Heavy silicone oil tamponade: a multicentre experience

Introduction

Rhegmatogenous retinal detachment (RRD) can cause significant vision loss, particularly if not treated in a timely fashion. For uncomplicated RRD, current surgical techniques carry a high anatomical success rate (>90%).1 2 Recognised risk factors for failure include choroidal detachment, hypotony, grade C proliferative vitreoretinopathy (PVR).3–5 RRD associated with inferior retinal breaks located at a position between 4 and 8 o’clock hours, particularly with concomitant PVR, remains a surgical challenge. Pars plana vitrectomy (PPV) with intraocular gases and standard silicone oil tamponades are unlikely to provide adequate support for such inferior peripheral retinal pathology.6 7 Recurrent RRD secondary to PVR is estimated to occur in about 5%–10% of all patients with failed primary RD repair.8

Retinal redetachment after surgery is likely to arise by several mechanisms: continued vitreoretinal traction, passage and propagation of fluid through an untreated retinal break, possibly due to intraocular fluid currents. In tamponade the gas or oil bubble prevents flow of fluid through tears in the retina.9 This allows retinal reattachment to be temporarily maintained while a permanent chorioretinal adhesion forms in response to retinopexy. Although inferior RD breaks without adequate tamponade have been reported to reattach,9 this is not replicated consistently in clinical practice and the search for adequate tamponade in these cases continues.

Endotamponades with a higher specific gravity within the vitreous cavity than aqueous are theorised to be useful for treating inferior retinal pathology.6 7 10 Examples include heavy oils and perfluorocabon liquids. The former are used for longer periods. First generation heavy silicone oils (HSO) such as fluorosilicone and perfluorocarbons (introduced in 1990s) were marked with complications such as fibrinous reactions, cataract or even retinal necrosis.11–13 A number of newer generations of HSO of different compositions are available for use: Densiron 68 (Geuder), Oxane HD (Bausch & Lomb, mixture of Oxane 5700, mixed fluorinated and hydrocarbonated olefin (RMN3)) and Alaheavy (Alamedics).6 7

Densiron 68 (Geuder AG, Neu-Ulm, Germany) is a mixture of perfluohexyloctane (F₆H₈) and siluron 5000, giving a mid-range viscosity of 1400 mPas and gravity of 1.06 g/cm³.14 It is licensed for the treatment of inferior or posterior RD in cases of PVR, proliferative diabetic retinopathy and giant retinal tears.14 It was chosen as one of the endotamponades in the HSO study.15 Anatomical success rates with Densiron 68 varies from 33.3% to 81.3%.7 15 16

Here, we present the largest evaluation of patterns of use, clinical outcomes and complications associated with the use of Densiron 68 as an internal tamponade in the treatment of RRDs among multiple vitreoretinal units within the UK.

Materials and methods

This was a retrospective, non-randomised study involving seven National Health Service vitreoretinal units within the UK; Bristol Eye Hospital (University Hospitals Bristol and Weston), Dundee eye clinic (Ninewells hospital, Dundee), Queen Elizabeth Hospital (King’s Lynn, Norfolk), Leicester Royal Infirmary, St Paul’s Eye Hospital (Royal Liverpool and Broadgreen University Hospitals), Norfolk & Norwich University Hospitals and Southend University Hospital. Surgical teams reported their outcomes using a standard anonymised spreadsheet. The tenets of the Declaration of Helsinki were followed in the study.

Inclusion criteria were patients receiving Densiron 68 during primary or recurrent RRD surgery during the period of January 2015 to December 2019 with a minimum follow-up of 3 months post-Densiron 68 removal. Cases receiving other heavy oil such as Densiron Xtra, Alaheavy or Oxane HD were excluded. Cases were collected via a combination of EPR, personal and theatre logs. All surgical procedures were performed by either vitreoretinal fellows or consultants.

Surgeons used their conventional vitreoretinal management to allow intraoperative reattachment of the retina, including use of retinotomies, drainage through breaks, retinectomies or use of perfluoroctane/perfluorodecalin. Patients were not instructed to undertake any particular posture.

Patients who required Densiron 68 tamponade were identified by the vitreoretinal units and their anonymised outcomes were pooled and analysed with Microsoft Excel 2020/GraphPad 2020. The primary outcome measures were primary success, final reattachment, qualified success and failure rates. Primary success was defined as achieving a flat retina at 3 months after one Densiron surgery and no subsequent retinal reintervention other than Densiron removal. Final reattachment was defined as achieving a flat retina at 3 months after repeated surgery following the introduction of Densiron. Qualified success referred to the proportion of patients with a flat retina and Densiron retained rather than removed at 3 months. Failure was defined as inability to achieve a flat retina despite all efforts.

The secondary outcome measures were median duration of tamponade, postoperative visual acuity (VA), initial postoperative uveitis, long-term anterior uveitis, ocular hypertension (OHT) and emulsification of oil. Adequate follow-up was defined as at least 3 months after Densiron 68 removal.

