Fovea Morphology After Vitrectomy in Eyes With Full-Thickness Macular Hole Coexisting With Diabetic Retinopathy

Abstract

Purpose:

This work presents the effects of vitrectomy with an inverted internal limiting membrane flap in full-thickness macular holes (FTMHs) in eyes with diabetic retinopathy (DR).

Methods:

Vitrectomy with the inverted and temporal inverted internal limiting membrane flap technique was performed in all cases. Inclusion criteria were FTMH, diabetes treated with oral drugs or with insulin for at least 5 years, spectral–domain or swept–source optical coherence tomography performed before surgery then 1 week, 1, 3, 6, 12, and 18 to 36 months after surgery.

Results:

FTMH with nonproliferative DR (NPDR) was noted in 20 eyes and with proliferative DR (PDR) in 5 eyes. In PDR the margins of the macular holes (MHs) were always detached, and the MHs tended to be larger than in NPDR. The minimum diameter–to–base diameter ratio was 1:2 in NPDR and 1:10 in PDR. Postoperatively central retinal thickness and visual acuity (VA) were significantly lower for PDR. Twenty–one of 25 FTMHs were closed after the first surgery, and all were closed after the second surgery. VA improved, final results did not depend on FTMH diameter. Photoreceptor defects decreased in size but were present in all cases 12 months after surgery.

Conclusions:

Morphology of MHs in NPDR and PDR was varied. FTMH in NPDR resembled idiopathic cases; morphology of FTMH with PDR was different. In PDR, FTMHs were larger and the fovea was usually detached; retina thinning was also observed. Final VA depended on initial VA and the severity of the disease (NPDR vs PDR).

Introduction

Full-thickness macular holes (FTMHs) have been reported to coexist with diabetic retinopathy (DR) as a rare complication observed in about 1.2% of cases scheduled for surgery. ¹ There are limited data on the prognosis and treatment of FTMHs in the course of diabetic disease. Although no longer used, it was historically advised to apply laser around the macular hole (MH). ^2,3 The literature reports that DR is a factor negatively influencing the results of FTMH surgery. ^4,5 This is probably due to a potentially different etiopathogenesis of diabetic MHs when compared with idiopathic cases. Idiopathic MHs occur in the course of anteroposterior and tangential traction, whereas in DR, an additional factor might be rupture of the cystoid macular edema, most commonly associated with retinal atrophy. ⁶ Retinal atrophy always coexists with loss of retinal outer layers, which might explain the limited prognosis in cases of coexisting MH and diabetic eye disease.

Our group first described the inverted internal limiting membrane (ILM) flap technique in a 2010 prospective, comparative study of large, chronic MHs. ⁷ Since that time, multiple comparative studies have also confirmed the technique’s results in MHs associated with high myopia with and without coexisting retinal detachment. ^{8 -10} As a result of the technique’s success in different types of MHs with worse prognoses, we concluded that it might also be successful in patients with coexisting diabetes and MHs, because diabetes worsens outcomes of MH. ^4,5

Thus, the aim of this study is to describe the results of the inverted ILM flap technique and the retina morphology of MHs coexisting with DR using swept-source optical coherence tomography (SS-OCT).

Methods

A retrospective, interventional study was conducted. The database of the Ophthalmic Clinic Jasne Blonia was reviewed to find the simultaneous diagnosis of MH and diabetes during the years 2012 to 2017.

All patients signed an informed consent. The Ophthalmic Clinic Jasne Blonia Institutional Review Board Ethics Committee retrospectively approved the study. Inclusion criteria were as follows: FTMH, diabetes, vitrectomy performed, and spectral-domain OCT (Copernicus HR, Optopol) or SS-OCT (Triton, Topcon) before surgery, then at 1 week (±3 days), 1 month (±1 week), 3 months (±1 month), 6 months (±1 month), 12 months (±2 months), and 18 to 36 months after surgery. Exclusion criteria were high myopia and chronic uveitis. In each eye, all examinations were performed with the same device.

All eyes had a complete ophthalmic examination with slit-lamp, intraocular pressure measurement, and a dilated fundus examination.

