Purpose
To evaluate the optimum medical strategy to prevent cystoid macular edema (CME) after cataract surgery.
Design
Systematic review and meta-analysis.
Methods
setting : Cochrane, MEDLINE, and EMBASE databases were searched to identify eligible randomized controlled trials (RCTs). study population : RCTs comparing medical strategies to prevent CME after uncomplicated cataract surgery in nondiabetic and diabetic patients. observation procedures : Data were extracted by 2 authors independently. Quality of individual RCTs was assessed using the Cochrane Collaboration’s tool for assessing risk of bias and Delphi criteria. main outcome measures : Odds of developing CME within 3 months postoperatively and foveal thickness, macular volume and corrected distance visual acuity change within 3 months postoperatively, as compared to baseline.
Results
Seventeen trials reported incidence rates. Topical nonsteroidal anti-inflammatory drugs (NSAIDs) significantly reduced the odds of developing CME as compared to topical corticosteroids in nondiabetic (odds ratio [OR] 0.11; 95% confidence interval [95% CI] 0.03–0.37) and mixed populations (OR 0.05; 95% CI 0.02–0.11). A combination of topical corticosteroids and NSAIDs significantly reduced the odds of developing CME as compared to topical corticosteroids in nondiabetic (OR 0.21; 95% CI 0.10–0.44) and diabetic patients (OR 0.17; 95% CI 0.05–0.50). Intravitreal corticosteroid or anti–vascular endothelial growth factor injections did not show any additional benefit in diabetic subjects.
Conclusions
Topical NSAIDs significantly reduced the odds of developing CME, as compared to topical corticosteroids, in nondiabetic and mixed populations. A combination of topical NSAIDs and corticosteroids reduced the odds of developing CME in nondiabetic and diabetic patients, as compared to topical corticosteroids.
For many decades, cataract has been the leading cause of blindness in the world. A major current focus in cataract surgery is how to minimize complications and improve postoperative visual recovery. Over the years, the incidence of complications has significantly decreased owing to more advanced surgical techniques. Nowadays, cystoid macular edema (CME) is one of the most prevalent postoperative complications after otherwise uncomplicated cataract surgery.
CME after cataract surgery was first reported in 1953 and is also known as the Irvine-Gass syndrome. It usually develops within 3 months postoperatively, with a peak incidence at 4–6 weeks after surgery. It is considered the most important cause of suboptimal visual acuity within the first weeks postoperatively and strongly affects early recovery. Although CME has been reported in up to 23% of nondiabetic subjects after regular uncomplicated cataract surgery, most cases are self-limiting and patients experience no or only minimal reduction in visual acuity. Approximately 0%–6% of nondiabetic subjects develop visual complaints and suffer from clinically significant macular edema (CSME). In contrast, incidence rates of CSME are up to 56% in diabetic patients with mild to moderate nonproliferative diabetic retinopathy (NPDR) and no CME preoperatively.
In 1998, a review by Rossetti and associates reported that prophylactic anti-inflammatory interventions are effective in reducing the incidence of CME after cataract surgery. Since then, many treatments have been studied in order to identify the optimal preventive treatment. Recently, Kessel and associates compared the efficacy of topical corticosteroids and topical nonsteroidal anti-inflammatory drugs (NSAIDs) in controlling postoperative inflammation and preventing CME after uncomplicated cataract surgery in nondiabetic patients. The systematic review showed less postoperative inflammation and less CME in the NSAID group. This study, however, did not compare any other intervention and did not address the efficacy in diabetic patients.
This systematic review was designed to collect and summarize the results of randomized controlled trials (RCTs) on the prevention of CME after cataract surgery. The current study compares the efficacy of preventive strategies on the odds of developing CME within 3 months after uncomplicated phacoemulsification cataract surgery with posterior chamber intraocular lens implantation in nondiabetic and diabetic patients with age-related cataract, without CME preoperatively and with no predisposing factors for developing CME. A meta-analysis and indirect treatment comparison was performed to compare the efficacy of various preventive treatments.
Methods
The systematic review and meta-analysis was designed using the guidelines of the Cochrane Handbook and result were reported following the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) statement. In accordance with Dutch guidelines, the medical ethics committee of the University Hospital Maastricht and Maastricht University decided that no institutional review board approval was required for this study.
Systematic Review Process
The Cochrane Library (1992 to present), MEDLINE (OVID, 1946 to present), and EMBASE (1947 to present) databases were searched in June 2013 and an update was conducted in July 2014. All search strategies included a combination of text words, including text words for cataract extraction (cataract extr*, phaco*), macular edema (irvine gass, edema, oedema), and study design (random*). The complete search strategy for the MEDLINE database is shown in Supplemental Table 1 (Supplemental Material available at AJO.com ). Similar search strategies were used for the Cochrane Library and EMBASE databases. To prevent exclusion of eligible articles, there were no language, publication status, or date restrictions. Reference lists of all included trials and previously published reviews were searched for additional RCTs by 2 review authors (L.W. and V.L.) independently. No trial registries were searched for unpublished trials and no study authors were contacted to identify additional studies. All records identified were managed using Endnote X7.
Titles and abstracts were scanned for eligibility by 2 review authors (L.W. and V.L.) independently. Discrepancies were resolved by discussion between the 2 authors. Full articles were obtained for all relevant abstracts and were reviewed by the 2 authors for eligibility. Both were unmasked to authors, journal, institution, and trial results during the assessment. In order to provide a complete overview of the available evidence, all RCTs comparing at least 2 preventive strategies of any type, dosage, or form were included in this systematic review. The authors excluded trials investigating the prevention of CME after intracapsular or extracapsular cataract extraction and trials including patients with preoperative CME or a high risk of developing CME postoperatively. Trials investigating the treatment of CME were also excluded.
Data Extraction
All data were extracted in duplicate by 2 authors (L.W. and V.L.) independently. A standard data extraction form was used, including the following items: study size, funding sources, eligibility criteria, type of participants, type of interventions, follow-up period, outcome definition, retinal thickness classified according to the Early Treatment Diabetic Retinopathy Study (ETDRS) retinal thickness map, macular volume (MV), and corrected distance visual acuity (CDVA).
All preventive strategies were classified into predefined treatment groups based on type of intervention and mode of administration (eg, topical corticosteroids or subconjunctival corticosteroids). All preoperative, intraoperative, and postoperative treatments were taken into account, except for once-only intraoperative eye drops that were thought not to influence the effect of additional preventive strategies. If an article included multiple study arms within the same treatment group, these arms were combined by adding the total number of participants in each group.
Risk of bias and quality of the included trials were assessed on study level by 2 reviewers independently of each other, using the Cochrane Collaboration’s tool for assessing risk of bias and Delphi criteria. The Delphi list assesses the quality of RCTs based on treatment allocation (randomization and allocation concealment); baseline prognostic factors; eligibility criteria; masking of outcome assessors, care providers, and patients; presentation of point estimates and measures of variability; and inclusion of an intention-to-treat analysis. The Cochrane Collaboration’s tool for assessing risk of bias assesses various types of bias, including selection bias (sequence generation and allocation concealment); performance bias (masking of participants and personnel); detection bias (masking of outcome assessment); attrition bias (incomplete outcome data); and other sources of bias.