Best-corrected VA (BCVA) was assessed using LogMAR. If Snellen VA had been used at any of the participating centres, this was converted to a LogMAR equivalent value. Very poor VA groups were assigned the following LogMAR equivalent: counting fingers (1.85), hand movement (2.3), light perception (2.6) and no light perception (2.9).17

The outcomes were analysed using Excel for mac 2020 and STATA statistical software V.14. A 5% significance level was used (two tailed). Continuous data were presented as mean (SD) or median (IQR) depending on the distribution of data and categorical variables were expressed as counts and percentages. χ² test or Fisher’s exact test was used to compare proportions between groups. The independent t-test and paired sample t-test were used to compare means from parametric datasets, and the Mann-Whitney and Wilcoxon tests were used to compare non-parametric data.

No patients nor the public were involved in the design, conduct or reporting of this study.

Results

One hundred and thirty four eyes of 134 patients were included in the study. During the study period, eleven patients (8.0%) had long-term Densiron 68 retention. The median age of patients was 64 years old (range 25–91 years), with 91 males (67.9%) and 43 females (32.1%).

The median (range) number of operations prior to Densiron 68 tamponade was 1 (0–4). All patients had RRD, with the most common indication being inferior retinal pathology (N=128; 93.4%) as displayed in table 1. Eighty-six cases (64.2%) had failed previous RD surgery secondary to inferior pathology. The median (range) number of inferior retinal tears (between 4 and 8 o’clock associated with subretinal fluid) was 1 (0–7). Densiron 68 was used in cases presenting with all grades of PVR, including no PVR as shown in table 2. All PVR C cases involved the inferior retina.

Seventy-two eyes (53.7%) were phakic, 63 (47.0%) were pseudophakic and 2 (1.5%) were aphakic at time of Densiron 68 insertion.

The median (IQR) duration for Densiron 68 tamponade was 107 days (70–156). Densiron 68 was used most commonly during recurrent RRD repair (87 eyes; 64.7%) with the remainder using it during primary RRD surgery.

Discussion

The total final reattachment rate of 73.9% from this study of heavy oil tamponade for RRD compare favourably with previous published reports of 33%–92%.7 15 18 19 This total comprised 64.6% of primary surgeries and 79.1% of surgeries for recurrent RRD. Although this difference did not quite reach statistical significance, it does suggest that those cases having Densiron 68 for primary RRD may have a higher failure rate. This may reflect the challenging nature of the cases with primary RRD that had Densiron 68 chosen as a tamponade. A similar reported primary success rate of 71.4% has previously been published in one large series.18 Primary surgery success for inferior PVR with light silicone oil has been reported to be as high as 82%,20 however, this rate was detected in a small single-centre retrospective subgroup of 22 patients and larger scale data are lacking. Our study is multicentred and larger than all previous reports and is therefore encouraging. The major indication was inferior PVR C. Pseudophakic patients achieved better results than phakic patients undergoing recurrent surgery in our cohort, while among males, those undergoing recurrent surgery had better outcomes than those undergoing primary surgery. As discussed above, this may reflect the nature of primary RRD cases and Densiron use, or the early use of Densiron 68 in failed cases. No other significant differences were seen between pooled or subgroup data.

Herbrig et al provided a large previously published series (99 eyes) on Densiron 68 for the treatment of complicated RRD. 74% of their cases had previously failed RRD surgery, which is in keeping with our indications and suggestive of the most common approach for use of HSO. Seventy-eight of 89 eyes (87.6%) that underwent PPV and Densiron 68 endotamponade showed retinal reattachment at 9 months, with 34 eyes (36%) requiring reoperation to achieve reattachment. These further operations included light silicone oil, second heavy oil or gas tamponade. In all eyes with a second Densiron 68 tamponade, the retina was completely reattached. This series compared 21 prospective patients with no previous history of retinal surgery before receiving Densiron 68, with a retrospective similar control group of 21 patients that had received light silicone oil tamponade. Fifteen eyes with Densiron 68 (71%) showed reattachment at 3 months while the heavy oil was still in situ, compared with only nine eyes (43%) with light silicone oil. While it failed to reach statistical significance (p=0.059), this difference may suggest greater efficacy of Densiron 68 in such cases.18 In contrast the HSO study, a multicentre randomised masked controlled trial by Joussen et al, suggested lower reattachment rates with HSO (28.3%) compared with standard silicone oil (1000 or 5000 centistokes; 40.4%) in the treatment of complicated inferior RRD.15 This trial also failed to show statistical difference between the two different oil tamponades in terms of reattachment rates or VA at 12 months follow-up. However, the trial was stopped early due to slow recruitment and may not have been adequately as powered as initially planned.15 In our series we had a higher anatomical success rate (61.5%) in inferior PVR C cases with Densiron 68 compared with the HSO study, accepting that our data is retrospective. Furthermore, we do not compare to light oil so any comparisons with historical data must be tentative. Our findings from the largest dataset on this topic (a decade later) provide consideration to perhaps review this subject prospectively. The causes for redetachments were not investigated as part of this study.

