Full-thickness macular hole formation in proliferative diabetic retinopathy

Twenty-one consecutive patients (21 eyes) having proliferative diabetic retinopathy (PDR) and fibrovascular proliferation (FVP) with optical coherence tomography (OCT) available before and after full-thickness macular hole (FTMH) formation were retrospectively reviewed. Four types of FTMH formation pathways in PDR were identified and were quite different from those in idiopathic conditions. The activity, severity and locations of FVP varied in PDR eyes destined to develop FTMHs. Type 1 was characterized by epiretinal membrane (ERM) and/or vitreomacular traction (VMT) inducing foveoschisis, intraretinal cysts or foveal detachment, followed by formation of a FTMH or macular hole retinal detachment (MHRD). In type 2, ERM and/or FVP induced lamellar macular hole (LMH) with foveoschisis, followed by the formation of FTMH or MHRD. Type 3 was characterized by the initial tractional retinal detachment (TRD) with foveal cysts and/or foveoschisis and the subsequent formation of MHRD. Type 4 was characterized by TRD associated with foveal thinning, ensued by the formation of MHRD. The severity of FVP was grade 2 in 66.7% of eyes in both types 1 and 4, and grade 3 in 75% of eyes in type 3 while the severity of FVP was more evenly distributed in type 2.

www.nature.com/scientificreports/ before FTMH formation with a minimum follow-up duration of 4 months after the development of FTMH or after operation. All patients received serial OCT examinations and for each patient, the same OCT device was used if possible to avoid error induced by using different devices. RTVue XR or 100 (Optovue, Inc., Fremont, CA, USA) was conducted in 19 eyes (90.5%) and Cirrus HD-OCT (Carl Zeiss Meditec, Inc., Dublin, CA, USA) was used in 1 eye (4.8%). Only one patient who lost to follow-up before 2010 and another 3 patients in their earlier periods (before 2013) were performed with Stratus OCT (Carl Zeiss Meditec, Inc., Dublin, CA, USA). The OCT devices adopted standardized protocols and software for imaging and measuring to ensure repeated evaluation of the same point at different time frames. The horizontal and vertical scans of OCT were used for serial follow-up and comparison. Pre-FTMH, post-FTMH and post-operative ophthalmological examination results including best-corrected visual acuity (BCVA), intraocular pressure, lens status (phakic or pseudophakic), severity, activity (active or mainly fibrotic), extent, and locations of FVP, and the extent of RD were recorded. Relevant OCT parameters such as vitreomacular traction (VMT), the presence of ERM, foveoschisis, intraretinal cysts, foveal thinning, lamellar macular hole (LMH), foveal detachment, central macular thickness (CMT), macular hole (MH) edges (flat or elevated), and MH minimal and basal diameters were documented for analysis. The CMT was provided automatically by the OCT device whereas the MH minimal and basal diameters were manually measured. The OCT images were reviewed and analyzed independently by two of the authors (MCT and LCW). In case of doubt, the senior author (CMY) was consulted and a panel discussion was held to reach a consensus. In this study, ERM was defined as the presence of an irregular and hyperreflective layer over the internal limiting membrane (ILM). Foveal thinning was defined as the CMT (from ILM to above retinal pigment epithelium) measuring 200 μm or less. The diagnosis of LMH was made based on the criteria proposed by the International Vitreomacular Traction Study (IVTS) group 15 , which include an irregular foveal contour, a break in the inner retinal layer at fovea, separation between the inner and outer foveal retinal layers, and an absence of a full-thickness foveal defect with preservation of foveal photoreceptors. FTMH was defined as a full-thickness retinal defect in the macular area with or without surrounding subretinal fluid (SRF). A macular hole retinal detachment (MHRD) was defined as SRF below FTMH extending