Corneal endothelial changes induced by pars plana vitrectomy with silicone oil tamponade for retinal detachment.

Silicone oils are effective intraocular tamponade agents in the treatment of severe retinal detachments, because they maintain the adhesion between neurosensory retina and retinal pigment epithelium, thanks to their ability to remove aqueous humor from the surface of the retina. To understand their effectiveness, it is important to know the characteristics of silicone oils. Patients should be closely monitored due to many complications associated with intraocular silicon oil, such as inflammatory reaction, raised intraocular pressure, refraction disorders, cataract, and emulsification. This study presents corneal endothelial changes and some intraocular complications caused by silicone oil used as an intraocular tamponade agent in the case of vitrectomy for complex retinal detachments. The aim of the study was to demonstrate the damage of corneal endothelial cells after the use of silicone oil in patients with retinal detachment surgery. Endothelial specular microscopy measurements were performed and the changes of the following parameters demonstrated the corneal damage: Mean cell density, coefficient of variation, average cell area, percentage of hexagonal cells, and corneal thickness. Three months postoperatively, a statistically significant decrease was observed in the following analyzed parameters: Mean cell density (P=0.04), and percentage of hexagonal cells (P=0.002); the remaining parameters also had a linear decrease (coefficient of variation, average cell area), but were statistically insignificant. Three months postoperatively, the corneal thickness presented a slight increase. Silicone oils are powerful tools when used wisely and within the limits of their use. These are often recommended in cases of severe detachment of the retina in patients at high risk of experiencing intraoperative complications. Copyright: © Coman (Cernat) et al.

Introduction

The research regarding intraocular tamponade products is one of the most challenging and interesting areas in ophthalmology; it began more than a century ago. In the early 1960s, Dr Paul Cibis began to inject silicon oil in vitro to animals, in order to provide a permanent support to the retina, but the anatomical and functional outcomes were not the expected ones (1-3). In The Netherlands, Zivojnović popularized the technique and contributed greatly to its use (4). In 1992, the Silicone Oil Study, a prospective, multicenter, randomized, controlled clinical trial showed that silicone oil was superior to SF6 and equivalent to the perfluorocarbon gas in the treatment of vitreoretinal proliferation. Following these studies, the USA Food and Drug Administration (FDA) approval for its use as a tamponade product in 1994 triggered marketing and use worldwide (5-9).

The first indications of silicone oil use were: Ocular trauma, severe proliferative diabetic retinopathy, complicated retinal detachment caused by proliferative retinopathy or viral retinitis and giant retinal tears, due to the silicone oil ability to displace aqueous humor from the retinal surface maintain the application of the neurosensory retina to the retinal pigment epithelium (10).

At present, new indications for silicone oil use are possible, such as chronic and persistent macular hole, chronic uveitis with hypotony, retinal detachment due to macular hole in myopic eyes, and colobomatous retinal detachment (11,12). A great advantage of the silicone oil is that it provides support over a long period of time until retinal recovery occurs.

In the case of retinal detachment, silicone oil is usually removed after 3-6 months, because it is thought that it is enough for the eye to recover with minimal risk for the development of proliferative vitreoretinopathy (13).

In addition, in patients who travel by air, in children and elderly patients (who cannot maintain a correct postoperative position), silicone oil is a first choice (14).

Retinal detachment surgery is a relative emergency surgery meaning that patients undergo this surgery within 24-48 h after the onset of vision loss. Furthermore, surgery is the only possible remedy in the event of retinal detachment. Different interventions exist and the choice is made according to the characteristics of the patient's pathology.

The objectives of the surgery sought in retinal detachment are: To heal or plug the tear in the retina, to puncture the liquid present between the two retinal layers, to re-approach the neuroepithelium of the pigment epithelium in order to reconstitute the anatomically and physiologically normal retina, and to create an adhesive scar between the retina and the fundus of the eye, in order to stabilize the retina and prevent recurrence (15,16).