In all eyes, vitrectomy with the inverted ILM flap technique was performed, usually within 3 months of presentation. After core vitrectomy, trypan blue staining (MembraneBlue Dual, Dutch Opthalmic Research Center) was administered to the vitreous cavity and left there for 45 seconds. Afterward, the ILM was not peeled completely but attached at the margins of the MH and turned over to cover the hole. Some eyes were operated on with the temporal inverted ILM flap technique, in which ILM was peeled only from the temporal side of the fovea. This modified technique was proved to be as efficacious as the original inverted ILM flap technique regarding rate of fovea closure and visual acuity (VA) results and has the advantage of resulting in fewer retinal nerve fiber layer defects. ¹¹

The following data were collected: age, sex, initial VA and VA during follow-up, type of DR (nonproliferative [NPDR] vs proliferative [PDR]), MH diameters, type of MH closure, and phakic status.

In OCT we measured the minimum and maximum diameter of the MH and analyzed the morphology of MHs with the minimum diameter–to–maximum (base) diameter ratio. For the maximum diameter, that is, base diameter, of each FTMH we measured the distance between the margins of the FTMH at the level of the retinal pigment epithelium. After surgery we recorded the presence of photoreceptor defects and the type of fovea contour (U-type, V-type, irregular type, “flap closure” type, flat closed, or elevated open). ^12,13 Additionally, we noted of the presence of any ILM flap remnants and monitored them for potential proliferation after surgery.

Statistical analysis was performed with SigmaStat 3.5 for Windows. Main outcome measures were closure of the MH and VA. Comparisons of continuous variables were performed using the independent-samples t test and Mann-Whitney U test. The Spearman correlation coefficient was used to analyze the correlation between the continuous variables. A value of P less than .05 was considered statistically significant.

Results

Twenty-five eyes of 21 patients (10 women, 11 men) were included in this study with a mean age of 66 years (range, 30-80 years, median 68 years). NPDR was noted in 20 eyes (Figures 1 and 2) and PDR was noted in 5 eyes (Figures 3 and 4).

Figure 1.

Macular hole (MH) in a 65-year-old man with nonproliferative diabetic retinopathy. Photoreceptor defects persisted up to 12 months after surgery (B-D, black arrow). (A) Swept-source optical coherence tomography (SS-OCT) before surgery. Visual acuity (VA) 20/250 (1.1 logMAR). Point a shows the base MH diameter; point b shows the minimum MH diameter; point c shows the left arm length; and point d shows the right arm length. (B) SS-OCT 1 month after surgery. VA 20/200 (1.0 logMAR). (C) SS-OCT 6 months after surgery. VA 20/125 (0.8 logMAR). (D) SS OCT 12 months after surgery. VA 20/63 (0.5 logMAR).

Figure 2.

Macular hole (MH) in a 73-year-old woman with nonproliferative diabetic retinopathy. (A) Full-thickness MH. Minimum diameter was 937 μm, and base diameter was 2072 μm. (B) One week after surgery, the MH was closed, and a photoreceptor defect was still visible. The inverted internal limiting membrane flap is indicated by the white arrow. (C) Six months after surgery. (D) One year after surgery, diabetic macular edema developed. Remnant of the inverted internal limiting membrane flap is indicated by the arrowhead.

Figure 3.

Macular hole (MH) with fovea detachment in a 68-year-old man with proliferative diabetic retinopathy (DR). Vitreous is still attached. The appearance of full-thickness MHs in proliferative DR differs from cases of nonproliferative DR and idiopathic cases. Usually the base MH diameter is 5 to 10 times wider than the minimum MH diameter. (A) Minimum diameter was 589 μm, base diameter 5485 μm, and visual acuity (VA) 20/1000 (1.7 logMAR). (B) One month after surgery, the MH was closed but subretinal fluid was still visible, and VA was 20/320 (1.2 logMAR). (C) Nine months after surgery, VA was 20/500 (1.4 logMAR). Retinal thinning at the fovea is visible.

Figure 4.