Meta-analysis
Subgroup analyses were performed for patients with and without diabetes mellitus (DM). An additional subgroup analysis was performed for all mixed populations, to compare data of studies that included both nondiabetic and diabetic patients. Trials comparing 2 identical treatment groups (eg, 2 corticosteroid eye drops) were excluded from the meta-analysis, as were data from trials that did not provide any measures of variability.
The primary outcome was the incidence of CME within 3 months after cataract surgery, using the diagnostic tools and definitions of the included trials. In case an article used separate definitions for CME and CSME, the occurrence of CSME was used in the meta-analysis, as this is the most clinically relevant outcome. Secondary outcome measures were the difference in optical coherence tomography (OCT)-measured central foveal thickness (FT) in the central 1-mm area of the macula and MV in the central 6-mm area of the macula within 3 months postoperatively, as compared to baseline. Moreover, this study describes the difference between treatment groups in CDVA change within 3 months postoperatively, as compared to baseline. CDVA data were converted to logMAR scale if necessary. If articles reported only absolute values of baseline and postoperative FT, MV, or CDVA, mean changes were calculated from the available information. The standard deviation (SD) for FT, MV, or CDVA change was imputed from the baseline and postoperative SD, using the methods described in the Cochrane Handbook. If an outcome was measured more than once within 3 months postoperatively, the latest follow-up moment was selected for inclusion in the meta-analysis, with the intention of comparing the most long-term outcome.
Analyses were performed using Review Manager. Odds ratios (ORs) and accompanying 95% confidence intervals (95% CIs) were calculated for all dichotomous outcomes, whereas mean differences with 95% CIs were analyzed for all continuous outcome measures. As described in the Cochrane Handbook, statistical heterogeneity was assessed using the χ 2 test. I 2 was used to describe the percentage of variability in effect estimate thought to be a result of heterogeneity. Whenever possible, treatment groups were compared directly using a classical pairwise meta-analysis. Some indirect comparisons could be performed using the Bucher method. This method can be used if no RCT compared 2 treatment groups directly, while both treatments have been compared to placebo or standard treatment in other trials.
Results
The literature search retrieved 2808 titles and abstracts. Thirty trials were included in this systematic review. Reasons for exclusion in each stage of the article selection process are shown in detail in the flow chart in Figure 1 . Characteristics of included studies are listed in Supplemental Table 2 (Supplemental Material available at AJO.com ). Eleven trials included only patients without DM and 7 trials included only diabetic patients. Twelve other trials included patients with and without DM or did not report the incidence of DM in the study population. These trials were clustered and referred to as “mixed populations.” In the nondiabetic and mixed populations, most trials compared the effect of topical corticosteroids, topical NSAIDs, or a combination of both. By contrast, many different treatments have been compared in diabetic subjects, including sub-Tenon corticosteroids, intravitreal corticosteroids, and intravitreal anti–vascular endothelial growth factor (anti-VEGF) injections. A complete overview of the studied treatment groups and definitions used in the included trials are given in Tables 1–3 for nondiabetic, mixed, and diabetic populations, respectively.
| Study | Definition CME | Treatment Group | Drug | Treatment Group | Drug |
|---|---|---|---|---|---|
| Dieleman et al, 2011 | CME on OCT (any increase in CPT >30% compared with the preoperative baseline value developing within 4 weeks after cataract surgery) in combination with a decrease in CDVA of 2 or more lines on the ETDRS chart | Topical CS | Dexamethasone 0.1% (postop) | Subconjunctival CS | Betamethasone 5.7 mg/mL (during surgery) |
| Donnenfeld et al, 2006 | BCVA worse than 20/30 at the 2-week postoperative visit and CME diagnosed using OCT | Topical CS | Prednisolone 1% (postop) | Topical CS + topical NSAID b | Prednisolone 1% (postop) + ketorolac 0.4% (preop and postop) |
| Donnenfeld et al, 2011 , a | Not given | Topical CS + topical NSAID | Difluprednate 0.05% (postop) + nepafenac 0.1% or ketorolac 0.4% (preop and postop) | Topical CS + topical NSAID | Prednisolone 1% (postop) + nepafenac 0.1% or ketorolac 0.4% (preop and postop) |
| Mathys et al, 2010 , a | An increase of >25 μm in macular thickness in the central 1 mm area on OCT | Topical CS + topical NSAID | Prednisolone 1% (postop) + nepafenac 0.01% (preop) | Topical CS + topical NSAID | Prednisolone 1% (postop) + nepafenac 0.01% (preop) + nepafenac 0.1% (postop) |
| Miyake et al, 1999 | Diagnosed on FA using the Miyake classification | Topical CS | Fluorometholone 0.1% (preop and postop) | Topical NSAID | Diclofenac 0.1% (preop and postop) |
| Miyake et al, 2001 | Diagnosed on FA using the Miyake classification | Topical CS | Fluorometholone 0.1% (preop and postop) | Topical NSAID | Diclofenac 0.5% (preop and postop) |
| Miyanaga et al, 2009 | Decreased VA and obvious CME by OCT | Topical CS | Betamethasone 0.1% and fluorometholone 0.1% (postop) | 1) Topical NSAID 2) Topical CS + topical NSAID | 1) Bromfenac 0.1% (postop) 2) Betamethasone 0.1% and fluorometholone 0.1% (postop) + bromfenac 0.1% (postop) |
| Moschos et al, 2012 | Not given | Topical CS | Dexamethasone 0.1% (postop) | Topical CS + topical NSAID | Dexamethasone 0.1% (postop) + diclofenac 0.1% (preop and postop) |
| Negi et al, 2006 | Snellen BCVA of 6/9 or less and CME on oral FA | Oral AZ + topical CS + subconjunctival CS | AZ 250 mg (end of surgery) + betamethasone 0.1% (postop) + betamethasone 4 mg (during surgery) | Oral AZ + subconjunctival CS + sub-Tenon CS | AZ 250 mg (end of surgery) + betamethasone 4 mg (during surgery) + triamcinolone acetonide 20 or 30 mg (during surgery) |
| Ticly et al, 2014 | Diagnosed on FA using the Miyake classification; CME on OCT defined as the presence of well-defined cystic fluid pockets or a CST above 315 μm | Topical CS | Prednisolone 1% (preop and postop) | Topical CS + topical NSAID | Prednisolone 1% (preop and postop) + ketorolac 0.4% (preop and postop) |
| Wang et al, 2013 | Impaired BCVA, macular alterations during fundus examination, and CRT of >250 μm on OCT and the presence of intraretinal cystoid space beneath the fovea | Oral CS + topical CS b | Prednisolone 15 mg (postop) + fluorometholone 0.1% or dexamethasone 0.1% (postop) | Oral CS+ topical NSAID b | Prednisolone 15 mg (postop) + bromfenac 0.1% (postop) |
| Yavas et al, 2007 | Diagnosed on FA as fluorescein leakage into the cystic space | Topical CS | Prednisolone 1% (postop) | Topical CS + topical NSAID b | Prednisolone 1% (postop) + indomethacin 0.1% (postop or preop and postop) |
a Not included in meta-analysis.