Silicone oil is injected at the end of the surgical procedure, when membrane dissection has already been performed, all lesions are sealed and released from the traction forces. Air-fluid exchange occurs and between 3-4 cm³ of silicone oil of 1,000 or 5,000 centistokes is injected (17).

In the postoperative period, the patient should be placed in ventral decubitus, in order to avoid the contact of silicone oil with the cornea and the posterior face of the intraocular lens (18).

Large population-based studies of retinal detachment have an annual incidence of about 1 in 10.000, and a family aggregation study estimated a 3% lifetime risk at 85.2 years (19). White and Asian populations have similar rates, with a lower incidence among individuals of African descent (20,21). The average age of presentation is approximately 60 years, the sexes being equally affected.

Due to the fact that the surgery for retinal detachment involves retrobulbar anesthesia and to the potential complications induced by the surgery and silicone oil, patients were screened for other associated disease and treatment options (22-29).

However, post-operatory complications of the silicone oil tamponade also occur. The use of silicone oils has no short- or medium-term side effects. Patients should be monitored regularly. They are usually seen several times in the first months, then 3-4 times a year, as long as silicone oil persists in the eye.

Early complications include: i) Post-operatory ocular inflammation that is almost constant. It is related to severe trauma and initial pathology. However, silicone oil, especially heavy silicone oil, is considered pro-inflammatory. Serum anti-silicon antibodies were found in 35.7% of patients with silicone tamponade, and up to 83% of those with intraocular silicone oil (30). ii) Variation of intraocular pressure pertains to early postoperative ocular hypertension which is often evident after silicone oil tamponade. It is related to immediate post-operatory inflammation and, more rarely, to excessive filling, which requires a partial discharge of silicone oil present. Chronic ocular hypertonia may occur, especially in the case of prolonged tamponade. This may be related to decompensation of pre-existing hypertension, prolonged steroid prescription, chronic trabeculitis through trabecular adhesions, or migration of silicon microemulsion particles into trabecular meshwork (31,32).

Medium- or long-term complications include: i) Refraction disorders: Due to its refractive index, silicone oil tamponade causes a change in the patient's refraction. Silicone is responsible for a hypermetropic of 3-7 diopters in phakic eyes and a myopia of 5 diopters in aphakic eyes. It is not just a complication, but rather a disorder that will last only during the intraocular presence of silicone (30). ii) The appearance of cataracts is a frequent complication after silicone oil tamponade. The contact between the silicone bubbles and the posterior capsule of the lens prevents the diffusion of nutrients and leads to the development of a posterior subcapsular cataract. After silicone oil tamponade 62.5% of patients develop cataracts up to 3 months, while the incidence is 100% over 6 years (33). iii) Emulsification is the second most common complication associated with silicone oil tamponade. It is defined as the fragmentation of a single silicone bubble in more bubbles of different diameters. Emulsification changes the silicone tamponade power resulting in a decrease in its ability to block dehiscence. Silicone microbubbles may be able to migrate to the anterior chamber and to the trabecular meshwork, and can cause edema keratopathy or intraocular hypertension (8). Some authors observed emulsification 2 weeks after injection. Previous findings showed that, the average emulsification time of silicone 1,000 was 13.2 months (5-24 months), which was significantly higher than the average tamponade time (34,35). iv) Contact of silicone oil with corneal endothelium can cause corneal decompensation and band keratopathy. Complications are secondary to the migration of silicone oil into the anterior chamber, resulting in discontinuity of endothelial metabolism and precipitation of calcium salts (36).

Literature reported an incidence of keratopathy of up to 30% of patients after 6 months of treatment (37,38).

Thus, the aim of the present study was to evaluate the loss of corneal endothelial cells in patients undergoing complex retinal detachment, which required internal tamponade with silicone oil of 1,000 centistokes.

Materials and methods

Ethics approval and patient consent

The present study is a retrospective, interventional, comparative assessment with consecutive enrolment of patients diagnosed with rhegmatogenous or tractional retinal detachment requiring surgery. All subjects provided written informed consent to be subjected to ocular surgery for retinal detachment, prior to enrolment. Ethics Committee (number 401) approval was obtained from ‘Dr. Carol Davila’ Central Military University Emergency Hospital Bucharest and was conducted in accordance with the Declaration of Helsinki and with the International Standard of Good Clinical Practice (ICH-GCP E6 Step 4).