Macular hole (MH) in a 72-year-old man with proliferative diabetic retinopathy. (A) A subfoveal hard exudate was visible under the MH. Visual acuity (VA) was 20/1000 (1.7 logMAR). (B) During surgery the exudate was evacuated during posterior hyaloid removal. A full-thickness MH with retina elevation was visible 1 week after surgery. The arrows indicate the inverted internal limiting membrane (ILM) flap. (C) One month after surgery the hole (ILM) still open, but the retina (ILM) only slightly elevated. Two months after surgery the retina was attached. Arrowheads indicate possible remnants of the internal limiting membrane. (D) Three months after surgery the MH was closed, and VA was 20/500 (1.4 logMAR). (E) Twelve months after surgery, the MH was closed. Photoreceptor defects were visible. Final VA was 20/500 (1.4 logMAR). Retinal thinning at the fovea is visible.

Panfundus laser photocoagulation was performed prior to vitrectomy in 3 eyes with NPDR and in 4 eyes with PDR.

In the PDR group, a tractional retinal detachment was present in addition to FTMH in 3 eyes before surgery, and silicone oil was used in each case. In 1 of them, FTMH formation was observed after silicone oil removal. In another eye, additional vitreous hemorrhage was present, and the MH was noted after blood removal. In 1 eye with PDR in which silicone oil was not used, subfoveal hard exudate was present (see Figure 4), which was evacuated during posterior hyaloid removal.

The mean minimum MH diameter (both groups combined) was 535 µm (range, 172-937 μm), and the mean maximum MH diameter was 1532 µm (range, 247-5584 μm).

MHs tended to be smaller in the NPDR group when compared with the PDR group. Mean minimum MH diameter was 504 µm in the NPDR (range, 139-937 μm, MHs of a minimum diameter less than 400 µm were noted in 7 patients) and 725 µm in the PDR (P = .05) cohorts. Mean maximum (base) MH diameter was 974 µm and 4880 μm in the NPDR and PDR groups, respectively (P = .002).

Not only the size but also the morphology of the MHs differed. In PDR the margins of the MH were elevated in all cases and the retina was detached in the foveal area. The difference between MHs in NPDR and PDR might be best presented as minimum diameter–to–base diameter ratios. The minimum diameter–to–base diameter ratio was 1:2 in NPDR (similar to idiopathic cases) and 1:10 in PDR cohorts.

The MHs were successfully closed in 21 cases (84%) after the first surgery and in all cases after a second surgery. Air was used as a tamponade in all cases with NPDR and in 2 cases with PDR; silicone oil was used in the remaining 3 cases with PDR. Silicone oil was removed within 3 to 6 months after surgery.

The closure process was different when we compared eyes with NPDR with those with PDR. In cases with NPDR, 1 week after surgery we observed 9 cases of U-type closure, 2 of V-type closure, 5 of flap closure, and surgical failure was noted in 4 cases. Flap closure was still present in 2 cases 6 months after surgery. At the end of follow-up, U-type closure was noted in 12 eyes; the rest had V-type or irregular closures (Table 1). In cases with PDR, the FTMH closed after surgery in 4 of 5 eyes 1 week after surgery. Foveal detachment was present postoperatively for up to 6 months in 3 of these (see Figure 3). In 1 eye FTMH closure was first noted 3 months after surgery (see Figure 4).

Table 1.

Type of Closure of Full-Thickness MHs After Vitrectomy With the Inverted Internal Limiting Membrane Flap Technique in Eyes With Nonproliferative DR.

	1 wk	1 mo	3 mo	6 mo	12 mo	18-36 mo
U-type	9 eyes	12 eyes	11 eyes	11 eyes	12 eyes	12 eyes
V-type	2 eyes	2 eyes	3 eyes	4 eyes	4 eyes	4 eyes
Irregular	0	1 eye	2 eyes	1 eye	4 eyes	4 eyes
Flap closure	5 eyes	3 eyes	2 eyes	2 eyes	0	0
Elevated open	4 eyes	2 eyes	2 eyes	2 eyes	0	0

Ellipsoid zone defects were present in all studied eyes 1 week after surgery (mean diameter 699 μm). They decreased in size but were still visible in all cases 12 months later (mean diameter 139 μm) (P<.05) (see Figure 1, black arrow, 2C, 3C, and 4E). Central retinal thickness was significantly lower for the PDR group when compared with the NPDR group 12 months after surgery (P<.05).

