Risk factors for ellipsoid zone integrity after macula-off rhegmatogenous retinal detachment repair

Background: ellipsoid zone (EZ) layer plays a vital role in visual performance of human retina, hereby we try to nd some risk factors for EZ integrity after primary macula-off rhegmatogenous retinal detachment (RRD) repair. Methods: a retrospective cases study. Patients with macula-off RRD undergoing successful primary retinal repair surgery were reviewed and spectral domain optical coherence tomography images of them were analyzed. Comprehensive preoperative, intraoperative and postoperative clinical factors were screened. Results: A total of 118 patients (118 eyes) were enrolled in this study. The mean age of those patients was 52.16 years old (52.16 ±12.87 years). The follow-up time ranged from 0.1 to 84 months (10.21 ±14.81 months). 54 cases (45.76%) had their EZ fully reconstructed at nal visit. The mid-quartile time of EZ reconstruction was 14.0 months (95% CI: 11.3 ~20.0 months). Multivariate proportional haphazard regression test revealed that the independent factors were: silicone oil tamponade (Hazard ratio=0.414, p=0.0400), posterior staphyloma (Hazard ratio=0.141, p=0.0021) and disorganization of retinal inner layer (Hazard ratio=0.167, p=0.0166). Conclusion: After successful retinal reattachment for macula-off retinal detachment, the mid-quartile time of EZ recovery was about 14 months. The independent risk factors for EZ recovery might include silicone oil tamponade, posterior staphyloma and disorganization of retinal inner layer.


Background
Ellipsoid zone (EZ) is a thin layer of outer retina, the component of which is considered as the mitochondrial complex recently. As we all know the mitochondria serve as the power plant in all animal cells, accordingly the EZ layer plays a key role in visual performance of human retina. When the neural retina detaches from the pigmental epithelium during retinal detachment, the EZ layer is partially or totally destroyed, with poor visual outcome during this pathophysiologic procedure. In the past several years, many studies have revealed that restoration of the integrity of EZ associated strongly with visual outcome after retinal detachment repair 1 2 . However, to our knowledge, there are rare studies carried on the reconstruction process of EZ reconstruction and the predictive factors for this process, especially in macula-off retinal detachment patients so far. So, in this study, we aimed to nd out some factors that could to some degree help clinicians to predict the EZ recovery after retinal detachment repair surgery.

Patients selection
Enrollment criteria: primary rhegmatogenous retinal detachment (RRD) undergoing uneventful retinal repair surgery; macula detachment con rmed by ultrasonic B scan or spectral domain optical coherence tomography (SD-OCT) before operation. Exclusion criteria: recurrent retinal detachment; traumatic retinal detachment; chronic retinal detachment, de ned as: no acute history of chief symptoms like vision loss or visual eld defect, chief symptoms lasted more than 3 months and corresponded to the detached retina, fundus examination revealed detached retina with concentric demarcation lines around retinal hole accompanied by retina thinning or xed; diabetes mellitus; macular hole or other maculopathy; other intraocular diseases or intraocular surgery that could affect the retinal structure. All the patients were selected between May 2011 and September 2017.
SD-OCT scan protocol and main measurements SD-OCT image of all patients at nal visit was reviewed. A 6mmX6mm image centered on the fovea was captured and the horizontal section passing through the foveal bulge was analyzed (in cases where the foveal bulge was not visible, a horizontal line passing through the steepest part of the foveal excavation was chosen). The disruption of EZ integrity was de ned as the cases in which the integrity could not be evaluated because of folding, undulation, edema, and/or the cases in which ELM continuity was disrupted, obviously. phacoemulsi cation and IOL implantation). DRIL was de ned as derangement of the normal laminar inner retinal structure, when it was not possible for the reader to identify any of the boundaries of the ganglion cell layer, inner plexiform layer, inner nuclear layer, and outer plexiform layer.

Statistical analysis
A survival analysis (K-M method) was used to describe the EZ reconstruction rate after surgery. Fully reconstruction of the EZ layer was considered as the endpoint event. Nonparametric Wilcoxon twosample test and Chi-square test was used to screen relative factors preliminarily. Cox proportional haphazard regression test was applied to con rm the independent risk factors for the EZ recovery. Signi cant value threshold was 0.05 during univariate analysis, while entry p value was 0.1 and elimination p value was 0.05 in multivariate analysis. All of those analyses were done by software SAS 9.2 (SAS Institute Inc, Cary, North Carolina, USA).