Patients

A total of 20 patients (7 males,13 females, aged 54-70 years) diagnosed with rhegmatogenous or traction retinal detachment that requires as a method of treatment posterior vitrectomy adjusted with silicone oil endotamponade, were selected from the Department of Ophthalmology of the ‘Dr. Carol Davila’ Central Military Emergency University Hospital in Bucharest.

The inclusion criterion was the diagnosis of rhegmatogenous or tractional retinal detachment requiring surgery. The exclusion criteria included any coexisting corneal or retinal disease, history of eye trauma or any other eye intervention performed in the past other than cataract.

The patients were divided into 2 groups: Group 1 (9 females, 3 males, aged, aged 54-70 years) included subjects who had a natural lens in the operated eye and group 2 (4 females, 4 males, aged 58-69 years) included those who were pseudophakic in the eye where the surgery was performed.

Methods

Non-contact corneal specular microscopy was used to measure the following parameters: Mean endothelial cell density (MCD), average cell area (AVG), coefficient of variation in cell size (CV), percentage of hexagonal cells (HEX) and corneal thickness (CT) at baseline representing surgery and then 3 months after the surgery was completed. Of the several measurements, the one that showed maximum counted endothelial cells was chosen. As a control method, the patient's other unoperated eye was used.

Surgery was performed under retrobulbar anesthesia by the same surgeon for all the patients. After topical disinfection with povidone-iodine, a sterile field and lid speculum were applied. The surgery consisted of: 25 gauge total posterior vitrectomy, locating the retinal hole/holes and performing a laser blockage around the hole and once the retinal attachment was obtained the air exchange was changed to silicone oil of 1,000 centistoke.

Preoperatively and 3 months postoperatively, bilateral corneal specular microscopy was performed in all the patients to count endothelial cells, coefficient of variation, central cell area, percentage of hexagonal cells and corneal thickness.

Patients were subsequently discharged after confirmation of retinal attachment and then re-evaluated after three months, using corneal specular microscopy; the results were recorded.

All determinations of the studied values were performed using similar working techniques. For the processing and systematization of the data, the Excel program of the Microsoft Office 365 suite was used. The graphical representations, as well as the statistical analysis of the data were performed using the same program, together with ‘add-ins’, such as WinStat and XL-stat. For the calculation of the statistical significance of the obtained results, online support was provided by Professor Richard Lowry-Vassar College Poughkeepsie (Poughkeepsie, NY, USA), through the link www.vassarstats.net.

Statistical analysis

In order to establish the relationships between various values of the analyzed coefficients, the average values were calculated, as well as the mean ± standard deviation or standard error of Student's test (t-test), and its statistical significance was represented by P-values. The statistical significance of the results was interpreted according to the value of the coefficient p: P>0.05 indicated the results were not statistically significant; for P-values between 0.05 and 0.001 the results were considered highly statistically significant, while P-values <0.001 were considered very highly statistically significant.

Results

Patient demographic data are presented in Fig. 1. Patient surgical indications were rhegmatogenous and tractional retinal detachment (Fig. 2). Patient distribution according to age is presented in Fig. 3.

Figure 1

Patient distribution by sex. Patients included in the study were predominantly men.

Figure 2

Patient distribution according to their diagnosis. In total, 20% of patients were diagnosed with rhegmatogenous retinal detachment and 80% with tractional retinal detachment.

Figure 3

Patient distribution according to age. The majority of patients were over 60 years of age.

Various parameters

MCD, CV, AVG, HEX and CT values of all groups at baseline, postoperative and three months postoperatively are shown in Fig. 4, Fig. 5, Fig. 6, Fig. 7 and Fig. 8. For the parameters MCD, AVG, CV, HEX a linear decrease was observed both immediately postoperatively (MCD with 1.22%, AVG with 12.05%, CV with 7.50%, HEX with 4.19%) (Table I) and after 3 months (MCD with 3.61%, AVG with 14.04%, CV with 14.08%, HEX with 7.78%) (Table II).