| Study | Definition CME | Treatment Group | Drug | Treatment Group | Drug |
|---|---|---|---|---|---|
| Almeida et al, 2008 , a | Not given | Topical CS | Prednisolone 1% (postop) | Topical CS + Topical NSAID | Prednisolone 1% (postop) + ketorolac 0.5% (preop and postop) |
| Almeida et al, 2012 | Not given | Topical CS | Prednisolone 1% (postop) | Topical CS + topical NSAID b | Prednisolone 1% (postop) + nepafenac 0.1% or ketorolac 0.5% (preop and postop) |
| Cable, 2012 , a | Not given | Topical CS + topical NSAID | Prednisolone 1% (during surgery) and difluprednate (postop) + bromfenac 0.09% (preop and postop) | Topical CS + topical NSAID | Prednisolone 1% (during surgery) and difluprednate (postop) + nepafenac 0.1% (preop and postop) |
| Cervantes et al, 2009 | CSME associated with vision loss | Topical CS | Dexamethasone 0.1% (postop) | Topical CS + topical NSAID | Dexamethasone 0.1% (postop) + nepafenac 0.1% (preop and postop) |
| Chatziralli et al, 2011 , a | Not given | Oral AZ + topical CS + topical NSAID | Acetozolamide 125 mg (preop) + dexamethasone 0.1% (preop and postop) + ketorolac 0.5% (preop) | Oral AZ + topical CS + topical NSAID | Acetozolamide 125 mg + dexamethasone 0.1% (preop and postop) + ketorolac 0.5% (preop and postop) |
| Miyake et al, 2000 | Diagnosed on FA using the Miyake classification | Topical CS | Fluorometholone 0.1% (preop and postop) | Topical NSAID | Diclofenac 0.1% (preop and postop) |
| Miyake et al, 2007 | Diagnosed on FA using the Miyake classification | Topical CS | Fluorometholone 0.1% (preop and postop) | Topical NSAID | Diclofenac 0.1% (preop and postop) |
| Miyake et al, 2011 | Diagnosed on FA using the Miyake classification | Topical CS | Fluorometholone 0.1% (preop and postop) | Topical NSAID | Nepafenac 0.1% (preop and postop) |
| Nishino et al, 2009 , a | FA was performed only when CME was suspected to worsen the VA to less than 0.7 | Topical NSAID | Bromfenac (preop and postop) | Topical NSAID + topical CS + subconjunctival CS | Bromfenac (preop and postop) + fluorometholone 0.1% (postop) + dexamethasone 0.5 mL (during surgery) |
| Weber et al, 2013 , a | Not given | Topical NSAID | Indomethacin 0.1% (preop and postop) | Topical NSAID | Ketorolac 0.5% (preop and postop) |
| Wittpenn et al, 2008 , a | Definite CME: presence of cystoid changes associated with substantial (≥40 μm) retinal thickening on OCT Probable CME: presence of changes in retinal contour and increased macular thickness relative to preoperative baseline, but without definite cystoid changes | Topical CS + topical NSAID | Prednisolone 1% (postop) + ketorolac 0.4% (preop) | Topical CS + topical NSAID | Prednisolone 1% (postop) + ketorolac 0.4% (preop and postop) |
a Not included in meta-analysis.
| Study | Definition CME | Treatment Group | Drug | Treatment Group | Drug |
|---|---|---|---|---|---|
| Ahmadabadi et al, 2010 | Subjective report of decreased vision by the patient, ophthalmoscopic detection of the presence of ME, and confirmation of the diagnosis by FA and OCT examinations at any postoperative visit | Topical CS | Betamethasone 0.1% (postop) | Topical CS + intravitreal CS | Betamethasone 0.1% (postop) + triamcinolone acetonide 2 mg (during surgery) |
| Chae et al, 2014 | >60 mm increase in CST relative to the screening CST value, as assessed by spectral-domain OCT. Diagnosed on FA using the Antcliff classification | Placebo | Intravitreal anti-VEGF | Ranibizumab 0.5 mg (during surgery) | |
| Endo et al, 2010 | Not given | Topical CS | Betamethasone 0.1% and fluorometholone 0.1% (postop) | Topical NSAID | Bromfenac 0.1% (postop) |
| Fard et al, 2011 | Increase in CPT on OCT after cataract surgery | Placebo | Intravitreal anti-VEGF | Bevacizumab 1.25 mg (during surgery) | |
| Kim et al, 2008 | Postsurgical CME was defined as decreased VA and CME on OCT | Topical CS | Prednisolone 1% (postop) | Topical CS + sub-Tenon CS | Prednisolone 1% (postop) + triamcinolone acetonide (during surgery) |
| Singh et al, 2012 | ≥30% increase in CSMT relative to the presurgical baseline measurement | Topical CS | Prednisolone 1% (postop) | Topical CS + topical NSAID | Prednisolone 1% (postop) + nepafenac 0.1% (preop and postop) |
| Udaondo et al, 2011 | ME involving or threatening the center of the macula as defined by the ETDRS | Topical CS | Dexamethasone 0.1% (postop) | Topical CS + intravitreal anti-VEGF | Dexamethasone 0.1 % (postop) + ranibizumab 5 mg (during surgery) |
Eight trials could not be included in the meta-analysis, since ORs and 95% CIs could not be calculated for the reported treatment comparisons. Quantitative measures could not be calculated because no patient developed CME in either treatment group, because there were no measures of variability given by the authors, or because the study compared 2 identical treatment groups as defined in this study (eg, comparing 2 different types of corticosteroid eye drops). A complete overview of all meta-analyses is provided in Supplemental Figures 1–10 (Supplemental Material available at AJO.com ) and a summary of all treatment comparisons is provided in Table 4 .
