Introduction

Diabetes mellitus is a main cause of sight impairment in people of working age. Ninety-three millions people around the world are estimated to have diabetic retinopathy (DR) [1]. Diabetic macular Oedema (DMO) affects 7% of diabetic patients and it is the main cause for not only decreased vision, but also loss of sight associated with DR. Anti-vascular endothelial growth factor (VEGF) therapy has become the first line treatment for centre-involved DMO and it is effective in improving and maintaining visual acuity, and this was shown in large-scale randomised controlled trials [2,3,4,5].

Patients suffering from DMO undergoing treatment with anti-VEGF require pars plana vitrectomy (PPV) for complications such as non-clearing vitreous haemorrhage or epiretinal membrane. Frequently, these patients need ongoing intravitreal treatment for DMO even after PPV.

The vitreous gel consists of collagen and glycosaminoglycans, and is believed to act as molecular barrier to drug diffusion [6]. The fine structure of the vitreous, the flow systems operating within it, its age-related structural changes, and the presence of inflammation may have a potential effect on the movement of particles targeting the retina [7]. Drug elimination from the vitreous occurs by the aqueous outflow into the anterior chamber and permeation through the retina via retinal-choroid-sclera pathways [8]. It has been shown that after vitreous removal, the clearance of VEGF is increased. This suggests that pharmokinetics of Anti-VEGF drugs and steroids could also be affected and therefore it may be more useful to predict the final outcome of their interactions except in terms of clinical outcomes as has been done in the present study [9].

Dexamethasone (DEX) implant is a long-lasting effect system made by a bioerodible copolymer. The MEAD trial [10] revealed that DEX implant might have an anatomical and functional effect for up to 6 months in non-vitrectomized eyes. The use of DEX implant in vitrectomized eyes has not been widely studied, and only a subgroup analysis of the CHAMPLAIN pivotal trial [11] and few scarce routine clinical care reports have been reports [12, 13].

We aimed to compare visual and anatomical outcome in vitrectomized and non-vitrectomized eyes treated with DEX implant due to DMO with a long follow-up in a large cohort.

Methods

Multicenter, retrospective, interventional study. Participants: 236 eyes from 234 patients with DMO with or without previous vitrectomy performed with a follow up of at least 12 months involving multiple sites from (1) Private Retina Service, Buenos Aires, Argentina; (2) Tel Aviv Medical Center, Israel; (3) Hospital Clínic of Barcelona, Spain; (4) University of Udine, Italy; (5) University of Perugia, S.Maria della Misericordia Hospital, Perugia, Italy; (6) The Macula Foundation, Genova, Italy; (7) University of Sydney, Sydney, Australia; (8) Istanbul University Istanbul Faculty Of Medicine, Turkey; (9) Susrut Eye Foundation & Research Centre, India; (10) Azienda Ospedaliera Universitaria Sassari, Italy; (11) La Fe University Hospital of Valencia, Spain

Ethics statement Institutional review board (IRB) approval was obtained through the individual IRBs at the participating institutes for a retrospective consecutive chart review. The research adhered to the tenets of the Declaration of Helsinki. All data discussed in this study were fully anonymized before they were accessed. Informed consent was taken. Patient records from January 1, 2014 to December 1, 2020 were reviewed for cases of DMO treated with DEX implant in vitrectomized and not vitrectomized eyes.

Study participants: The following were set as study inclusion criteria: (1) age 18 years or older; (2) type 1 or 2 diabetes mellitus; (3) DMO causing visual loss (BCVA ≤ 20/32; 0.2 logMAR) in vitrectomized or not vitrectomized eyes; foveal-involving macular oedema defined by central macular thickness (CMT) of >300 μm in the central subfield; and intra- or subretinal fluid seen on SD-OCT; (4) treatment with DEX implant; (5) at least 12 months of follow-up after first DEX implant. Exclusion criteria were (1) other concomitant ocular diseases causing macular oedema (i.e. neovascular age-related macular degeneration or choroidal neovascularization due to other reasons, retinal vein occlusion, uveitis and recent intraocular surgery possibly causing), (2) uncontrolled glaucoma.