Figure 4

Evolution of mean MCD values for the operated eye.

Figure 5

Evolution of mean CV values for the operated eye.

Figure 6

Evolution of mean AVG values for the operated eye.

Figure 7

Evolution of mean HEX values for the operated eye.

Figure 8

Evolution of mean CT values for the operated eye.

Table I

Comparison between the postoperative values of the analyzed parameters for all patients.

Variables	Operated eye	Non-operated eye
MCD average postoperative	2,171.95 (-1.22%)	2,525.6
Standard deviation	564.92	466.69
Standard error	163.851
95% CI	21.9507-685.3493
T-test	2.158
P-value	0.04[a]
CV average postoperative	25.3 (-7.5%)	28.2
Standard deviation	6.197580173	4.843552415
Standard error	1.759
95% CI	-0.6606-6.4606
T-test	1.649
P-value	0.1
AVG average postoperative	403.75 (-12.5%)	418.3
Standard deviation	118.5752398	125.7756336
Standard error	38.652
95% CI	-63.6970-92.7970
T-test	0.376
P-value	0.7
HEX average postoperative	65.25 (-4.19%)	69.15
Standard deviation	6.08173495	4.475209492
Standard error	1.688
95% CI	0.4820-7.3180
T-test	2.31
P-value	0.03[a]
CT average postoperative	572.95	567.8
Standard deviation	59.23721381 (+2.52%)	66.45720427
Standard error	19.907
95% CI	-45.4492-35.1492
T-test	-0.259
P-value	0.8

aStatistically significant. PFK-CP, Pseudophakic; MCD, mean endothelial cell density; AVG, average cell area; CV, coefficient of variation; HEX, hexagonal cells; CT, corneal thickness; CI, confidence interval.

Table II

Comparison between the 3 months postoperative values of the analyzed parameters for all patients.

Variables	Operated eye	Non-operated eye
MCD average 3 months postoperative	2,119.55 (-3.61%)	2,478.6
Standard deviation	587.5921609	455.2281186
Standard error	166.207
95% CI	22.5810-695.5190
T-test	2.16
P-value	0.04[a]
CV average 3 months postoperative	23.5 (-14.08%)	28.5
Standard deviation	5.599107072	4.82182538
Standard error	1.652
95% CI	-0.3448-6.3448
T-test	1.816
P-value	0.07[a]
AVG average 3 months postoperative	394.6 (-14.04%)	418
Standard deviation	116.9621306	125.8046104
Standard error	38.41
95% CI	-54.3575-101.1575
T-test	0.609
P-value	0.5
HEX average 3 months postoperative	62.8 (-7.78%)	68.65
Standard deviation	6.257795139	4.901785389
Standard error	1.777
95% CI	2.2517-9.4483
T-test	3.291
P-value	0.002[a]
CT average 3 months postoperative	582.6 (+4.25%)	575.85
Standard deviation	60.75063786	65.25049808
Standard error	19.935
95% CI	-47.068-33.3068
T-test	-0.339
P value	0.74

aStatistically significant. Pseudophakic PFK-CP, Pseudophakic; MCD, mean endothelial cell density; AVG, average cell area; CV, coefficient of variation; HEX, hexagonal cells; CT, corneal thickness; CI, confidence interval.

For the CT parameter, a slight increase was observed both immediately postoperatively (with 2.52%) (Table I) and after 3 months (with 4.25 %) (Table II).

Postoperatively, a statistically significant decrease in the parameters MCD (P=0.04) and HEX (P=0.03) (Table I) was identified. A linear decrease in the other parameters (CV, AVG) was evident, but these were statistically insignificant (Table I).

After 3 months postoperatively, a statistically significant decrease was observed at the following analyzed parameters: MCD (P=0.04), HEX (P=0.002) (Table II). The remaining parameters also had a linear decrease (AVG, CV), but were statistically insignificant. Furthermore, after 3 months postoperatively, the CT parameter had a slight increase (Table II).