| Population Studied (No. of Studies) | OR (95% CI) | |||
|---|---|---|---|---|
| Topical NSAID | > | Topical CS | Nondiabetic (3) | 0.11 (0.03–0.37) |
| Mixed population (3) | 0.05 (0.02–0.11) | |||
| Topical NSAID | ≈ a | Topical CS & topical NSAID | Nondiabetic (n.a.) | 0.54 (0.13–2.20) |
| Topical NSAID & topical CS | > | Topical CS | Nondiabetic (4) | 0.21 (0.10–0.44) |
| Diabetic (1) | 0.17 (0.05–0.50) | |||
| Topical CS & intravitreal anti-VEGF | ≈ | Topical CS | Diabetic (1) | 0.13 (0.02–1.21) |
| Topical CS & intravitreal anti-VEGF | ≈ a | Topical CS & intravitreal CS | Diabetic (n.a.) | 0.71 (0.02–28.76) |
| Topical CS & intravitreal CS | ≈ | Topical CS | Diabetic (1) | 0.09 (0.00–1.89) |
| Subconjunctival CS | ≈ | Topical CS | Nondiabetic (1) | 1.18 (0.53–2.62) |
| Subconjunctival CS & sub-Tenon CS & oral AZ | ≈ | Topical CS & subconjunctival CS & oral AZ | Nondiabetic (1) | 0.31 (0.03–3.16) |
| Topical NSAID & oral CS | ≈ | Topical CS & oral CS | Nondiabetic (1) | 0.06 (0.00–1.10) |
Quality of Evidence
Only RCTs were selected for inclusion in this study. Supplemental Table 2 (Supplemental Material available at AJO.com ) contains an assessment of the risk of bias within studies and the quality of the included RCTs. As evident from Supplemental Figures 11 and 12 (Supplemental Material available at AJO.com ), the overall quality of evidence of the included studies was low to moderate. Eleven trials used a table of random numbers to randomize the included patients, 1 trial used centralized randomization by the pharmacy, and 3 trials used envelopes. The method of treatment allocation was unclear in 15 other trials. Six trials stated that the treatment allocation was concealed from the investigators. Six trials were open-label studies.
Nondiabetic Patients
As shown in Table 1 , 9 trials reporting the incidence of CME after cataract surgery in nondiabetic patients could be included in this meta-analysis. The meta-analysis in Figure 2 shows that topical NSAIDs significantly reduced the odds of developing CME after cataract surgery, as compared to topical corticosteroids. The OR was 0.11 (95% CI 0.03–0.37; I 2 0%). The difference in efficacy between topical corticosteroids and NSAIDs was not statistically significant if patients also received oral prednisolone for 7 days postoperatively (OR 0.06; 95% CI 0.00–1.10). A combination of topical corticosteroids and topical NSAIDs reduced the odds of developing CME as compared to topical corticosteroids as a single-drug treatment, with an OR of 0.21 (95% CI 0.10–0.44; I 2 18%). Only 1 RCT provided a direct comparison between topical NSAIDs and a combination treatment of topical corticosteroids and NSAIDs. Unfortunately, it was not possible to include this comparison in the quantitative analyses, since no patient in either treatment group developed CME. Nevertheless, using the common comparator of topical corticosteroids, it was possible to perform an indirect comparison of topical NSAIDs vs a combination treatment of topical NSAIDs and corticosteroids. This comparison showed no statistically significant difference in the odds of developing CME after cataract surgery with an OR of 0.54 (95% CI 0.13–2.20).
One study compared the efficacy of postoperative corticosteroid eye drops to subconjunctival corticosteroids at the end of cataract surgery in nondiabetic patients. This study showed no statistically significant difference in the odds of developing CME between both treatment groups (OR 1.18; 95% CI 0.53–2.62).
None of the studies investigated the change in FT or MV after cataract surgery. Eight trials reported the change in CDVA within 3 months postoperatively or both the preoperative and postoperative CDVA. None of the treatment comparisons showed significant differences in CDVA change from baseline. An overview of these meta-analyses can be found in Figure 3 .
Mixed Populations
Four trials reported the incidence of CME after cataract surgery in mixed populations, including both diabetic and nondiabetic subjects. Table 2 provides a complete overview of the included studies. The meta-analysis shown in Figure 2 suggests that topical NSAIDs significantly reduce the odds of developing CME after cataract surgery, as compared to topical corticosteroids (OR 0.05; 95% CI 0.02–0.11; I 2 0%). This finding was also confirmed by OCT, showing a significantly smaller increase in FT in the topical NSAID group as compared to topical corticosteroids (mean difference −23.20 μm; 95% CI −42.95 to −3.45 μm). Other trials showed that a combination of topical corticosteroids and NSAIDs significantly reduced the postoperative change in MV, as compared to topical corticosteroids as a single-drug treatment. The mean difference was −0.25 mm 3 (95% CI −0.36 to −0.13 mm 3 ; I 2 0%). There was no statistically significant difference between treatment groups in FT change from baseline (mean difference −6.00 μm; 95% CI −15.17 to 3.17 μm). An indirect comparison between topical NSAIDs and a combination of topical corticosteroids and NSAIDs showed no significant difference in the FT change from baseline, with a mean difference of −17.2 μm (95% CI −38.97 to 4.57 μm).
Two trials reported change in CDVA within 3 months postoperatively. None of the comparisons showed significant differences between treatment groups in CDVA change from baseline.
Diabetic Patients
Seven trials included diabetic patients with mild to severe NPDR and no diabetic macular edema preoperatively. Five trials reported the incidence of CME after cataract surgery. A complete overview of all articles is presented in Table 3 . As can be seen in Figure 2 , a combination of topical corticosteroids and topical NSAIDs reduced the odds of developing CME after cataract surgery as compared to topical corticosteroids as a single-drug treatment (OR 0.17; 95% CI 0.05–0.50). Only 1 study compared the efficacy of topical NSAIDs vs topical corticosteroids in diabetic patients, but did not report on the incidence of CME. The difference in FT change from baseline between both treatment groups was not statistically significant (mean difference −17.00 μm; 95% CI −36.37 to 2.37 μm). Nevertheless, the study did show a statistically significantly larger improvement in CDVA in the NSAID group as compared to the topical corticosteroid group (mean difference −0.13 logMAR; 95% CI −0.24 to −0.02 logMAR). Preoperative CDVA in the NSAID and corticosteroid group was 0.24 and 0.16 logMAR (20/35 and 20/29 Snellen), respectively, and improved to −0.09 and −0.04 logMAR (20/16 and 20/18 Snellen) at 6 weeks postoperatively.
While topical treatments have been studied most extensively, several trials investigated the effect of intravitreal treatments on the incidence of CME after cataract surgery in diabetic patients. Two trials showed that intravitreal anti-VEGF injections at the end of cataract surgery did not cause a statistically significant reduction in the odds of developing CME, as compared to placebo (OR 0.68; 95% CI 0.21–2.19) or as an additional treatment to topical corticosteroids (OR 0.13; 95% CI 0.02–1.21). One RCT studied the efficacy of intravitreal triamcinolone acetonide in 41 eyes of diabetic patients who also received topical corticosteroids. Four eyes in the control group and no eyes in the triamcinolone group developed CME, but the treatment effect was not statistically significant (OR 0.09; 95% CI 0.00–1.89). CDVA improved from 0.81 and 0.98 logMAR (20/129 and 20/191 Snellen) preoperatively to 0.13 and 0.09 logMAR (20/27 and 20/25 Snellen) at 3 months postoperatively in the control group and intravitreal triamcinolone group, respectively. The difference between treatment groups in postoperative CDVA change from baseline (mean difference −0.13 logMAR; 95% CI −0.40 to 0.14) was not statistically significant.
