Steroid Drugs as an Adjunct for Reducing the Incidence of Proliferative Vitreoretinopathy after Rhegmatogenous Retinal Detachment Surgery: A Meta-Analysis of Randomized Controlled Studies

Abstract

Introduction

The efficacy and influence of steroids for reducing the incidence of proliferative vitreoretinopathy (PVR) after rhegmatogenous retinal detachment (RRD) surgery remain controversial. Systematic review and meta-analysis were conducted to explore the effect of steroids versus placebo on risk of PVR.

Methods

We searched PubMed, Embase, Web of science, EBSCO, and Cochrane library databases through September 2020 for randomized controlled trials (RCTs), assessing the effect of steroid drugs as an adjunct for reducing the incidence of PVR after RRD surgery. This meta-analysis was performed using the random-effect model. Data were extracted by two reviewers independently; the quality of RCTs was assessed by the Cochrane risk-of-bias tool. We calculated risk ratio (RR) and the 95% confidence intervals (CIs) of all outcomes and plotted on forest plots. I² accessed using the χ² test was applied to quantify the degree of heterogeneity.

Results

Four RCTs involving 478 patients (478 eyes) are included in the meta-analysis. There was no significant difference in the incidence of PVR recurrence between steroid groups and control groups (RR: 0.87, 95% CI: 0.70–1.08, p = 0.19). However, the incidence of recurrent PVR was lower in the steroid group (RR: 0.67, 95% CI: 0.46–0.99, p = 0.04) than in the control group when only PVR grades A and B were taken into consideration. Besides, steroids could significantly reduce the incidence of macular edema after surgery (RR: 0.64, 95% CI: 0.47–0.88, p = 0.007). The steroid group and control group had comparable outcomes of retinal reattachment rate and reoperation rate after primary surgery. Additionally, there was no significant difference of the incidence of epiretinal membrane, and the incidence of surgery required by epiretinal membrane.

Conclusion

This meta-analysis reveals that RRD surgery combined with steroid drugs administration could significantly reduce the recurrence in PVR grade A and B subgroup, as well as the incidence of macular edema after surgery.

Introduction

Rhegmatogenous retinal detachment (RRD), the most common type of retinal detachment (RD), is a sight-threatening, serious ocular disorder, which can cause blindness unless surgical management is received [1]. Although pars plana vitrectomy (PPV) is one of the major treatments for RRD, there will be increased risks for proliferative vitreoretinopathy (PVR) after the surgery. PVR occurs in about 10% of RRD patients and can lead to the reoccurrence of RD [2]. Further surgery is needed for PVR patients. However, with the increase in the number of operations, the postoperative anatomic success rate after surgery is about 45–80%, and rates are lower than 60% for severe PVR stages [3].

With deep investigation on the pathogenesis and mechanism of PVR, several drugs such as adjuvant therapies during the perioperative period have been reported to show potential benefit and reduce postoperative risk of PVR. Administration of 5-fluorouracil and low-molecular-weight heparin can reduce the re-detachment rate and rate of reoperations to achieve stable reattachment after 6 months of high-risk eyes for PVR [4], though others found no significant difference between the group with 5-fluorouracil and low-molecular-weight heparin and with placebo [4–7]. Use of oral 13-cis-retinoic acid decreases the rate of macular pucker formation [8], and a recent study has shown that oral 13-cis-retinoic acid significantly improves the single surgery success rate and visual acuity [9].

Recently, steroids, such as dexamethasone, prednisone, and triamcinolone acetonide (TA), are considered beneficial for reducing the risk of PVR [10–14]. Steroids are mainly used to reduce intraocular inflammation [15]. Considering that inflammatory factors play a vital role in the pathogenesis of PVR [16], steroids are also used in the perioperative treatment of RRD patients to reduce the risk of PVR. The aim of this systematic review and meta-analysis of randomized controlled trials (RCTs) was to assess the impact and effect of steroids and compare different interventions of steroid on the outcomes of RRD after PPV.

Methods

This systematic and meta-analysis review was based on the Preferred Reporting Items for Systematic Reviews and Meta-Analysis and the Cochrane Intervention Systematic Review Manual [17, 18]. Ethical approval and patient consent were not required since all analyses were based on previously published research.