The HSO study did not find any difference in vision comparing HSO against standard silicone oil, but there was an improvement from baseline in both groups.15 While our study showed a VA improvement from baseline to post-oil removal, it failed to show statistical significance, supporting the findings of Kocak and Koc.21 This may reflect the complex nature of these eyes in which retinal reattachment may be a primary goal. However, other HSO studies did show statistically significant improvement in vision and anatomical reattachment.18 19 22

The reported adverse effects of HSO tamponade include raised IOP, inflammatory response, cystoid macular oedema, oil dispersion/emulsification and cataract formation.23 24

In our series, 49 patients (36.6%) required topical treatment for IOP control. Of those, long-term raised IOP persisted in 24 eyes (17.9%). In their smaller series, Caporossi et al found a rate of 34.7% of patients requiring topical treatment, reducing to 26.5% after Densiron 68 removal.25 Other published series showed OHT rates of 12%–27%.15 18 23 26 Eight patients in our cohort (6.0 %) required surgical intervention for IOP management, whereas Stappler et al’s prospective cohort of 122 patients did not require any surgical IOP-lowering treatment, describing a steady decline in OHT rate over 3 months post-HSO removal.7

Nguyen et al found that 42% of patients required removal of oil and/or glaucoma surgery to lower their IOP during the early phase, and 32% required medication for chronic OHT.27 This study had a much higher reported incidence of OHT (48%) and this could be attributed to having a small cohort of 47 patients. Our rates were lower and may represent a closer reflection of the true rate, considering the size of our cohort. Whether Densiron 68 has a greater effect on IOP than light silicone oil is unclear; but our data suggest similar rates, although we acknowledge we have no direct control group.

HSOs are suggested to promote an inflammatory response.6 28 In our series, 49 patients (35.7%) developed anterior uveitis within the postoperative period. Following Densiron 68 removal, 8% of patients required long-term topical anti-inflammatory treatment. Auriol et al reported a 40.7% rate of anterior chamber inflammation with fibrin accumulation and a mean time to removal of Densiron 68 of 14 weeks.29 However, Kocak and Koc found a much lower rate (3%) of anterior inflammation with fibrin.21 Russo et al’s review of HSO concluded that inflammation may depend on tamponade duration.28 However, data analysis of our cohort did not reveal any association between inflammation and duration of Densiron 68 tamponade. Our data must be interpreted with caution, as prospective objective analysis of inflammation was not undertaken. Data on long-term inflammation rates with light silicone oil is limited.

Emulsification is also recognised as a complication of heavy and light silicone oil.6 It has been reported to occur in 8.3%–16% of patients with HSO.7 18 19 30 In our series, the emulsification rate was 10.1%, in keeping with reported emulsification rates of Densiron 68. Densiron Xtra (a newer tamponade) is promoted as being more resistant to emulsification due to its higher viscosity and longer silicone oil molecular chain.31 In vivo confirmation is still awaited.32 Emulsification of light silicone has been reported to occur in up to 40% of cases with long chain silicone oil (5700 cSt)33 and 63.4% in shorter chain silicone oil (1300 cSt). Other series have reported similar disparities over a mean period of 7.3 months (SD 4.2 months, range 1–17 months).34 Our rates of emulsification with Densiron 68 are therefore reassuring.

Cataract is a recognised consequence of PPV.26 While our study did not specifically look at progression of cataract, 85.7% of patients required lens extraction. This could be secondary to multiple factors: pre-existing lens status, previous tamponades, multiple surgeries or use of steroids.6 Our series has similar lens findings to a series of 122 patients by Stappler et al,7 however, Joussen et al did not show conclusive evidence of cataract progression in their randomised control trial.15

There are several limitations to this study. Its retrospective nature limited more detailed interrogation of the outcomes and complications of Densiron 68. Unlike previous published case series, our data offer a high proportion of patients receiving Densiron 68 as a primary endotamponade. We did not collect data on superior retinal breaks, therefore while it is unlikely that any patient with superior pathology received Densiron as primary endotamponade, we cannot exclude this as a possibility that may have contributed to our relatively high redetachment rate. Therefore, this patient group is significantly different from others in published literature, so we recommend cautious interpretation of our results when compared with other studies. Subgroups for analysis were retrospective and therefore limited in interpretation. It is possible that not all cases of Densiron 68 use were identified across all centres, which may represent a source of bias in the study. However, this process was optimised using a combination of personal logbooks, theatre records and electronic patient records. We were unable to compare Densiron to light silicone oil in this study, nor were we able to investigate causes for retinal redetachment.

However, this is the largest series presenting real-world, long-term, multicentre UK data on the use and outcomes of surgery with Densiron 68. Densiron 68 is perhaps not widely used due to lack of familiarity with the injection and removal process. The data presented here should reassure surgeons that Densiron 68 is a useful tamponade alternative for the management of complex inferior RRD, with similar morbidity to standard silicone oil.