Comparison of values of the analyzed parameters depending on the lens

Prior to retinal detachment surgery, pseudophakic patients had a greater decrease in MCD, AVG, CV, HEX parameters (Table III, Table IV, Table V, Table VI and Table VII) in comparison with those who had their own lens and underwent the same surgery for retinal detachment. The decrease was observed both immediately postoperatively and after 3 months postoperatively, but this decrease was statistically insignificant. In all patients, the presence of a slight constant postoperative inflammation and increased intraocular pressure immediately postoperatively was observed.

Table III

Comparison between the values of the analyzed parameters, depending on the type of lens.

	Non-operated eye		Operated eye
Variable	Phakic	PFK-CP	Phakic	PFK-CP
MCD average postoperative	2,309.33	1,965.87	2,517.83	2,537.2
Standard deviation	494.86	599.77	515.21	382.27
Standard error	245.61		213.62
95% CI	-844.14 to-246.48		-394.34 to -480.25
T-test	-1.39		0.09
P-value	0.18		0.93
MCD average 3 months postoperative	2,239.08	1,940.25	2,461.41	2,504.37
Standard deviation	554.69	589.96	498.87	379.02
Standard error	259.56		208.14
95% CI	-844.14 to -246.48		-394.34 to -480.25
T-test	-1.15		0.21
P-value	0.26		0.84

PFK-CP, Pseudophakic; MCD, mean endothelial cell density; CI, confidence interval.

Table IV

Comparison between the values of the analyzed parameters, depending on the type of lens.

	Non-operated eye		Operated eye
Variables	Phakic	PFK-CP	Phakic	PFK-CP
CV average postoperative	27.41	22.12	28.5	27.75
Standard deviation	6.31	4.39	5.61	3.34
Standard error	2.57		2.21
95% CI	-10.71 to -0.12		-5.40 to -3.90
T-test	-2.05		-0.33
P-value	0.05[a]		0.74
CV average 3 months postoperative	24.25	22.37	28.83	28
Standard deviation	6.64	3.15	5.01	4.47
Standard error	2.53		2.19
95% CI	-7.2 to-3.45		-5.44 to-3.77
T-test	-0.73		-0.38
P-value	0.47		0.71

aStatistically significant. PFK-CP, Pseudophakic; CV, coefficient of variation; CI, confidence interval.

Table V

Comparison between the values of the analyzed parameters, depending on the type of lens.

	Non-operated eye		Operated eye
Variables	Phakic	PFK-CP	Phakic	PFK-CP
AVG average postoperative	414.16	388.12	428	403.75
Standard deviation	142.68	64.85	150.86	71.1
Standard error	54.15		57.5
95% CI	-139.81 to -87.73		-145.07 to -96.57
T-test	0.48		-0.42
P-value	0.60	0.67
AVG average 3 months postoperative	401.08	384.87	428.75	401.87
Standard deviation	141.8	62.28	150.44	71.99
Standard error	53.61		57.45
95% CI	-128.84 to -96.43		-147.59 to -93.84
T-test	-0.3		-0.46
P-value	0.76		0.64

PFK-CP, Pseudophakic; AVG, average cell area; CI, confidence interval.

Table VI

Comparison between the values of the analyzed parameters (HEX), depending on the type of lens.

	Non-operated eye		Operated eye
Variables	Phakic	PFK-CP	Phakic	PFK-CP
HEX average postoperative	66.25	63.75	69.67	68.37
Standard deviation	4.88	7.27	3.19	5.81
Standard error	2.70		2.01
95% CI	-8.18 to -3.18		-5.51 to -2.93
T-test	0.92		-0.64
P-value	0.36		0.52
HEX average 3 months postoperative	63.75	61.37	65.58	67.25
Standard deviation	3.58	8.67	3.09	6.51
Standard error	2.78		2.15
95% CI	-8.21 to -3.46		-6.86 to -2.2
T-test	-0.85		-1.08
P-value	0.4		0.29

PFK-CP, Pseudophakic; HEX, hexagonal cells; CI, confidence interval.