Literature Search and Selection Criteria

To find the observational studies on the effects of the steroid drugs as an adjunct for reducing the incidence of PVR after RRD surgery, we systematically searched individual databases including PubMed, Embase, Web of science, EBSCO, and Cochrane library from inception to September 2020 for RCTs. We also reviewed reference lists of the identified publications for additional pertinent studies. The keywords for searching include “proliferative vitreoretinopathy” OR “rhegmatogenous retinal detachment” OR “steroids” OR “dexamethasone” OR “prednisone” OR “triamcinolone acetonide.” At the same time, we also manually searched the reference lists of retrieved research and related reviews and repeated the above search process to ensure that all the studies that meet the requirements were included.

Given different surgery strategies for PVR and various steroids as adjuvant drugs to surgery, the inclusion criteria were proposed to fully evaluate the therapeutic effect of steroids for PVR. The classifications of PVR grades A, B, and C were according to the acknowledged diagnostic criteria [19]. The inclusion criteria were presented as follows: (1) study design is RCTs; (2) patients who undergo PPV, PPV and silicone oil removal, or scleral buckling; (3) intervention treatments are steroids as adjunct to surgery; and (4) measurement of at least one of the outcomes that included in the present study. The exclusion criteria were presented as follows: (1) non-RCT study and (2) different evaluation outcomes. If duplicate publications are retrieved from literature search, only the most complete one would be included for pooled meta-analyses. All contents, including full texts, abstracts, and titles, were independently screened by two authors (Manhong Xu and Xiaoe Fan). If there was a discrepancy in the results, an additional author (Xinyuan Huang) would be added to draw the conclusion through consultation.

Data Extraction and Outcome Measures

The relevant data from the articles were extracted by two reviewers (Manhong Xu and Xiaoe Fan) independently, using a standard data extraction form. The extracted data included the first author(s) or the information provider, publishing date, study design, sample size, geographical location of the research, interventions details, age, sex, outcomes, and follow-up periods. Postoperative PVR recurrence rates were extracted as the primary outcome of the meta-analysis. Secondary outcomes included retinal reattachment (complete retinal reattachment) rate, posterior retinal reattachment (post equatorial retinal reattachment/incomplete reattachment) rate, incidence of macular edema, reoperation rate after primary surgery, macular pucker or epiretinal membrane (ERM), and surgery rate of macular pucker or ERM.

Quality Assessment in Individual Studies

The methodological quality of each RCT was assessed by the Cochrane risk-of-bias tool (RoB 2.0) which consists of three evaluation elements: randomization, allocation concealment, blinding (both participants and outcome assessment), dropouts, selective reporting, and other bias. Risk of bias for different aspects was assessed as high risk, low risk, and few issues. The results of the risk of bias assessment were summarized using the risk of bias plot and the risk of bias summary plot. Quality assessment of eligible studies was performed by two authors separately. The authors would write to the corresponding authors of eligible studies for additional information if we fail to extract all necessary data from included studies.

Statistical Analysis

RevMan 5.5.3 (Cochrane Collaboration, Nordic Cochrane Centre, Copenhagen, Denmark) was applied to integration collected data statistics and analysis. The mean difference and risk ratio (RR) were used to assess continuous variable outcomes and dichotomous outcomes with a 95% confidence interval (CI), respectively. The heterogeneity of studies was accessed using the χ² test based on the values of P and I². The random-effects model was applied for the meta-analysis. I² was calculated to quantify the degree of heterogeneity across studies. I² < 25%, 25–50%, and >50% indicate low, moderate, or high heterogeneity, respectively. p values <0.05 were considered statistically significant.

Results

Literature Search

The detailed flowchart of the search and selection results is shown in Figure 1. A total of 130 potential records up to September 2020 were identified initially (PubMed = 130, and no more articles are included in other databases, including Embase, Web of science, EBSCO, and Cochrane library). After removing 13 duplicates, a total of 117 potential records were assessed by certain criteria. Finally, 4 RCTs out of 130 studies were included in this meta-analysis (Fig. 1) [20–23].

Fig. 1.

Flow diagram of the literature search and selection. PVR, proliferative vitreoretinopathy; RCT, randomized controlled trial.