An indirect comparison showed no statistically significant difference in the odds of developing CME after intravitreal corticosteroid vs intravitreal anti-VEGF injections in diabetic patients who also received topical corticosteroids (OR 0.71; 95% CI 0.02–28.76).
Results
The literature search retrieved 2808 titles and abstracts. Thirty trials were included in this systematic review. Reasons for exclusion in each stage of the article selection process are shown in detail in the flow chart in Figure 1 . Characteristics of included studies are listed in Supplemental Table 2 (Supplemental Material available at AJO.com ). Eleven trials included only patients without DM and 7 trials included only diabetic patients. Twelve other trials included patients with and without DM or did not report the incidence of DM in the study population. These trials were clustered and referred to as “mixed populations.” In the nondiabetic and mixed populations, most trials compared the effect of topical corticosteroids, topical NSAIDs, or a combination of both. By contrast, many different treatments have been compared in diabetic subjects, including sub-Tenon corticosteroids, intravitreal corticosteroids, and intravitreal anti–vascular endothelial growth factor (anti-VEGF) injections. A complete overview of the studied treatment groups and definitions used in the included trials are given in Tables 1–3 for nondiabetic, mixed, and diabetic populations, respectively.
| Study | Definition CME | Treatment Group | Drug | Treatment Group | Drug |
|---|---|---|---|---|---|
| Dieleman et al, 2011 | CME on OCT (any increase in CPT >30% compared with the preoperative baseline value developing within 4 weeks after cataract surgery) in combination with a decrease in CDVA of 2 or more lines on the ETDRS chart | Topical CS | Dexamethasone 0.1% (postop) | Subconjunctival CS | Betamethasone 5.7 mg/mL (during surgery) |
| Donnenfeld et al, 2006 | BCVA worse than 20/30 at the 2-week postoperative visit and CME diagnosed using OCT | Topical CS | Prednisolone 1% (postop) | Topical CS + topical NSAID b | Prednisolone 1% (postop) + ketorolac 0.4% (preop and postop) |
| Donnenfeld et al, 2011 , a | Not given | Topical CS + topical NSAID | Difluprednate 0.05% (postop) + nepafenac 0.1% or ketorolac 0.4% (preop and postop) | Topical CS + topical NSAID | Prednisolone 1% (postop) + nepafenac 0.1% or ketorolac 0.4% (preop and postop) |
| Mathys et al, 2010 , a | An increase of >25 μm in macular thickness in the central 1 mm area on OCT | Topical CS + topical NSAID | Prednisolone 1% (postop) + nepafenac 0.01% (preop) | Topical CS + topical NSAID | Prednisolone 1% (postop) + nepafenac 0.01% (preop) + nepafenac 0.1% (postop) |
| Miyake et al, 1999 | Diagnosed on FA using the Miyake classification | Topical CS | Fluorometholone 0.1% (preop and postop) | Topical NSAID | Diclofenac 0.1% (preop and postop) |
| Miyake et al, 2001 | Diagnosed on FA using the Miyake classification | Topical CS | Fluorometholone 0.1% (preop and postop) | Topical NSAID | Diclofenac 0.5% (preop and postop) |
| Miyanaga et al, 2009 | Decreased VA and obvious CME by OCT | Topical CS | Betamethasone 0.1% and fluorometholone 0.1% (postop) | 1) Topical NSAID 2) Topical CS + topical NSAID | 1) Bromfenac 0.1% (postop) 2) Betamethasone 0.1% and fluorometholone 0.1% (postop) + bromfenac 0.1% (postop) |
| Moschos et al, 2012 | Not given | Topical CS | Dexamethasone 0.1% (postop) | Topical CS + topical NSAID | Dexamethasone 0.1% (postop) + diclofenac 0.1% (preop and postop) |
| Negi et al, 2006 | Snellen BCVA of 6/9 or less and CME on oral FA | Oral AZ + topical CS + subconjunctival CS | AZ 250 mg (end of surgery) + betamethasone 0.1% (postop) + betamethasone 4 mg (during surgery) | Oral AZ + subconjunctival CS + sub-Tenon CS | AZ 250 mg (end of surgery) + betamethasone 4 mg (during surgery) + triamcinolone acetonide 20 or 30 mg (during surgery) |
| Ticly et al, 2014 | Diagnosed on FA using the Miyake classification; CME on OCT defined as the presence of well-defined cystic fluid pockets or a CST above 315 μm | Topical CS | Prednisolone 1% (preop and postop) | Topical CS + topical NSAID | Prednisolone 1% (preop and postop) + ketorolac 0.4% (preop and postop) |
| Wang et al, 2013 | Impaired BCVA, macular alterations during fundus examination, and CRT of >250 μm on OCT and the presence of intraretinal cystoid space beneath the fovea | Oral CS + topical CS b | Prednisolone 15 mg (postop) + fluorometholone 0.1% or dexamethasone 0.1% (postop) | Oral CS+ topical NSAID b | Prednisolone 15 mg (postop) + bromfenac 0.1% (postop) |
| Yavas et al, 2007 | Diagnosed on FA as fluorescein leakage into the cystic space | Topical CS | Prednisolone 1% (postop) | Topical CS + topical NSAID b | Prednisolone 1% (postop) + indomethacin 0.1% (postop or preop and postop) |
a Not included in meta-analysis.