Table VII

Comparison between the values of the analyzed parameters (CT), depending on the type of lens.

	Non-operated eye		Operated eye
Variables	Phakic	PFK-CP	Phakic	PFK-CP
CT average postoperative	578.41	564.75	571.91	561.62
Standard deviation	45.59	74.44	46.8	87.69
Standard error	26.71		30.03
95% CI	-69.79 to -42.45		-73.39 to -52.81
T-test	-0.51		-0.34
P value	0.61		0.73
CT average 3 months postoperative	587.91	574.62	579.66	570.12
Standard deviation	44.21	78.67	47.94	84.51
Standard error	27.39		29.51
95% CI	-70.83 to -44.25		-71.55 to -52.47
T-test	-0.48		-0.32
P-value	0.63		0.75

PFK-CP, Pseudophakic; CT, corneal thickness; CI, confidence interval.

Discussion

The direct effect of silicone oil on the corneal endothelium in direct touch (39-41) has been studied before, but there are very few studies showing the effect of silicone oil in the vitreous cavity of phakic and pseudophakic eyes on endothelium. Consequently, in the present study, the changes that may occur in the corneal endothelium following retinal detachment surgery adjusted with silicone oil in phakic or pseudophakic patients were examined.

In the present study, we found a significant loss of endothelial cells, as demonstrated by the decrease in MCD, CV, AVG and HEX parameters. During the follow-up period, there were no significant complications caused by the use of silicone oil, except for a slight ocular inflammation, present almost constantly in all operated patients.

In 2014, Goezinne et al (42) who studied five groups of patients showed that the highest loss in MCD at postoperative 12 months was observed in aphakic eyes (39.2%), followed by pseudophakic eyes (19.2%) that underwent cataract surgery during the follow-up period. The eyes that were pseudophakic at the beginning of the study showed an MCD loss of 4.6%. In any case, phakic eyes showed no significant difference. The silicone oil tamponade eyes in the present study included both phakic and pseudophakic eyes, in which a statistically significant reduction (3.61%) in MCD at 3 months postoperatively was found, compared with baseline.

In 2017, Shaheer et al (43) who studied two groups of patients reported that MCD was decreased in both the groups showing a cell loss of 30.48±25.78 in phakic patients group and 77.52±40.03 in pseudophakic patients group, but the decrease in the endothelial cell count was statistically insignificant. Findings of the present study demonstrate statistically significant reduction on MCD both in the phakic and pseudophakic groups.

Silicone oils are powerful tools when used wisely and within the limits of their use. These are often recommended in cases of severe detachment of the retina in patients at high risk of experiencing intraoperative complications. Regular monitoring of these patients is therefore essential, especially when prolonged tamponade is required.

In the present study, we have found a linear decrease in all parameters followed (MCD, AVG, CV, HEX), both immediately postoperatively and after 3 months postoperatively.

The decrease was not statistically significant at all. At MCD, it was statistically significant both immediately postoperatively (P=0.04) and after 3 months postoperatively (P=0.04). For HEX, we also observed a statistically significant decrease immediately postoperatively (P=0.03), and a statistically significant high after 3 months (P=0.002). There was also a greater decrease in patients previously operated on for cataracts; thus, the irido-crystalline diaphragm provides protection against the decrease in the number of endothelial cells during retinal detachment surgery (44).

The study has some limitations. First of all, the low number of patients on whom it was performed. In addition, the duration of the surgery was not measured, which also contributes to the decrease in the number of endothelial cells. Another limitation is the low number of pseudophakic patients included in the study. In conclusion, posterior vitrectomy with internal silicone oil tamponade causes a decrease in the number of endothelial cells. Other factors, such as fluid turbulence, phototoxicity, changes in temperature and pH can also affect the corneal endothelium (45). Further studies are needed to demonstrate potential side effects on the anterior ocular segment of interventions in the posterior ocular segment.