Study Characteristics

The baseline characteristics of the included four RCTs are shown in Table 1. These studies were published between 2008 and 2018, and two of the four trails were finished in Iran, one was in Switzerland and the other was in the UK. The sample sizes ranged from 52 to 220, and totally, 487 participants with 487 eyes were included in these trails, comprising 164 (33.7%) females and 323 (66.3%) males. Besides, there were 243 patients in the steroid arm and 244 patients in the control arm. PVR grades A, B, and C are included. In 2008, patients received PPV, and 4 mg TA was injected into the silicone-filled vitreous cavity at the end of the procedure [23]. In 2010 and 2012, scleral buckling was performed in all patients [21, 22]. The patients in the treatment group in the study performed by Mohammad H. Dehghan et al. [22] received 1 mg/kg oral prednisolone for 10 days postoperatively, while the patients in the steroid group in the study of Koerner et al. [21] received prednisone for 15 days starting with 100 mg at the day of surgery and being tapered to 12.5 mg. And in the study of Banerjee et al. [20], the adjunct group received an injection of 0.7 mg of slow-release dexamethasone (Ozurdex) at the time of vitrectomy surgery and silicone oil removal. Although there were differences among the follow-up durations of four studies, the primary outcome and secondary outcomes in this study were observed at 6 months. Notably, there were no results with the Ozurdex injection at the time of silicone oil removal within 6 months [20].

Table 1.

Characteristics of included studies

Study	Place	Disease	Intervention details		Age (mean±SD), years		Gender, female, n (%)		Surgery	Intervention		Follow-up duration, months
			steroid arm	control arm	steroid arm	control arm	steroid arm	control arm		steroid arm	control arm
Banerjee et al. 2017 [20]	UK	RRD (PVR grade C)	70	70	60.6±14.3	61.6±13.9	24 (34.3)	30 (43)	PPV and silicone oil removal	0.7 mg slow-release dexamethasone (Ozurdex) implant	No adjunctive medication	6, 24
Koerner et al. 2012 [21]	Switzerland	RRD (PVR grades A and B)	110	110	54.5±13.8	54.1±14.4	42 (38.2)	34 (30.9)	Scleral buckling	100 mg–12.5 mg oral prednisone for 15 days before surgery	Placebo	1, 3, and 6
Dehghan et al. 2010 [22]	Iran	RRD (PVR grades A and B)	25	27	48±14	42±17	8 (32%)	10 (37%)	Scleral buckling	1 mg/kg oral prednisolone plus antacid for 10 days	Vitamin BS6 tablets	3 and 6
Ahmadieh et al. 2008 [22]	Iran	RRD (PVR grade C)	38	37	54.5±18.8	45.7±20.7	8 (21.1%)	8 (21.6%)	PPV	4 mg TA was injected into the silicone-filled vitreous cavity at the end of the procedure	No adjunctive medication	1, 3, and 6

RRD, rhegmatogenous retinal detachment; PVR, proliferative vitreoretinopathy.

Quality Assessment

All of 4 RCTs used a random block permutation method to divide the patients into the steroid arm or control arm. The outcomes were assessed in a blinded fashion. The quality of the eligible studies is shown in Figure 2.

Fig. 2.

Methodological quality summary and risk of bias in the studies. a Authors’ judgments about each risk-of-bias item for each included study. +: low risk of bias; −: high risk of bias; ?: unclear risk of bias. b Risk-of-bias graph: review authors’ judgments about each risk-of-bias item are presented as percentages across all included studies.

Primary Outcomes: PVR Recurrence

The random-effect model is used for the analysis of primary outcomes. As shown in Figure 3 (datasets on top), there is no statistical difference of the recurrence of PVR after surgery between steroid and control groups (RR: 0.87, 95% CI: 0.70–1.08, p = 0.19) with no heterogeneity among the studies (I² = 0%, heterogeneity p = 0.41) when RRD (PVR grade A), RRD (PVR grade B), and RRD (PVR grade C) were included. But when RRD patients with PVR grade C were excluded, the incidence of PVR recurrence was significantly decreased in the steroid group compared with the control group (RR: 0.67, 95% CI: 0.46–0.99, p = 0.04) with no heterogeneity among the studies (I² = 0%, heterogeneity p = 0.57, Fig. 3, datasets in the middle). Consistent with the results before, with the inclusion of only RRD (PVR grade C), there was no significant difference of the PVR recurrence incidence in the steroid group compared with the control group (RR: 1.1, 95% CI: 0.80–1.26, p = 0.96) with no heterogeneity among the studies (I² = 0%, heterogeneity p = 0.92, Fig. 3, datasets at the bottom).

