GCL-IPL. Our findings are in line with previous studies suggesting that SiO tamponade may affect retinal structure, particularly leading to thinning in the macular region and a reduction in foveal sensitivity, which in turn affects the BCVA [7, 10‒12]. Additionally, beyond the duration of tamponade, differences in filling materials have also been explored, with the CMT and GCL-IPL being significantly thinner in eyes subjected to SiO tamponade compared to perfluoropropane filling [10]. However, Dubroux et al. [3] found that SiO filling led to retinal layer thinning, predominantly in IRLT, which resolved upon SOR and was not exacerbated by longer tamponade duration. This contradicts our findings, possibly due to longer SiO exposure in their study cohort (<6 months versus ≥6 months).

The mechanisms behind retinal thinning caused by SiO tamponade are not fully understood. Proposed mechanism includes a retinal response to SiO that triggers macrophage-mediated inflammation and subsequent apoptosis, along with Müller cell dysfunction and potential retinal toxicity. Some hypothesize that the toxicity may result from impaired sodium and potassium pump activity in Müller cells, with prolonged vitreous cavity presence disrupting ion exchange and leading to increased retinal potassium, IRLT thinning, and outer retinal layer degeneration [11, 13]. Besides, studies have shown that SiO tamponade can induce ferroptosis in retinal Müller cells, thereby accelerating retinal damage [14].

In this study, we observed that the postoperative IOP was higher compared to the preoperative levels, with group 3 exhibiting the highest IOP 1 month after SOR. The correlation between SiO tamponade and elevated IOP has been well documented. SiO droplets can enter the anterior chamber from the vitreous cavity, obstructing the trabecular meshwork and causing toxicity, leading to increased IOP [15]. SiO emulsification is a main cause of both anterior and posterior ocular complications, with its duration of tamponade being a critical factor in emulsification and IOP rise [11‒13]. Thus, the longer the SiO tamponade duration, the higher the incidence of complications, with elevated IOP being the most significant. Furthermore, Flaxel et al. [16] reported that elevated IOP persisted even after SOR. Therefore, meticulous monitoring of IOP is essential during and after SiO tamponade, with medication or surgery for IOP management when needed.

Our research on RRD treatment with PPV and SiO tamponade showed significant postsurgical BCVA improvement in most patients, but no correlation with tamponade duration, which are in line with previous studies [3, 17, 18]. However, some researchers found poorer BCVA and reduced retinal blood flow with SiO tamponade, indicating a need for further investigation into its duration’s effect on postoperative BCVA [10, 19].

There are several limitations in our study. Being a retrospective analysis, it is subject to missing patient data and a limited sample size, notably in group 3. Besides, the determination of SiO tamponade and surgery timing by the ophthalmologist may have introduced selection bias. Future research should adopt a prospective design with large cohorts to validate our findings.

In conclusion, our finding indicates that prolonged SiO tamponade is correlated with increased IOP and the thinning of the CMT and GCL-IPL over time, with favorable postoperative outcomes for 1-month SiO tamponade. However, it does not appear to impact the IRLT, ORLT, or final visual acuity. Given the potential risks of extended tamponade, it is advisable to remove SiO as soon as anatomically feasible. Further validation of our findings via larger, prospective studies is warranted.

We would like to thank all participants in this study.

The study adhered to the tenets set out in the Declaration of Helsinki and was approved by the Medical Ethics Committee, Zhujiang Hospital, Southern Medical University (2024-KY-424-01). The need for written informed consent was waived by the Medical Ethics Committee, Zhujiang Hospital, Southern Medical University.

The authors have no conflicts of interest to declare.

This work was supported by President Foundation of Zhujiang Hospital, Southern Medical University (grant No. yzjj2023ms19) (to S.F.), the Science and Technology Plan Project of Guangzhou city (grant No. 202201011443) (to J.L.), the Basic and Applied Basic Research Project of Guangzhou Science and Technology Plan jointly the City and University (college) (grant No. 202201020008) (to W.W.), and the Basic and Applied Basic Research Project of Guangzhou Science and Technology Plan jointly the City and University (college) (grant No. 2023A03J0584) (to L.L.).

S.F. and P.S. conceived the idea of this article. J.L., Y.M., P.G., and C.H. collected the data. Q.L., J.L., P.K., and Y.Y. analyzed the data. J.L. and W.W. wrote the manuscript. L.L. and J.L. revised the manuscript. All authors contributed to the review and editing, and read and approved the final manuscript.

Jingyi Li and Wei Wu contributed equally to this work.

All data generated or analyzed during this study are included in this article. Further inquiries can be directed to the corresponding author.