Consecutive patient charts were reviewed for demographic data; previous treatments given for DMO; stage of retinopathy (diagnosed by clinical examination, as per the International Classification) [14]; BCVA in logarithm of the minimum angle of resolution (logMAR) scale and CMT before and at last follow up after DEX implant; IOP before DEX implant and use of IOP-lowering treatment during study period; lens status and cataract grading before and at last follow up after DEX implant; number of DEX implants given trough the study period; additional treatments given for DMO during study period.

Refractory DME was defined as worsening of BCVA by 2 Early Treatment Diabetic Retinopathy Study (ETDRS) lines or reduction of less than 10% of retinal thickness on SD-OCT measured 1 month after at least 3 anti-VEGF injections that were given at monthly intervals.

All included subjects were required to have OCT scans obtained using horizontal raster pattern scans cantered on the fovea, using spectral domain-OCT (Spectralis; Heidelberg Engineering, Heidelberg, Germany, and HD-OCT Cirrus 5000, Carl Zeiss Meditec, Dublin, CA, USA). Retinal thickness was analysed using the retinal thickness map analysis protocol with nine ETDRS subfields. Central foveal subfield thickness (CST) was defined as average retinal thickness of the circular area with 1 mm diameter around the foveal centre, recorded at baseline and at last follow-up visit after DEX implant.

IOP increase was considered significant in case of (1) eyes with pre-existing glaucoma and increase in number of IOP-lowering drops or surgery during follow-up, and (2) every eye without pre-existing glaucoma that needed any IOP-lowering therapy during follow-up.

Outcome measures

Main outcome measures were BCVA and CST change before and at last follow up after DEX implant compared between vitrectomized and non-vitrectomized eyes. Secondary outcomes were number of DEX implants, additional treatments needed, the proportion of cataract surgery, and the use of IOP-lowering treatment between the two study groups during the study period.

Statistical analysis

The demographics and clinical characteristics of our study cohort were evaluated using traditional descriptive methods. To control for inter-eye correlation, we used a generalised estimating equations (GEE) procedure. Differences in baseline characteristics between the groups as well as differences in BCVA and CST from baseline to last follow-up were analysed by an univariable GEE model. Differences in outcome variables (BCVA change, CST change) between the groups were analysed by a multivariable regression model, including baseline BCVA or baseline CST respectively. All statistics were computed with SPSS Statistics 25 (IBM, Armonk, NY).

Results

The study included 236 eyes from 234 patients, with mean age of 69.5 ± 13.8 years, 120 male patients (51.3%). The non-vitrectomized group included 130 eyes, 71 (54.6%) were treatment naive, and 59 (54.4%) were previously treated for DMO. The vitrectomized group included 106 eyes, 76 (71.7%) were treatment naive, and 30 (28.3%) were previously treated for DMO. Causes for PPV were vitreous haemorrhage (95.3%, n = 225), and vitreomacular interface abnormalities in (4.7%, vitreomacular traction n = 11, epiretinal membrane n = 1). Baseline characteristics are detailed in Table 1. There was a difference between the two groups in gender (p = 0.540). However, we believe that that would not have changed the results materially.

Table 1 Baseline characteristics.
Full size table

In the non-vitrectomized group, mean baseline BCVA was 0.59 ± 0.15, and 0.61 ± 0.19 for naive and refractory eyes, respectively. In the vitrectomized group, mean baseline BCVA was 0.57 ± 0.17 logMAR, and 0.58 ± 0.40 logMAR for naive and refractory eyes, respectively (p = 0.810). In the non-vitrectomized group, mean baseline CST was 583 ± 94 µm, and 565 ± 107 µm for naive and refractory eyes, respectively. In the vitrectomized group, mean baseline CST was 693 ± 159 µm, and 490 ± 133 µm for naive and refractory eyes, respectively. CST was significantly higher in the vitrectomized eyes and it did significantly differ between the groups (p < 0.001). The proportion of pseudophakic eyes and glaucomatous eyes was significantly higher in the vitrectomized eyes and it did significantly differ between the subgroups (p = 0.009 and p = 0.003, respectively).