| Study | Definition CME | Treatment Group | Drug | Treatment Group | Drug |
|---|---|---|---|---|---|
| Almeida et al, 2008 , a | Not given | Topical CS | Prednisolone 1% (postop) | Topical CS + Topical NSAID | Prednisolone 1% (postop) + ketorolac 0.5% (preop and postop) |
| Almeida et al, 2012 | Not given | Topical CS | Prednisolone 1% (postop) | Topical CS + topical NSAID b | Prednisolone 1% (postop) + nepafenac 0.1% or ketorolac 0.5% (preop and postop) |
| Cable, 2012 , a | Not given | Topical CS + topical NSAID | Prednisolone 1% (during surgery) and difluprednate (postop) + bromfenac 0.09% (preop and postop) | Topical CS + topical NSAID | Prednisolone 1% (during surgery) and difluprednate (postop) + nepafenac 0.1% (preop and postop) |
| Cervantes et al, 2009 | CSME associated with vision loss | Topical CS | Dexamethasone 0.1% (postop) | Topical CS + topical NSAID | Dexamethasone 0.1% (postop) + nepafenac 0.1% (preop and postop) |
| Chatziralli et al, 2011 , a | Not given | Oral AZ + topical CS + topical NSAID | Acetozolamide 125 mg (preop) + dexamethasone 0.1% (preop and postop) + ketorolac 0.5% (preop) | Oral AZ + topical CS + topical NSAID | Acetozolamide 125 mg + dexamethasone 0.1% (preop and postop) + ketorolac 0.5% (preop and postop) |
| Miyake et al, 2000 | Diagnosed on FA using the Miyake classification | Topical CS | Fluorometholone 0.1% (preop and postop) | Topical NSAID | Diclofenac 0.1% (preop and postop) |
| Miyake et al, 2007 | Diagnosed on FA using the Miyake classification | Topical CS | Fluorometholone 0.1% (preop and postop) | Topical NSAID | Diclofenac 0.1% (preop and postop) |
| Miyake et al, 2011 | Diagnosed on FA using the Miyake classification | Topical CS | Fluorometholone 0.1% (preop and postop) | Topical NSAID | Nepafenac 0.1% (preop and postop) |
| Nishino et al, 2009 , a | FA was performed only when CME was suspected to worsen the VA to less than 0.7 | Topical NSAID | Bromfenac (preop and postop) | Topical NSAID + topical CS + subconjunctival CS | Bromfenac (preop and postop) + fluorometholone 0.1% (postop) + dexamethasone 0.5 mL (during surgery) |
| Weber et al, 2013 , a | Not given | Topical NSAID | Indomethacin 0.1% (preop and postop) | Topical NSAID | Ketorolac 0.5% (preop and postop) |
| Wittpenn et al, 2008 , a | Definite CME: presence of cystoid changes associated with substantial (≥40 μm) retinal thickening on OCT Probable CME: presence of changes in retinal contour and increased macular thickness relative to preoperative baseline, but without definite cystoid changes | Topical CS + topical NSAID | Prednisolone 1% (postop) + ketorolac 0.4% (preop) | Topical CS + topical NSAID | Prednisolone 1% (postop) + ketorolac 0.4% (preop and postop) |
a Not included in meta-analysis.
| Study | Definition CME | Treatment Group | Drug | Treatment Group | Drug |
|---|---|---|---|---|---|
| Ahmadabadi et al, 2010 | Subjective report of decreased vision by the patient, ophthalmoscopic detection of the presence of ME, and confirmation of the diagnosis by FA and OCT examinations at any postoperative visit | Topical CS | Betamethasone 0.1% (postop) | Topical CS + intravitreal CS | Betamethasone 0.1% (postop) + triamcinolone acetonide 2 mg (during surgery) |
| Chae et al, 2014 | >60 mm increase in CST relative to the screening CST value, as assessed by spectral-domain OCT. Diagnosed on FA using the Antcliff classification | Placebo | Intravitreal anti-VEGF | Ranibizumab 0.5 mg (during surgery) | |
| Endo et al, 2010 | Not given | Topical CS | Betamethasone 0.1% and fluorometholone 0.1% (postop) | Topical NSAID | Bromfenac 0.1% (postop) |
| Fard et al, 2011 | Increase in CPT on OCT after cataract surgery | Placebo | Intravitreal anti-VEGF | Bevacizumab 1.25 mg (during surgery) | |
| Kim et al, 2008 | Postsurgical CME was defined as decreased VA and CME on OCT | Topical CS | Prednisolone 1% (postop) | Topical CS + sub-Tenon CS | Prednisolone 1% (postop) + triamcinolone acetonide (during surgery) |
| Singh et al, 2012 | ≥30% increase in CSMT relative to the presurgical baseline measurement | Topical CS | Prednisolone 1% (postop) | Topical CS + topical NSAID | Prednisolone 1% (postop) + nepafenac 0.1% (preop and postop) |
| Udaondo et al, 2011 | ME involving or threatening the center of the macula as defined by the ETDRS | Topical CS | Dexamethasone 0.1% (postop) | Topical CS + intravitreal anti-VEGF | Dexamethasone 0.1 % (postop) + ranibizumab 5 mg (during surgery) |
Eight trials could not be included in the meta-analysis, since ORs and 95% CIs could not be calculated for the reported treatment comparisons. Quantitative measures could not be calculated because no patient developed CME in either treatment group, because there were no measures of variability given by the authors, or because the study compared 2 identical treatment groups as defined in this study (eg, comparing 2 different types of corticosteroid eye drops). A complete overview of all meta-analyses is provided in Supplemental Figures 1–10 (Supplemental Material available at AJO.com ) and a summary of all treatment comparisons is provided in Table 4 .
| Population Studied (No. of Studies) | OR (95% CI) | |||
|---|---|---|---|---|
| Topical NSAID | > | Topical CS | Nondiabetic (3) | 0.11 (0.03–0.37) |
| Mixed population (3) | 0.05 (0.02–0.11) | |||
| Topical NSAID | ≈ a | Topical CS & topical NSAID | Nondiabetic (n.a.) | 0.54 (0.13–2.20) |
| Topical NSAID & topical CS | > | Topical CS | Nondiabetic (4) | 0.21 (0.10–0.44) |
| Diabetic (1) | 0.17 (0.05–0.50) | |||
| Topical CS & intravitreal anti-VEGF | ≈ | Topical CS | Diabetic (1) | 0.13 (0.02–1.21) |
| Topical CS & intravitreal anti-VEGF | ≈ a | Topical CS & intravitreal CS | Diabetic (n.a.) | 0.71 (0.02–28.76) |
| Topical CS & intravitreal CS | ≈ | Topical CS | Diabetic (1) | 0.09 (0.00–1.89) |
| Subconjunctival CS | ≈ | Topical CS | Nondiabetic (1) | 1.18 (0.53–2.62) |
| Subconjunctival CS & sub-Tenon CS & oral AZ | ≈ | Topical CS & subconjunctival CS & oral AZ | Nondiabetic (1) | 0.31 (0.03–3.16) |
| Topical NSAID & oral CS | ≈ | Topical CS & oral CS | Nondiabetic (1) | 0.06 (0.00–1.10) |
Quality of Evidence
Only RCTs were selected for inclusion in this study. Supplemental Table 2 (Supplemental Material available at AJO.com ) contains an assessment of the risk of bias within studies and the quality of the included RCTs. As evident from Supplemental Figures 11 and 12 (Supplemental Material available at AJO.com ), the overall quality of evidence of the included studies was low to moderate. Eleven trials used a table of random numbers to randomize the included patients, 1 trial used centralized randomization by the pharmacy, and 3 trials used envelopes. The method of treatment allocation was unclear in 15 other trials. Six trials stated that the treatment allocation was concealed from the investigators. Six trials were open-label studies.