more than one-disc diameter. Edges of MH were defined as flat when MH edges were not elevated and contained minimal cystic spaces. MH minimal and basal diameters were measured on OCT scans. In cases with MHRD, only MH minimal diameters were measured. MH minimal diameter was defined as the shortest diameter of MH in an area excluding the operculum and could be in the outer or inner retina on OCT while basal diameter was defined as the length of the base of MH. FVP severity was classified into 4 grades based on the severity of vitreoretinal adhesion: multiple-point adhesions with or without 1-site plaque-like broad adhesion (Grade 1), broad adhesions in more than 1 but fewer than 3 sites, located posterior to the equator (Grade 2), broad adhesions in more than 3 sites, located posterior to the equator or extending beyond the equator within 1 quadrant (Grade 3), and broad adhesions extending for more than 1 quadrant anterior to the equator (Grade 4) 16 . Broad adhesion was defined as fibrovascular proliferative membrane having multiple point adhesions and occupying more than 2-disc areas 16 . The locations of FVP were divided into 5 categories 17 : (1) complete arcade type: FVP encompassed both upper and lower arcade vessels; (2) incomplete arcade type: FVP encompassed either upper or lower arcade vessels or their adjacent areas with or without disc involvement; (3) juxtapapillary type: FVP existed around the disc and nasal to the disc; (4) central type: FVP was seen mainly in the macular area; (5) widespread type: FVP involved upper and lower arcade vessels, macular area, and juxtapapillary area or extended beyond equator for more than 1 quadrant. The 'macular area' was defined as the area with a diameter of about 5.5 mm centered on the foveal depression.
Surgical procedure. Surgeries were performed by using a standard three-port pars plana vitrectomy only when clinically indicated (when anatomical and functional improvements after operation were expected) and informed consents were obtained preoperatively. ERM and ILM in the macular areas were removed with or without an inverted ILM flap for premacular membrane which caused traction and structural changes. Visually significant cataracts were removed, and intraocular lenses were implanted when indicated. If necessary, supplementary panretinal photocoagulation (PRP) was done, and intravitreal injection of bevacizumab 1.25 mg was performed at the end of surgery.
Statistical analysis. Means  Before the formation of FTMHs, PRP was performed in 16 eyes (76.2%) and intravitreal injection of antivascular endothelial growth factor (anti-VEGF) was performed in 9 eyes (42.9%). One patient lost to followup after the diagnosis of FTMH was made. The average follow-up duration of the remaining 20 eyes was 44.8 ± 35.5 months (range: 4-117 months; median: 42 months). Spontaneous closure of FTMH was found in 3 of the 20 eyes (15.0%) with a duration of 2, 4 and 19 months, respectively. The other 17 eyes (85.0%) received vitrectomy for treatment of FTMH. Table 1 summarized the data for each eye in this study.
Structural changes of formation of FTMH. Four different types of FTMH formation pathways in PDR could be identified. Type 1 was observed in 6 eyes (28.6%) and was characterized by ERM and/or VMT causing foveoschisis, intraretinal cysts or foveal detachment, followed by formation of a FTMH or MHRD (Fig. 1). Type 2 was found in 8 eyes (38.1%), and in this type, LMH with foveoschisis was the characteristic feature, which was induced by ERM (Fig. 2) or tractional retinoschisis (TRS) (Fig. 3) and it finally progressed into a FTMH or MHRD. www.nature.com/scientificreports/ Type 3 was detected in 4 eyes (19.0%), and was characterized by the initial tractional retinal detachment (TRD) with foveal cysts and/or foveoschisis and subsequent formation of MHRD (Fig. 4). Type 4 was noticed in 3 eyes (14.3%) and was characterized by TRD causing foveal thinning, ensued by the formation of MHRD because of persistent tractions (Fig. 5).