Results
A total of 118 patients (118 eyes) were enrolled in this study, with 72 males and 46 females, 73 right eyes and 45 left eyes. The mean age of these patients was 52.16 years old (52.16 ±12.87 years). The followup time ranged from 0.1 to 84 months (10.21 ±14.81 months). 54 cases (45.76%) had their EZ fully reconstructed at nal visit. The mid-quartile time of EZ recovery was 14.0 months, with a 95% con dential interval from 11.3 to 20.0 months (Figure 1

Discussions
In this study, we described the EZ reconstruction process after macula-off retinal detachment surgery. To our knowledge, it has the largest population of cases so far. After retinal detachment, retinal neural cells apoptosis was observed in 24 hours, which peaked by 2 days and dropped to a low level by 7 days 3 .
Lewis GP found that the inner and outer retinal could regenerate after retina reattached in animal experiment, and recently Ra E also con rmed in vivo that the photoreceptor outer segment could regenerate using adaptive optics fundus camera 4 5 . However, this microstructure recovery after successful repair is a time-consuming process. According to the survival function curve in our study, the mid-quartile time was about 14 months, which meant that generally about half of the patients would get their EZ zone fully reconstructed about one year after surgery. Moreover, this recovery rate was much faster in the rst two years, and dropped to a much lower level after two or three years postoperatively.
SO has been used world widely in complicated vitreoretinopathy, meanwhile many toxic effects on intraocular structures were substantialized, including iris, ciliary body, trabecular meshwork, retina and optic nerve, etc 6 7 . Recently many authors have proved foveal thickness decreasing, EZ disruption, and even choroidal thickness decreasing after SO endotamponade 8-10 . The exact mechanism of SO toxicity is still unknown. Slight to moderate immunologic in ammation of SO microbubble in different area is probably acceptable as proved by Wickham L 11 . Another explanation is that the ions exchange function of the muller cells between retina and vitreous uid is barred by the SO which constricts the little space between these two interfaces, reported in vitro-experiments 12 . It should be highlighted that postponed SO endotamponade could damage the whole retinal structure, which should be avoided by clinicians.
A posterior staphyloma is an outward protrusion of all layers of the posterior eye, usually occurring in pathological myopia, in which all layers of retina and choroid degenerate irreversibly, from microvasculature to large vessels in fundus. Yuichiro Tanaka has reported extreme thinning or loss of inner neural retina along the edge of staphyloma 13 . As a result of the whole eyewall atrophy, the regenerative potential of the neuroretina is very limited obviously, so it is very di cult for those cases to aqcuire a satis ed EZ integrity. Meanwhile, in some previous studies the variable of axis length was revealed as a risk factor for visual function recovery or retinal microstructural integrity, but it was not the case in this study, either in univariate (p = 0.6002) or multivariate analysis (p = 0.7774, not presented). In our point, though posterior staphyloma often accompanies with longer axis, but the damages to retinal structure of them are not parallel. The patient with long axis but without posterior staphyloma could still remain prosperous anatomical outcomes after retina repair, maybe due to less atrophic changes on the fundus structure.
Recently DRIL get its focus in many studies. It was reported that DRIL associated with cystoid macular edema in uveitis, and also correlated with anatomical and functional outcomes of macular epiretinal membrane peeling surgery [14][15][16] . DRIL is a condition of disarrangement of inner retinal layers, which represent the transmission pathway from photoreceptor to ganglion cells. However, the mechanism of DRIL formation is also unclear. DRIL may be a generic nding of tissue damage in a variety of retinal pathophysiological procedures, including ischemia, vasculopathy, and blunt trauma. In this study, we found a strong correlation between DRIL and EZ integrity, however it could not be further clari ed whether DRIL was a causative factor for EZ recovery or both of them just shared a common pathogenesis process, which needed future researches We presumed before this study that postoperative subretinal uid might affect EZ recovery, as some other authors proposed. However, in this study subretinal uid was not signi cantly correlated with EZ integrity. The relatively rare cases of subretinal uid in this study might contribute to this discrepancy, as we only con rmed 9 cases (9/118) of subretinal uid (6 in fully reconstruction group and 3 in the other group) through OCT scan, so even larger population should be planned.
As a retrospective study, though we have a relatively large number of cases in this study, it was still not satis ed to comprehensively con rm the risk factors for EZ integrity after RRD repair. Furthermore, as mentioned above, we cannot con rm the causal relationship of some key factors in this study. So, prospective studies or multicenter studies need to be carried out.     Figure 1 Survival analysis of EZ integrity K-M method Fully reconstruction of the EZ layer was considered as the endpoint event. In the 118 patients, 54 cases (45.76%) had their EZ fully reconstructed at nal visit. The mid-quartile time of EZ recovery was 14.0 months, with a 95% con dential interval from 11.3 to 20.0 months. This recovery rate was much faster in the rst two years, and dropped to a much lower level after two or three years postoperatively.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. supplementarymaterialoriginaldataR2.xlsx