Nondiabetic Patients
As shown in Table 1 , 9 trials reporting the incidence of CME after cataract surgery in nondiabetic patients could be included in this meta-analysis. The meta-analysis in Figure 2 shows that topical NSAIDs significantly reduced the odds of developing CME after cataract surgery, as compared to topical corticosteroids. The OR was 0.11 (95% CI 0.03–0.37; I 2 0%). The difference in efficacy between topical corticosteroids and NSAIDs was not statistically significant if patients also received oral prednisolone for 7 days postoperatively (OR 0.06; 95% CI 0.00–1.10). A combination of topical corticosteroids and topical NSAIDs reduced the odds of developing CME as compared to topical corticosteroids as a single-drug treatment, with an OR of 0.21 (95% CI 0.10–0.44; I 2 18%). Only 1 RCT provided a direct comparison between topical NSAIDs and a combination treatment of topical corticosteroids and NSAIDs. Unfortunately, it was not possible to include this comparison in the quantitative analyses, since no patient in either treatment group developed CME. Nevertheless, using the common comparator of topical corticosteroids, it was possible to perform an indirect comparison of topical NSAIDs vs a combination treatment of topical NSAIDs and corticosteroids. This comparison showed no statistically significant difference in the odds of developing CME after cataract surgery with an OR of 0.54 (95% CI 0.13–2.20).
One study compared the efficacy of postoperative corticosteroid eye drops to subconjunctival corticosteroids at the end of cataract surgery in nondiabetic patients. This study showed no statistically significant difference in the odds of developing CME between both treatment groups (OR 1.18; 95% CI 0.53–2.62).
None of the studies investigated the change in FT or MV after cataract surgery. Eight trials reported the change in CDVA within 3 months postoperatively or both the preoperative and postoperative CDVA. None of the treatment comparisons showed significant differences in CDVA change from baseline. An overview of these meta-analyses can be found in Figure 3 .
Mixed Populations
Four trials reported the incidence of CME after cataract surgery in mixed populations, including both diabetic and nondiabetic subjects. Table 2 provides a complete overview of the included studies. The meta-analysis shown in Figure 2 suggests that topical NSAIDs significantly reduce the odds of developing CME after cataract surgery, as compared to topical corticosteroids (OR 0.05; 95% CI 0.02–0.11; I 2 0%). This finding was also confirmed by OCT, showing a significantly smaller increase in FT in the topical NSAID group as compared to topical corticosteroids (mean difference −23.20 μm; 95% CI −42.95 to −3.45 μm). Other trials showed that a combination of topical corticosteroids and NSAIDs significantly reduced the postoperative change in MV, as compared to topical corticosteroids as a single-drug treatment. The mean difference was −0.25 mm 3 (95% CI −0.36 to −0.13 mm 3 ; I 2 0%). There was no statistically significant difference between treatment groups in FT change from baseline (mean difference −6.00 μm; 95% CI −15.17 to 3.17 μm). An indirect comparison between topical NSAIDs and a combination of topical corticosteroids and NSAIDs showed no significant difference in the FT change from baseline, with a mean difference of −17.2 μm (95% CI −38.97 to 4.57 μm).
Two trials reported change in CDVA within 3 months postoperatively. None of the comparisons showed significant differences between treatment groups in CDVA change from baseline.
Diabetic Patients
Seven trials included diabetic patients with mild to severe NPDR and no diabetic macular edema preoperatively. Five trials reported the incidence of CME after cataract surgery. A complete overview of all articles is presented in Table 3 . As can be seen in Figure 2 , a combination of topical corticosteroids and topical NSAIDs reduced the odds of developing CME after cataract surgery as compared to topical corticosteroids as a single-drug treatment (OR 0.17; 95% CI 0.05–0.50). Only 1 study compared the efficacy of topical NSAIDs vs topical corticosteroids in diabetic patients, but did not report on the incidence of CME. The difference in FT change from baseline between both treatment groups was not statistically significant (mean difference −17.00 μm; 95% CI −36.37 to 2.37 μm). Nevertheless, the study did show a statistically significantly larger improvement in CDVA in the NSAID group as compared to the topical corticosteroid group (mean difference −0.13 logMAR; 95% CI −0.24 to −0.02 logMAR). Preoperative CDVA in the NSAID and corticosteroid group was 0.24 and 0.16 logMAR (20/35 and 20/29 Snellen), respectively, and improved to −0.09 and −0.04 logMAR (20/16 and 20/18 Snellen) at 6 weeks postoperatively.
While topical treatments have been studied most extensively, several trials investigated the effect of intravitreal treatments on the incidence of CME after cataract surgery in diabetic patients. Two trials showed that intravitreal anti-VEGF injections at the end of cataract surgery did not cause a statistically significant reduction in the odds of developing CME, as compared to placebo (OR 0.68; 95% CI 0.21–2.19) or as an additional treatment to topical corticosteroids (OR 0.13; 95% CI 0.02–1.21). One RCT studied the efficacy of intravitreal triamcinolone acetonide in 41 eyes of diabetic patients who also received topical corticosteroids. Four eyes in the control group and no eyes in the triamcinolone group developed CME, but the treatment effect was not statistically significant (OR 0.09; 95% CI 0.00–1.89). CDVA improved from 0.81 and 0.98 logMAR (20/129 and 20/191 Snellen) preoperatively to 0.13 and 0.09 logMAR (20/27 and 20/25 Snellen) at 3 months postoperatively in the control group and intravitreal triamcinolone group, respectively. The difference between treatment groups in postoperative CDVA change from baseline (mean difference −0.13 logMAR; 95% CI −0.40 to 0.14) was not statistically significant.
An indirect comparison showed no statistically significant difference in the odds of developing CME after intravitreal corticosteroid vs intravitreal anti-VEGF injections in diabetic patients who also received topical corticosteroids (OR 0.71; 95% CI 0.02–28.76).
Discussion
The current paper compared the efficacy of predefined preventive strategies on the incidence of CME after uncomplicated cataract surgery in nondiabetic, mixed, and diabetic populations.
In nondiabetic patients, it was found that topical NSAIDs significantly reduced the odds of developing CME, as compared to topical corticosteroids. These findings are in line with the results of a recent systematic review by Kessel and associates, reporting a significantly higher prevalence of CME in the corticosteroid group as compared to the NSAID group. The systematic review by Kessel and associates did not investigate the additive effects of combining topical corticosteroids and NSAIDs vs single-drug treatment. The current study demonstrated that a combination of topical NSAIDs and corticosteroids significantly reduced the odds of developing CME as compared to topical corticosteroids, while combination treatment did not show any benefit over topical NSAIDs in an indirect treatment comparison. This suggests that a topical NSAID should always be part of the preventive treatment after cataract surgery in nondiabetic patients. Whether the use of corticosteroid eye drops can be avoided cannot be concluded from these results. To establish a better evidence-based preventive strategy, the European Society of Cataract & Refractive Surgeons (ESCRS) decided to design a large multicenter study to compare the efficacy of topical NSAIDs, topical corticosteroids, and a combination treatment of both drugs to prevent the occurrence of CME after cataract surgery in nondiabetic patients.
Several factors may influence the interpretation of the current study results. Firstly, one should consider the difference in potency of various corticosteroid eye drops. It is known that fluorometholone is a corticosteroid with only low potency, owing to a lower penetration of the cornea. Two of 3 RCTs comparing the efficacy of topical corticosteroids vs topical NSAIDs in the nondiabetic study population used low-potency corticosteroids, which might have caused an overestimation of the efficacy of topical NSAIDs as compared to topical corticosteroids.
