Fig. 1
Improvement of diffuse DME after vitrectomy and posterior hyaloid elevation in a patient with taut posterior hyaloid. (a) Preoperative fundus photograph. (b) Late frame of preoperative FA showing diffuse leakage. (c) Preoperative OCT showing thick hyaloid and macular edema. (d) Postoperative fundus photograph. (e) OCT at postoperative month 3 after vitrectomy and posterior hyaloid elevation with significant improvement
Attached Vitreous with Vitreomacular Traction
Vitreomacular traction (VMT) is associated with foveal distortion, and eyes with this condition often respond favorably to surgical intervention. Best visualized by OCT, VMT is defined as vitreofoveal attachment and traction with perifoveal vitreoretinal separation [29, 34, 35]. Figure 2 offers an illustrative example of a diabetic patient with DME and VMT where the posterior hyaloid is attached at the fovea but the perifoveal hyaloid is elevated.
Fig. 2
Improvement of DME after vitrectomy and posterior hyaloid elevation in a diabetic patient with vitreofoveal traction. (a) Preoperative fundus photograph. (b) Preoperative fluorescein angiogram. (c) Preoperative OCT showing posterior hyaloid attachment at fovea with surrounding perifoveal hyaloid detachment and DME. (d) Postoperative fundus photograph. (e) OCT showing resolution of vitreofoveal traction and DME after vitrectomy without ILM peeling
The Diabetic Retinopathy Clinical Research Network (DRCRnet) Vitrectomy Study was a large prospective study that examined vitrectomy for DME in eyes with at least moderate vision loss and VMT [36]. The study included 87 eyes with VMT based on the “investigator’s evaluation,” baseline VA 20/63 to 20/400, and OCT central subfield thickness >300 μm. Surgical intervention beyond vitrectomy was not standardized. Membrane peel (ERM) was performed in 61 % and ILM peeling in 54 % of cases. At 6 months postoperatively, median OCT thickness decreased by 160 μm and 68 % of eyes had ≥50 % reduction in macular thickness. VA improved by ≥10 letters in 38 % of eyes but deteriorated by ≥10 letters in 22 % of eyes [36]. After separation of vitreofoveal traction, improvement of macular edema was detected on OCT (Fig. 2). Based on this study, vitrectomy for DME associated with VMT appears beneficial; however, the study has several shortcomings. There was no control group, and VMT was defined by clinical judgment rather than a standardized definition. Finally, surgical interventions were not standardized [36]. Taken together, these results suggest that vitrectomy with posterior hyaloid elevation and removal can be beneficial in the setting of DME with VMT.
Attached Vitreous and No Observable Traction
In addition to cases with a taut posterior hyaloid or VMT, there is also support in the literature for vitrectomy in some patients with an attached hyaloid but no observable traction. Ikeda et al. described three DME eyes without clinical evidence of traction in the pre-OCT era that underwent vitrectomy. The cystoid changes had disappeared by 5 days postoperatively in all eyes. The diffuse macular edema had resolved within 2 weeks and VA was maintained or improved [37]. Otani et al. subsequently evaluated 13 DME eyes with retinal swelling on OCT before and after vitrectomy. At 6 months postoperatively, the mean foveal thickness decreased significantly from 630 to 350 μm. The best corrected visual acuity (BCVA) improved by more than two lines in 38 % of eyes and remained the same in 54 % [38]. Additionally, La Heij et al. found resolution of macular edema in all patients after a median period of 3 months and improvement of VA (median improvement of five lines) after vitrectomy in 21 eyes with DME with an attached hyaloid but no known traction [39]. Taken together, these three studies suggest that DME may improve in patients with an attached hyaloid, even without known traction.
There are several reasons that these patients with an attached hyaloid without traction may benefit from vitrectomy. First, there may be subclinical traction from the posterior hyaloid that may not be detected clinically, as Ikeda et al. and La Heij did not use OCT. In addition, vitrectomy may serve to increase vitreous oxygen tension. Regardless, these results on DME are less impressive than those involving patients with known traction but may be considered, especially in refractory cases.
Detached Vitreous (PVD)
Based on the above studies, traction appears to be a significant cause of diffuse retinal leakage in DME that can improve with vitrectomy. If the hyaloid is detached and there is no other known etiology to cause traction, such as an epiretinal membrane (ERM), there is less of a rationale for surgery. Several studies have examined this.