Fig. 3.

Forest plot diagram of postoperative PVR recurrence rate of RRD patients at 6 months. Datasets on the top, in the middle, and at the bottom represent the data of PVR grades A, B, and C; PVR grades A and B; and PVR grade C. M–H, Mantel-Haenszel test; CI, confidence interval; PVR, proliferative vitreoretinopathy.

Secondary Outcomes

Retinal Reattachment

Dehghan et al. [22] did not investigate the retinal reattachment rate in the original publication, so the study was excluded when we analyzed the secondary outcomes/retinal reattachment. In total, 308 eyes (75.3%) achieved retinal reattachment after primary surgery in two groups, with 163 eyes (78.0%) in the steroid group and 145 eyes (72.5%) in the control group. Although the retinal reattachment rate in the steroid arm was higher than in the control arm, there was no significant difference between the two groups (RR: 1.07, 95% CI: 0.99–1.15, p = 0.10) with no heterogeneity among the studies (I² = 0%, heterogeneity p = 1.00, Fig. 4, datasets on top). Dehghan et al. [22] as well as Ahmadieh et al. [23] did not report the results of the posterior (post-equatorial) retinal reattachment rate; therefore, these studies were excluded for the analysis. In the steroid group, the posterior retinal reattachment rate was 85.4%, which was slightly lower than that in the control group (85.9%). But there was no difference between the two groups (RR: 1.00, 95% CI: 0.96–1.04, p = 0.99) with no heterogeneity among the studies (I² = 0%, heterogeneity p = 0.45, Fig. 4, datasets at the bottom).

Fig. 4.

Forest plot diagram of the retinal reattachment rate and posterior retinal reattachment rate of RRD patients (PVR grades A, B, and C) at 6 months. M–H, Mantel-Haenszel test; CI, confidence interval; PVR, proliferative vitreoretinopathy.

Macular Edema

Two out of 4 studies did not report the situation of macular edema, so we only included the other 4 studies [24, 25], which were with no heterogeneity among the studies (I² = 0%, heterogeneity p = 0.99, Fig. 5). Out of 94 patients, 32 patients were manifested as macular edema in the steroid group at 6 months after surgery, and the incidence of macular edema was 34.0%. However, 50 patients in the control group experienced macular edema in a total of 96 control patients, and the incidence was 52.1%. And there was significant difference between the steroid and control groups (RR: 0.64, 95% CI: 0.47–0.88, p = 0.007).

Fig. 5.

Forest plot diagram of the macular edema incidence of RRD patients (PVR grades A, B, and C) at 6 months. M–H, Mantel-Haenszel test; CI, confidence interval; PVR, proliferative vitreoretinopathy.

Reoperation after Primary Surgery

One study [22] was excluded since the researchers did not include the information of reoperation after primary surgery. From the results of other three RCTs, a total 151 (36.9%) eyes experienced reoperation in both groups. For the steroid group, the reoperation rate was 33.5% (70 eyes), and for the control group, the reoperation rate was 40.5% (81 eyes). The reoperation rate in the steroid group did not significantly decrease compared with the control group (RR: 0.84, 95% CI: 0.61–1.16, p = 0.28), but the data showed significant heterogeneity (I² = 32%, p = 0.23, Fig. 6).

Fig. 6.

Forest plot diagram of reoperation after primary surgery within 6 months. M–H, Mantel-Haenszel test; CI, confidence interval; PVR, proliferative vitreoretinopathy.

Macular Pucker or ERM

In the study by Ahmadieh et al. [23], a higher incidence of macular pucker or ERM was shown in the control group (35.1%) compared to the steroid group (25.8%). But in the study by Banerjee et al. [20], almost same incidences of macular pucker or ERM were observed in the two groups (56.0% vs. 59.4%). There was no statistically significant difference of the macular pucker or ERM incidence between the two groups (RR: 0.87, 95% CI: 0.57–1.32, p = 0.50), but the data showed significant heterogeneity (I² = 33%, p = 0.22, Fig. 7). Similarly, patients with no statistically significant difference of macular pucker or ERM required surgery between the two groups (RR: 1.04, 95% CI: 0.73–1.48, p = 0.81), with no heterogeneity among the studies (I² = 0%, heterogeneity p = 0.58, Fig. 8).