Functional and anatomical change after DEX implant

Mean baseline BCVA was 0.59 ± 0.15 logMAR, 0.61 ± 0.19 logMAR, 0.57 ± 0.17 logMAR, and 0.58 ± 0.40 logMAR and changed by −0.23 ± 0.20, −0.15 ± 0.25, −0.21 ± 0.15, and −0.03 ± 0.29 for naive and refractory non-vitrectomized and naive and refractory vitrectomized eyes, respectively (p < 0.001 for naive and refractory non-vitrectomized and naive-vitrectomized eyes; p = 0.533 for refractory vitrectomized eyes, Fig. 1A). There was no significant difference in BCVA change between vitrectomized and non-vitrectomized eyes in both groups (naive eyes: p = 0.89, refractory eyes: p = 0.10). There was a significant difference in BCVA change between naive and refractory DMO in the vitrectomized group (p = 0.001). Notably, refractory vitrectomized eyes did not gain vision after DEX-implant treatment. However, an anatomical response was noted with reduction of retinal thickness (Table 2).

Fig. 1: Funtional and anatomical changes after DEX implant.
figure 1

Change in visual acuity (VA, A) and central subfield thickness (CST, B) from baseline to last follow-up. Data are mean ± SD.

Full size image
Table 2 Treatment characteristics, outcome and safety profile.
Full size table

Mean baseline CST was 583 ± 94 µm, 565 ± 107 µm, 693 ± 159 µm, and 490 ± 133 µm and changed by −304 ± 117 µm, −252 ± 180 µm, −437 ± 181 µm, and −173 ± 170 µm for naive and refractory non-vitrectomized and naive and refractory vitrectomized eyes, respectively (for all groups: p < 0.001, Fig. 1B). There was no significant difference in anatomical outcome between vitrectomized and non-vitrectomized eyes in both groups (naive eyes: p = 0.21, refractory eyes: p = 0.65,). Moreover, there were no significant differences in anatomical outcome between naive and refractory DMO in vitrectomized eyes (p = 0.059). (Table 2)

Safety profile

Sixteen out of 24 phakic vitrectomized eyes (66.7%) underwent cataract surgery during follow up (Fig. 2A): Within the naive eyes group 11/12 (91.7%), and within the refractory eyes group 5/12 (41.7%). In the non-vitrectomized group, 22 out of 38 eyes (57.9%) underwent cataract surgery during follow up (Fig. 2A): Within the naive eyes group 15/16 (93.8%), and within the refractory eyes group 7/22 (31.8%).

Fig. 2: Treatment characteristics, outcome and safety profile.
figure 2

Eyes that underwent cataract surgery (A) and eyes that received IOP-lowering treatment (B) within the follow-up in %.

Full size image

IOP increased significantly in vitrectomized eyes during follow-up in 14/106 eyes (13.2%), out of which 6/14 eyes (42.9) had controlled pre-existing glaucoma. IOP increase was treated with topical treatment in 11/14 (78.6%), with surgery in 2/14 (14.3%) and without need for treatment in 1 case (7.1%). In non-vitrectomized eyes, 18/127 (14.2%) received IOP-lowering treatment during follow-up (Fig. 2B).

The mean number of DEX implants given during follow-up was 3.5 (3.9 ± 0.5, 3.1 ± 1.2, 3.1 ± 1.1, 3.9 ± 2.9, for naive and refractory non-vitrectomized and naive and refractory vitrectomized eyes, respectively, Fig. 3A, B). There was no significant difference between vitrectomized and non-vitrectomized eyes in both groups (p = 0.81). The proportion of patients who needed additional anti-VEGF injections was 1 (1.4%), 10 (16.9%), 2 (2.6%), 6 (20%) for naive and refractory non-vitrectomized (p = 0.002) and naive and refractory vitrectomized eyes, respectively (p = 0.014).

Fig. 3: Ultra widefield fundus imaging obtained few minutes after DEX implant procedure, injection was performed in the superotemporal quadrant.
figure 3

A Left eye, non-vitrectomized: The implant is maintained in the superotemporal quadrant. B Right eye of another patient, vitrectomized: The implant remained in the inferior part of the eye.