Discussion
In this study, we reported four different types of FTMH formation pathways in PDR. In type 1, the ERM or posterior hyaloid membrane causing vitreofoveal traction showed similar pattern to the prehole condition in idiopathic cases, except that the traction was contributed by multi-layered membranes. Therefore, unlike the idiopathic conditions in which traction is usually released after forming a gap in the inner cysts, traction persists even after separation of the posterior hyaloid from the fovea in PDR cases. The persistent traction exerted from FVP or ERM to the foveal area induced progressive foveoschisis and/or intraretinal cystic change, leading to FTMH or MHRD. One unique characteristic of the FTMH in PDR was that in 83.3% of cases the hole edge was flat, instead of being elevated with obvious intraretinal cysts. We speculated that the tangential traction, and not the anterior-posterior or oblique traction, played a more important role in the FTMH formation. In type 2, LMH with foveoschisis was the specific feature, which developed from tractional retinoschisis secondary to ERM and/or fibrovascular traction. This configuration was quite different from idiopathic LMH which was rarely associated with retinoschisis, and whose natural history was considered to be more stable 18 . Moreover, all eyes in this type had foveoschsis and/or intrareitnal cysts with tractional membrane, suggesting that the mechanism of LMH was a tractional rather than a degenerative process.   www.nature.com/scientificreports/ In type 3, TRD first developed by the severe oblique macular traction from the condensed hyaloid and FVP. However, instead of tension release after macular detachment, the nondetached vitreous hyaloid along with FVP tissue maintained a persistent tangential and anterior-posterior traction on the macula, causing disruption of the fovea and formation of MHRD. In our series, all eyes with this pattern developed MHRD instead of simple FTMH. The ultimate development of MHRD in this type emphasized the persistent and strong nature of the traction force in TRD. In type 4, there was TRD associated with foveal thinning before FTMH formation. One eye in this pattern presented with macula-on TRD initially and subsequent foveal thinning before FTMH formation. Foveal thinning may occur before or after TRD involving the macula. MHRD rather than the simple FTMH ensued in all cases in this type. We attributed the mechanism of foveal thinning mainly to the tractional force on the fovea, but ischemia and neurodegeneration might also play some roles. The inner retina was relatively hypoxic and more vulnerable to metabolic stress induced by diabetes, tissue degeneration or cell loss might lead to retinal thinning 19 . Apoptosis of neuroglial cells, induced by hyperglycemia and advanced glycation end products 20,21 might also contribute to foveal thinning.
Locations, severity and activity of FVP varied in each case. In 66.7% of eyes, the FVP severities were either grade 1 or grade 2 at the time of FTMH formation. This observation indicated that FTMHs could develop even in eyes with mild FVP and mild PDR. The key factor was the traction force exerted on the fovea area. Further, FTMH formation might occur in either active (38.1%) or fibrotic (61.9%) stage of the disease. In the active stage, the traction force on the fovea might be getting stronger as the disease progressed, while in the fibrotic stage, the atrophic macula may contribute to the vulnerability of the structure against a milder traction. We observed at least 4 months after operation since our previous study showed most of the postoperative complications in PDR were managed within 4 months postoperatively 16 . The MH closure rate in this series was high compared with the previous studies 22 .
Improvements in surgical instruments and techniques may play important roles in increasing the MH closure rate, although case selection bias could not be ruled out. Compared to eyes with FTMH without RD, those with MHRD had poorer BCVA, shorter duration of FTMH formation, higher proportion of grade 3 and grade 4 FVP, higher rate of active FVP, lower rate of ERM/VMT, and higher rate of TRD, all these indicating that MHRD was more likely to occur in eyes with more severe PDR.
Our study is limited by the retrospective nature and the small number enrolled. However, in this series, we were able to identify four different types of FTMH formation pathways in PDR, which were quite different from those in idiopathic conditions. The foveal might be subject to multi-directional tractions by the ERM, posterior hyaloid, and/or fibrovascular membrane (types 1 and 2) or in the presence of TRD (types 2, 3 and 4), and these complex and strong tractions could induce retinal tears, LMH (type 2), and foveal thinning (type 4) either from the attached retina or after macular detachment. MHRD was more likely to occur in eyes with severe PDR. Spontaneous closure of FTMHs in PDR might be observed. The activity, severity and locations of FVP varied in PDR eyes destined to develop FTMHs. The understanding of the various configuration patterns of FTMH formation may help in guiding follow-up schedule, surgical planning and management. Whether our classification of MH formation pathways is adequate and thorough enough requires further study with a larger number of cases.