Fig. 7.

Forest plot diagram of macular pucker or ERM within 6 months. M–H, Mantel-Haenszel test; CI, confidence interval; PVR, proliferative vitreoretinopathy; ERM, epiretinal membrane.

Fig. 8.

Forest plot diagram of the surgery rate of macular pucker or ERM patients within 6 months. M–H, Mantel-Haenszel test; CI, confidence interval; PVR, proliferative vitreoretinopathy; ERM, epiretinal membrane.

Conclusion

This meta-analysis investigated the effectiveness of steroids as an adjunct to RRD surgery. Four RCTs of small or medium sample size were included, and the quality of most of the evidence in this meta-analysis was moderate. Overall, the four publications included in this review were of good quality.

Surgery is considered to be the best therapeutic approach to RRD, and with the innovation of surgical technology and the update of surgical instruments, the primary retinal reattachment rate has been very high, reaching over 90% [24]. PVR is a common cause of RRD recurrence, causing additional surgeries or clinical interventions. Patients experienced worse vision after multiple surgeries, and the cost of treatment increases significantly. Therefore, prevention of PVR after RRD surgery has always been a clinical problem explored by ophthalmologist.

Increasing evidence indicated that the activation of inflammatory processes, induced by the presence of RRD and surgeries themselves, is particularly associated with the occurrence of PVR and retinal/choroidal detachment [25, 26]. At present, some studies have shown that steroids, as an adjuvant drug of RRD, can effectively reduce inflammation. Intravitreal injection of 40 mg TA after vitrectomy of RRD can reduce intraocular inflammation [27] and improve visual acuity [11]. One pilot study shows that intravitreal injection of crystalline cortisone and removal of most of the vehicle are nontoxic to the intraocular structure, which can reduce postoperative intraocular inflammation [10]. Preoperative subconjunctival injection of 10 mg dexamethasone can reduce the breakdown of blood-retinal barrier in patients undergoing conventional scleral buckling RD surgery [28]. Ozurdex is a 6-mm implant containing 700 mg of dexamethasone in a biodegradable polymer (Novadur, Allergan, Irvine, CA, USA) [29]. Due to the short half-life period of dexamethasone, this sustained-release dexamethasone preparation has been developed and applied for the treatment of RVO [30], noninfectious posterior uveitis [31], diabetic macular edema [32], as well as PVR [24]. In addition to the intravitreal injection, systemic administration of steroids, such as oral corticosteroids before surgery, can improve the retinal reattachment rate of RRD after PPV [33].

Our meta-analysis failed to find the advantage of using steroids as adjuvant drugs for RRD, that is, steroid adjuvant drugs did not reduce the PVR recurrence rate of RRD patients. This result may be due to the heterogeneity of the included study, and our inclusion criteria led to a significant difference between the preoperative risk of the inclusion study and previous systematic examinations. In fact, the treatment group showed a lower albeit not statistically significant PVR rate when compared with the control group. However, when PVR grade C was excluded, patients treated with steroids showed a significantly lower PVR rate. There are two main reasons for this: (1) although steroids were used in several studies included in this review, the medication administration was different, and the sample size was also small. Therefore, although the treatment arm showed a lower PVR recurrence rate, there was no significant difference between the two groups. Larger and more rigorous clinical trials are needed to truly reflect the differences between with or without steroids; (2) the earliest stages of PVR, grades A and B, are characterized by the presence of pigment clumps in the vitreous, caused by pigmented cell proliferation and by the retinal surface wrinkling, respectively [34]. Although the pathological cause of PVR is not yet clear, it is recognized that growth factors and cytokines in the vitreous may be important drivers in the pathogenesis of this blinding disease. When retinal pigment epithelium (RPE) cells are destroyed, retinal vascular endothelial cells and RPE cells are continuously exposed to the environment with a large amount of growth factors [35–38], such as epidermal growth factor (EGF), fibroblast growth factor-2 (FGF-2), transforming growth factor-beta (TGF-b), interferon gamma (IFNg), interleukins, and monocyte chemotactic protein-1 (MCP-1). On the one hand, these factors promote the migration and proliferation of these cells and at the same time, aggravate the formation of extracellular matrix, which is essential to the formation of PVR membrane; on the other hand, these factors also drive endothelial cells and RPE cells to continue to secrete cytokines [39]. This may explain why the use of steroid drugs can help reduce the recurrence rate of PVR in the early stage. Therefore, while using steroid drugs after RRD surgery, individualized treatment for each patient and different stages of PVR should be considered.