Ikeda et al. in 2000 described five DME eyes that had a detached hyaloid without ERM on exam and confirmed intraoperatively. After vitrectomy, four eyes had resolution of the DME and all had improved VA. This was attributed to removal of cytokines and an increase in vitreous oxygen tension after surgery [37]. Other studies have failed to replicate these findings. Massin et al. evaluated eight eyes with diffuse DME and detached hyaloid without ERM before and after vitrectomy using OCT. While retinal thickness decreased from 522 to 428 μm after surgery, median VA actually worsened from 20/100 to 20/200 [5]. These studies suggest that vitrectomy is generally not indicated for mild DME in patients with a detached hyaloid without traction.
Detached Vitreous (PVD) with Epiretinal Membrane
Like a taut posterior hyaloid or VMT, an epiretinal membrane (ERM) can also exert traction on the retina and contribute to DME. ERM peeling has been suggested as an adjunct to vitrectomy in select cases of DME where an ERM is present. Although the majority of studies examining surgical intervention for DME have focused on cases of traction from the posterior hyaloid, some groups have looked at the utility of ERM peeling in eyes with DME. For instance, a subgroup of DME patients examined by Yamamoto et al. had a PVD and ERM before undergoing vitrectomy and membrane peel. This subgroup of five patients had significant improvement in postoperative mean VA, and the final VA improved by two or more lines in 60 % of eyes. Although mean foveal thickness decreased from 448 to 238 μm, this difference was not statistically significant [40]. Figure 3 offers an illustrative example of diabetic patient with a PVD and ERM on OCT; foveal contour improved following vitrectomy and membrane peel. Vitrectomy with membrane peel could be considered in select cases. The same criteria should be used as for nondiabetic ERM. Surgery should be considered for symptomatic visual loss associated with obvious ERM.
Fig. 3
Improvement of macular edema and foveal contour after vitrectomy and ERM peel in diabetic patient with ERM and PVD. (a) Preoperative fundus photograph. (b) Preoperative fluorescein angiogram. (c) Preoperative OCT showing PVD and ERM. (d) Postoperative fundus photograph. (e) Postoperative fluorescein angiogram. (f) OCT showing resolution of vitreofoveal traction and DME after vitrectomy without ILM peeling
The Role of Internal Limiting Membrane Peeling During Diabetic Vitrectomy
Rationale for Surgical Intervention
As an adjunct to vitrectomy, peeling the ILM has been recommended in select cases of DME. The ILM may thicken in DME due to cellular proliferation and deposition of extracellular matrix. This leads to decreased water movement between the vitreous and retina, build-up of proteins in the interstitial space, decreased diffusion of proteins to the vitreous space, and macular edema [41, 42]. Removal of this thickened ILM eliminates a possible barrier to cytokines and oxygen [43, 44]. It can also help to ensure complete removal of residual cortical vitreous [41, 45]. Similar to a taut posterior hyaloid or VMT, tangential traction can also be exerted by an ILM. ILM peeling can also ensure complete removal of epiretinal cells. This may limit postoperative ERM formation by removing the scaffold for proliferating cells [46]. For these reasons, ILM peeling has been proposed as an adjunct to vitrectomy in select cases of DME.
In order to better understand how ILM peeling is beneficial, several studies have thoroughly investigated changes in pathology and imaging. Gentile et al. described two cases of diffuse DME after vitrectomy that demonstrated a taut ILM. After repeat vitrectomy with ILM peeling, macular edema and VA improved. The ILM was analyzed with immunostaining and revealed an inner monolayer of cytokeratin-positive (retinal pigment epithelial (RPE) cells) and/or glial fibrillary acidic protein-positive cells with smooth muscle actin (SMA) immunoreactivity. As SMA indicates myofibroblast differentiation and the contractile ability of the RPE and glial cells, these changes likely caused tangential traction which was relieved by ILM peeling [47]. The tangential traction that can be exerted by the ILM was also imaged in a study by Abe et al. They performed a retrospective case series of 26 DME eyes imaged with OCT to identify both traction seen on tomography and fine folds seen on three dimensional imaging. After ILM peeling, the fine folds resolved, even in those eyes without traction on tomography. Surgically obtained specimens confirmed that the fine folds involved the ILM [48]. This suggests that ILM peeling can help resolve tangential traction in DME, even when not obvious on standard tomography.