Remnants of ILM were observed in 7 eyes 12 months after surgery. Neither its proliferation nor epiretinal membrane growth was observed in any case, and VA did not differ in these eyes when compared with the rest of the study eyes (P = .4).

VA improved continuously after surgery (Table 2) and the improvement was statistically significant (P = .002: .001 for NPDR group, .2 for PDR), but there was no statistically significant correlation observed between initial MH diameters and final VA (P >.1). The Spearman rank order test confirmed correlation between the severity of disease and final VA (P = .01).

Table 2.

Improvement of Visual Acuity in Eyes With DR.

	NPDR	PDR
Before surgery	20/400 (1.3 logMAR)	20/500 (1.46 logMAR)
1 wk postoperatively	20/160 (0.95 logMAR)	20/400 (1.32 logMAR)
1 mo postoperatively	20/125 (0.85 logMAR)	20/400 (1.32 logMAR)
3 mo postoperatively	20/125 (0.83 logMAR)	20/400 (1.35 logMAR)
6 mo postoperatively	20/125 (0.81 logMAR)	20/250 (1.15 logMAR)
12 mo postoperatively	20/100 (0.73 logMAR)	20/320 (1.26 logMAR)
Final	20/63 (0.5 logMAR)	20/200 (1.0 logMAR)

Abbreviations: NPDR, nonproliferative diabetic retinopathy; PDR, proliferative diabetic retinopathy.

Conclusions

We noticed in our study of FTMH coexisting with diabetes that the ratio of female to male patients was 10:11, whereas idiopathic FTMH is far more frequently observed in the female population than the male. ¹⁴ Furthermore, although the morphology of FTMH in NPDR is similar to idiopathic cases, the morphology of FTMH in PDR differs significantly from idiopathic cases described in the literature.

One of the most important mechanisms of idiopathic FTMH formation is vitreomacular traction (anteroposterior and tangential). In diabetes the situation may differ somewhat. FTMH formation after rupture of macular edema has been reported. ⁶ In our study, we observed FTMH formation in an eye after silicone oil removal, thus we had to exclude tractional forces. In another eye (see Figure 3) the vitreous was still attached. Thus, defects in nutrition and oxygenation of the central retina and choroid must be taken into consideration when investigating FTMH formation in diabetic patients.

The healing processes were significantly altered in all diabetic FTMHs, fovea shape regeneration was limited, and photoreceptor defects persisted for many months after surgery in all eyes.

FTMHs in the PDR Group Before Surgery

Before surgery, in PDR with FTMH we noted the coexistence of additional foveal detachment in all eyes and a different relation of minimum and base diameter when compared with NPDR with FTMH and idiopathic FTMH. In eyes with NPDR the ratio of minimum diameter to base diameter was usually 1:2, which is also typical for idiopathic FTMH cases. ^15,16 In the cases of PDR in the present study, the ratio was 1:10. Additionally, foveal detachment was visible in all cases of coexisting PDR and FTMH. Foveal detachment was also sometimes visible in eyes with FTMH and high myopia; however, the minimum diameter–to–base diameter ratio was reported to be 1:2 in myopic eyes. ⁹ In summary, FTMHs coexisting with PDR seemed to present a different morphology from all previously described types of FTMH.

FTMHs in the NPDR Group After Surgery

Postoperatively, the healing process seemed to be prolonged. In idiopathic cases it has been reported that U-type closure is observed in about 30% of eyes 1 week after surgery, and in about 85% 1 year later. ^12,17 In the present study the number of eyes with U-type closure barely increased over time (see Table 1). Flap closure is a recently described closure type, probably occurring in holes that might have remained open if the inverted ILM flap technique had not been used. ¹³ One week after surgery for idiopathic FTMHs we saw flap closure in about 15% of cases. No additional flap closures were reported at the 1-month follow-up, but many of the MHs closed with flap only developed into a U-type closure. ¹³ In cases of diabetic FTMHs, flap closure was present in 25% of eyes with NPRD, and half of them persisted up to month 6.