The rate of retinal reattachment in the steroid treatment group was higher, and the rate of reoperation in the steroid treatment group was lower, but there was no significant difference. Acar et al. [12] found that the TA group (87.50%) had a higher rate of reattachment than that of the control group (78.12%), although it was not statistically significant. Two of the studies also involved reattachment of the posterior retina, and there was no statistical difference between the two groups. In addition, there is no difference in the number of second operations after surgery in RRD patients between the two groups. The reason why steroid treatment effect is not significantly better than that of the control group may be the following two aspects: (1) comparing the retinal reattachment rate, RRD (PVR grades A, B, and C) patients were included. Comparing the rate of reoperation, only RRD (PVR grade C) patients were included. From the conclusion of the present meta-analysis, RRD surgery combined with steroid drugs administration could significantly reduce the recurrence in PVR grade A and B subgroups, while not affecting that of the PVR grade C subgroup; (2) steroids have low solubility in silicone oil, which may possibly cause steroids hard to diffuse freely in the eye to achieve effective therapeutic concentrations and to show a beneficial effect [40].

The regression rate of macular edema in the steroid group was higher than that in the control group. No statistical difference of the incidence of macular pucker (or ERM) between the two groups was observed, but the rate of surgery caused by macular pucker (or ERM) at 6 months was significantly decreased by steroid treatment. Increased chemotactic factors and inflammatory cytokines levels is one of the major pathological mechanisms of PVR. The epithelial-mesenchymal transition of RPE cells, the hypertrophy of glial cells, and photoreceptor apoptosis mediated by activated macrophages are all involved in the pathological progression of PVR [2]. PVR is a multifactorial process, and a single drug cannot cover all the mechanisms.

Although this meta-analysis only includes RCTs, there are some limitations that cannot be ignored: (1) only 4 clinical studies were included in this study, and 478 eyes were evaluated in total, which limits our ability to summarize the results, especially with regard to differences between drugs. Due to the small sample size, some conclusions are difficult to draw; (2) the duration of some clinical trials is short, and some observation indexes cannot be concluded. In addition, longer term follow-up is also very necessary, which is helpful to obtain the information of adverse reactions of hormone medication and more objective research results; (3) due to the different observation indexes of each experiment, and considering the real situation of surgery, we did not report the changes of cataract and IOP after surgery; (4) types of steroids and administration methods of the four included clinical studies were different, and other factors, such as the skill of the surgeon, the grads of PVR, the method of surgery, or the dose of steroids used, can lead to the existence of heterogeneity and will all affect our results. In conclusion, in order to enhance the effectiveness of meta-analysis, clinical trials should be conducted with expanded sample size and intraocular injection of steroids to evaluate the efficacy of steroids as an adjuvant after RRD in the future.

Acknowledgments

The authors thank Dr. Lai Xiang for helping with the statistical analysis.

Statement of Ethics

An ethics statement is not applicable because this study is based exclusively on published literature.

Conflict of Interest Statement

The authors have no conflicts of interest to declare.

Funding Sources

This work is supported by the National Natural Science Foundation of China (82171085), the National Natural Science Foundation of China (81900891), and the Tianjin Research Innovation Project for Postgraduate Students (2021YJSB271).

Author Contributions

Manhong Xu designed this study. Manhong Xu and Xiao’e Fan collected and analyzed the data. Xinyuan Huang and Xin Chen generated the figures. Yan Shao reviewed and revised the manuscript. Xiaorong Li and Yan Shao contributed to manuscript development and proofreading and approved the final version of the manuscript. All the authors read and approved the final manuscript.

Funding Statement

This work is supported by the National Natural Science Foundation of China (82171085), the National Natural Science Foundation of China (81900891), and the Tianjin Research Innovation Project for Postgraduate Students (2021YJSB271).

Data Availability Statement

All data generated during this study are included in this article. Further inquiries can be directed to the corresponding author (xiaorli@163.com).