Full size image

Discussion

To the best of our knowledge this is the largest study that directed to compare the functional and anatomical results with the DEX implant in vitrectomized and non-vitrectomized eyes in a multicentric cohort of DMO eyes. We showed that DEX implant caused significant functional and anatomical improvement in both groups when compared to the baseline measurements. The two groups fared the same, i.e. there was no statistical significant difference between them. Furthermore, we measured the safety profile in this multicentre international study.

Following PPV surgery, frequently repeated application of intravitreal medication may be required [15]. Previous studies have shown a shorter half-life of bevacizumab, ranibizumab and aflibercept in eyes which have undergone PPV [16, 17]. This has led to an expert opinion that fixed monthly dosing, with low threshold for increasing frequency of injection even to 2-weekly may be required, as well as close monitoring of patients to establish individual response and customise injection frequency as needed [15]. This remains controversial, as animal models have shown that intraocular pharmacokinetic properties of anti-VEGF agents in vitrectomized eyes were similar to those in non-vitrectomized eyes [18]. Therefore, it is important to know how these drugs will act on vitrectomized eyes and compare with non-vitrectomized eyes. After the vitreous is removed, the eye becomes less viscous and loses its natural role of reservoir. Therefore, a clearance of intravitreal drugs from the vitreous cavity is accelerated [9, 19].

This may be one of the reasons why in vitrectomized eyes, DEX implant has the potential to be more predictive in terms of functional and anatomical outcomes. It has been shown that DEX implant has the potential to work properly by itself, without needing in this case the vitreous [20]. DEX implant is a sustained-release preparation of DEX embedded in a bioerodible copolymer consisting of poly (lactic-co-glycolic acid). Pharmacokinetic and pharmacodynamic data suggest that when injected into the posterior segment, the DEX implant releases the active component into the vitreoretinal tissues for up to 6 months [21]. In an experimental study, the behaviour of DEX implants injected into vitrectomized eyes was modelled, and DEX implant was injected into a BSS-filled box [20]. These preclinical data are supported by clinical studies. The CHAMPLAIN study was conducted on 55 vitrectomized eyes and evaluated the safety and efficacy of DEX implant over a 26-week period [11]. The findings showed a decrease in macular thickness and an increase in BCVA after 8 weeks, and these effects continued for 26 weeks. Moreover, the combination of intravitreal DEX injection with PPV had been shown to be safe and effective for some underlying conditions that result in macular oedema, like DR, retinal vein occlusion, and uveitis [22]. The findings of our current study are in concordance with previous reports about the efficacy and safety of DEX implant in vitrectomized eyes [11,12,13].

Our study had certain limitations, such as its retrospective nature, however, we report comparative analysis with long-term effectiveness and safety of DEX implant in vitrectomized and non-vitrectomized eyes. Due to lack of standardised treatment or re-treatment protocol, we cannot make conclusive remarks on its efficacy and suggest treatment protocols in vitrectomized eyes using DEX implant. Moreover, CST was statistically significantly higher in the vitrectomized eyes and it did significantly differ between the groups. This might interfere with the anatomical outcome. As this was an international multicentre retrospective study, selectivity of patients may have played a part in the results, as patients with a shorter follow-up were excluded from the study.

In conclusion, this study has demonstrated similar effectiveness of DEX in non-vitrectomized and vitrectomized eyes.

Although there are postulated reasons why anti-VEGF agents may be less effective in vitrectomised eyes the clinical data is not conclusive

We strongly believe that in suitable vitrectomized patients, DEX implant might be considered as treatment option due to its efficacy and safety profile. A randomised clinical trial is needed in order to evaluate the impact of vitrectomy in the current algorithm of DMO treatment.

Summary table

What was known before

  • There was no certainty that Dex implant works the same in vitrectomized and non-vitrectomized eyes.

What this study adds

  • We demonstrated similar anatomical and functional efficacy of DEX implant in non-vitrectomized and vitrectomized eyes. Its efficacy was not influenced by full vitrectomy for Diabetic retinopathy complications. Safety profile was well balanced between groups.