In the NPDR group we observed photoreceptor defects, decreasing in size but still present in all cases 12 months after surgery (Figure 1, Ab-Ad). In idiopathic cases it has been presented that the ellipsoid zone is continuous in 70% to 88% of cases 12 months after surgery. ^12,18 Foveal lucencies, as previously described by other authors, were also noted to disappear within 3 to 11 months after surgery. ^19,20 Ellipsoid zone recovery was found to correlate with improvement in VA in idiopathic cases. These findings, in turn, might explain why VAs remain decreased in eyes with FTMH coexisting with DR. In 2 eyes with NPDR, diabetic macular edema developed after successful FTMH closure (6 months in 1 case, 24 months in another).

FTMHs in the PDR Group After Surgery

Postoperatively in PDR eyes, despite successful closure of FTMH, we observed prolonged subretinal fluid in some cases (see Figure 3). In other cases, reabsorption of subretinal fluid occurred first and enabled closure of the FTMH (see Figure 4).

In addition to photoreceptor defects, retinal thinning also occurred in all eyes (see Figure 3C and 4E). This might be due to vascular abnormalities, increased foveal avascular zone, and insufficient nourishment and oxygen supply to the central retina. As a consequence, even if the MH remains closed, VA remains limited in such cases.

We confirmed the correlation between the severity of disease and final VA (P = .01).

Hyperreflective Remnants of ILM

Theoretically, when using the inverted ILM flap technique we leave remnants of the ILM on the macular surface. There may be concern, especially in diabetic eyes, regarding the potential for proliferation. In this study, we noted hyperreflective remnants of the ILM in 28% of eyes (7 of 25). In an earlier study we observed such remnants in 44% of idiopathic cases. ¹³ The ILM remnants did not grow over time, did not proliferate or progress to epiretinal membrane, and did not influence VA.

FTMHs in DR—Anatomical Results

MHs coexisting with DR were previously described as having a worse prognosis than idiopathic cases. Ghoraba reported 14 cases of FTMHs and DR (without OCT examination). ¹ Owing to the short observation period, he was able to report on only 8 patients, 6 of whom experienced hole closure (75%) and none of whom reached a VA of more than 20/100. Yeh and colleagues reported a success rate of 82.6% in their series of 23 eyes. ²¹ Yan et al reported similar anatomical results (83%). ²² These data must be interpreted with caution because OCT was available in less than half of this group. Moreover, ILM peeling was not performed in all eyes. More recently, Chen and Yang reported a closure rate of 100% with the inverted ILM flap technique in their series of 10 eyes. ⁴ They reported that autologous flap transplants were performed in 3 cases, however, which might suggest that those were repeated surgeries. In our series, in which all 25 eyes were treated with the inverted ILM flap technique, we noted that SS-OCT confirmed closure of the MH after the first surgery in 84% of cases and in all eyes after a second intervention. Those results are worse than those reported in idiopathic cases.

FTMHs in DR—Functional Results

Visual prognosis was reported to be limited in all previous studies. ^1,2,21,23 Ghoraba reported that none of his patients achieved VA better than 20/100 (Snellen equivalent, 0.2). ¹ Yeh et al and Chen and Yang reported even worse results of 1.43 logMAR postoperatively (Snellen, 0.04) and 1.1 logMAR (Snellen, 0.1), respectively. ^4,21 In the present study we noted continuous VA improvement (P < .05) (see Table 2). The improvement was far greater in the NPDR group (final VA 20/63 [0.5 logMAR]) than in the PDR group (final VA 20/200 [1.0 logMAR]). A bias of this study is its retrospective nature. The results need to be confirmed in future prospective evaluations.

Summary

In summary, our study offers the new finding that diabetic FTMHs seem to have a different morphology and healing process from what previously have been described for idiopathic cases. In the NPDR group, the minimum diameter–to–base diameter ratio is 1:2, similar to idiopathic cases. In PDR, FTMHs usually coexist with retinal detachment in the fovea and have a minimum diameter–to–base diameter ratio of 1:10. FTMHs coexisting with NPDR have better visual prognosis than with PDR. Restoration of the photoreceptors might take longer than 12 months. Postoperative foveal contour improves slowly, with some cases exhibiting a “flap closure” type for several months. In PDR subretinal fluid might persist for many months despite successful closure of the MH. Even if it is possible to close the MH with the inverted ILM flap technique and improve VA, visual prognosis is still limited, especially in cases with PDR.