Evidence for Surgical Intervention
Peeling of the ILM has been proposed as a helpful adjuvant to vitrectomy for DME but results in the literature are mixed. Kamura et al. evaluated 34 DME eyes treated with ILM peeling during vitrectomy compared to eyes treated with vitrectomy alone and found that VA improved significantly after vitrectomy regardless of ILM peeling and without a significant difference between the groups [49]. Bahadir et al. examined 17 DME eyes that underwent ILM peeling during vitrectomy, and comparing them to eyes with vitrectomy alone found a significant improvement in postoperative VA in both groups, but no difference between them [50]. Rosenblatt et al. reviewed 26 eyes with refractory DME without traction that were treated with vitrectomy and ILM peel. There was a statistically significant improvement of mean VA (50 % of eyes gained at least two lines of VA) and mean foveal thickness (311 μm from 575 μm) [51]. Patel et al. evaluated ten eyes with diffuse refractory DME which underwent vitrectomy and ILM peeling compared to vitrectomy alone, finding that ILM peeling was associated with a significant improvement in foveal thickness and macular volume, but not with change in VA [52]. Additionally, Recchia et al. examined ten patients after vitrectomy and ILM removal with diffuse DME refractory to laser, finding both improvement in central macular thickness and VA [53]. Finally, Yanyali et al. treated 12 DME eyes with vitrectomy and ILM peel compared to controls treated with laser in this prospective study, finding a significant improvement in mean foveal thickness and VA in the surgical group but not in the laser group [54]. In a later study, Yanyali et al. reviewed 27 DME eyes that underwent vitrectomy with ILM peeling, finding a significant decrease in foveal thickness and improvement in VA [55]. In summary, the majority of these studies report some additional benefit with ILM peeling; however, restoration of foveal anatomy was more common than improvement in VA. In practice, employment of ILM peeling for diffuse DME appears mixed, as in the DRCRnet Vitrectomy Study that showed 54 % of surgeons elected to peel the ILM [36].
Prognostic Factors
Several prognostic factors for favorable outcomes after surgical intervention for DME have been identified, perhaps most importantly preoperative VA and early surgical intervention. Pendergast et al. showed a strong correlation between preoperative and postoperative VA. They examined 55 DME eyes that underwent vitrectomy with stripping of a taut posterior hyaloid and found that eyes with preoperative BCVA of 20/200 or worse responded less favorable to vitrectomy. Eyes with preoperative BCVA of 20/100 or better improved by a median of 60 % compared to 18 % the eyes with VA of 20/200 or worse [7]. Harbour et al. examined ten DME patients who underwent vitrectomy for a taut posterior hyaloid and found that the three eyes with rapid deterioration of vision from DME followed by prompt surgical intervention (less than 1 month) experienced the most improvement in final BCVA [30].
Other studies have used OCT to delineate prognostic factors for DME and surgical intervention by identifying markers for photoreceptor damage that would limit visual potential. Maheshwary et al. found a statistically significant correlation between percentage disruption of the IS/OS junction and VA in 62 DME eyes using OCT [56]. Additionally, Chhablani et al. found that external limiting membrane (ELM) integrity correlated with postoperative outcome in their study of 34 eyes with resistant DME treated with vitrectomy [57]. Finally, Nishijima et al. identified hyperreflective foci in the outer retina that were predictive of photoreceptor damage and poor vision in their study of 32 DME eyes that underwent vitrectomy [58].
Additionally, other ocular and systemic prognostic factors have been identified. Longer axial length was found to be associated with better VA after vitrectomy by Wakabayashi et al. in 51 eyes with DME that underwent vitrectomy [59]. Better glycemic control also correlated with better outcomes. Yamada et al. examined 44 diabetic eyes that underwent vitrectomy with ILM peeling for DME and found that the postoperative macular thickness was significantly thicker with higher glycosylated hemoglobin levels [60]. These studies suggest that there are retinal, ocular, and systemic factors that can help identify patients who could benefit from surgical intervention for DME.
Summary
When considered by categories of vitreoretinal interface problems, the utility of vitrectomy in select DME cases becomes clearer. Vitrectomy has been shown to be beneficial in most DME cases where a taut posterior hyaloid or vitreomacular traction is present. It is beneficial in select cases where the posterior hyaloid is attached, even if there is no observable traction. When separation of the posterior hyaloid has occurred, vitrectomy can be beneficial in select cases where an ERM is present.
Favorable anatomic results are more common than visual results when vitrectomy and other surgical interventions are performed in select cases of DME. As discussed, VA can improve 5–15 letters postoperatively but may worsen in some cases. Despite this limited improvement in VA, OCT results are often more impressive. Foveal thickness usually decreases postoperatively by 100–250 μm on OCT or greater than 50 % reduction of retinal thickening. The fact that improvement on OCT does not translate to significant visual results may reflect that vitrectomy is often performed for refractory DME cases with long-standing edema with irreversible macular damage.
In summary, eyes with observable vitreous and/or epiretinal traction are most likely to improve after vitrectomy. Eyes with refractory edema and no observable traction, however, are less likely to improve. Unfortunately, improvement in retinal thickening is often more impressive than improvement in VA even in these select cases. However, vitrectomy and other surgical interventions may be beneficial for select cases of DME, especially when surgical intervention is undertaken early, before photoreceptor damage has occurred.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