The comparison of topical NSAIDs vs combination treatment with topical corticosteroids and NSAIDs was based on an indirect comparison through the common comparator of topical corticosteroids. In the absence of a direct comparison, the Bucher method can be used to compare results of different study populations. However, it should be noted that the Bucher method assumes treatment effects to be constant across different populations. Most trials compared the effect of topical NSAIDs to low-potency corticosteroids, whereas the effect of topical combination treatment was compared to higher-potency corticosteroids. This may have caused an overestimation of the effect of topical NSAIDs as compared to combination treatment with topical corticosteroids and NSAIDs.
A major limitation of this study is the use of various detection methods and definitions to diagnose CME after cataract surgery. Any differences between trials in the efficiency of detecting CME may lead to inaccurate estimates of direct and especially indirect comparisons. Four studies in the nondiabetic population used similar definitions to detect CME on FA. OCT was used in 2 other studies, but was only performed in patients with a decreased VA. As CME often occurs without any visual complaints, the incidence of CME will be higher in studies using FA in all patients. One should note the large differences in the incidence of CME in 3 studies using FA to detect CME (15.6%–19.2%), as compared to 2 studies performing an OCT only in patients with decreased VA (1.4%–4%). The studies using FA in all patients most likely include patients with normal CDVA and may overestimate the incidence of clinically relevant CME, since CME will resolve spontaneously in many cases. Future studies would therefore benefit from standardization of the definition of CME after cataract surgery.
The timing of follow-up visits is a third factor that might influence the interpretation of direct and indirect treatment comparisons. It is known that most cases of CME occur within 3 months after cataract surgery, with a peak incidence at 4–6 weeks postoperatively. Therefore, the detection rate of CME will be highest in studies with a follow-up of 4–6 weeks postoperatively. The RCTs investigating the efficacy of topical single-drug treatments scheduled their follow-up visits closer to 4–6 weeks postoperatively, as compared to studies investigating the effect of combination treatments. Consequently, in an indirect comparison, the effect of combination treatments of topical corticosteroids and NSAIDs might be overestimated when compared to topical NSAIDs.
This study was not designed to investigate the optimal duration of pre- or postoperative topical treatments. Although the anti-inflammatory treatment is usually initiated postoperatively, it is thought that the use of preoperative NSAIDs limits the release of prostaglandins during surgery. Two studies found a significantly lower incidence of CME after cataract surgery in patients starting NSAID treatment 1–3 days preoperatively as compared to patients treated only postoperatively.
In mixed populations, topical NSAIDs significantly reduced the odds of developing CME, as compared to topical corticosteroids. All studies were performed by the same author and used the Miyake classification to diagnose CME on FA. As mentioned previously, the definition used to detect CME after cataract surgery may highly influence the reported incidence rates of CME. Therefore, it would be helpful to use an objective outcome measurement (eg, change in postoperative FT or MV) to compare the efficacy of various treatments. One study showed a significantly smaller increase in FT after cataract surgery and postoperative topical nepafenac as compared to postoperative treatment with topical fluorometholone. This may suggest that a combination of topical NSAIDs and corticosteroids would also reduce the change in FT after cataract surgery as compared to topical corticosteroid treatment. However, another study could not find a statistically significant difference between eyes treated with a combination of topical nepafenac plus dexamethasone vs dexamethasone eye drops as a single-drug therapy. This deviation may be caused by the use of corticosteroids of various potencies.
The present study also compared the efficacy of various treatments to prevent the occurrence of CME after cataract surgery in diabetic patients. The odds of developing CME were significantly lower after topical combination treatment with an NSAID and corticosteroid, as compared to a single-drug treatment with topical corticosteroids. One multicenter trial treating all patients with prednisolone eye drops for 2 weeks postoperatively, or longer if considered necessary to treat anterior segment inflammation, provided evidence for this observation. However, an underestimation of the overall incidence of CME in this study may have occurred, as both anterior segment inflammation and CME are a consequence of the underlying inflammatory process after cataract surgery. Although none of the trials studied the efficacy of topical single-drug treatments on the odds of developing CME after cataract surgery, 1 study did investigate the efficacy of bromfenac vs fluorometholone eye drops in reducing postoperative foveal thickening. This trial reported no significant differences in FT change from baseline between the NSAID and corticosteroid treatment groups. A subanalysis including only patients with NPDR showed a statistically smaller FT change in the bromfenac group at 4 and 6 weeks postoperatively. As mentioned previously, fluorometholone can be considered a corticosteroid of only low potency, which might have overestimated the effect of NSAID eye drops in this study.
Whereas different postoperative treatments did not affect CDVA in nondiabetic and mixed populations, diabetic patients showed a statistically significantly larger improvement in CDVA using topical NSAIDs as compared to topical corticosteroids.
Although topical treatments have been studied most extensively, some comparisons could be made regarding the use of intravitreal treatments in diabetic patients. Three studies showed that intravitreal corticosteroid and anti-VEGF treatments might be useful to reduce the odds of developing CME after cataract surgery, but none of the studies was able to show statistically significant results owing to small sample sizes of the included studies. One study showed that a preoperative injection of 0.5 mg ranibizumab prevents the occurrence of CME after cataract surgery at 1 month postoperatively, but this difference disappeared at 3 months postoperatively. This may be caused by the limited duration of the treatment effect of intravitreal ranibizumab. An indirect treatment comparison could not find a statistically significant difference in the efficacy of intravitreal corticosteroid vs intravitreal anti-VEGF injections in diabetic patients who also received a topical corticosteroid. When considering the abovementioned treatment strategies, cataract surgeons are obliged to carefully consider the adverse events reported after intravitreal injections. It should be noted that increased intraocular pressure is reported in 41.2 per 100 patients after intravitreal corticosteroid injection, whereas the incidence is much lower (0.12–0.49 per 100 patients) after intravitreal bevacizumab or ranibizumab injection. Other safety concerns with intravitreal injections are mainly associated with the injection procedure. Endophthalmitis, the most dreaded complication after intravitreal injection, was reported in 0.04–0.05 per 100 injections in a recent systematic review summarizing the adverse events after intravitreal anti-VEGF injections reported in 278 articles. Given the paucity of studies on the prevention of CME after cataract surgery in diabetic patients, the ESCRS also decided to set up a European multicenter study investigating the efficacy of intravitreal anti-VEGF and subconjunctival corticosteroid injections to prevent the occurrence of CME after cataract surgery in diabetic patients.
The purpose of this study was to summarize and integrate the results of previous RCTs on the prevention of CME within 3 months after uncomplicated phacoemulsification cataract surgery with posterior chamber intraocular lens implantation in nondiabetic and diabetic patients with age-related cataract, without CME preoperatively, and without risk factors for developing CME. Results of this meta-analysis show that topical NSAIDs significantly reduced the odds of developing CME, as compared to topical corticosteroids in nondiabetic and mixed populations. Furthermore, a combination of topical NSAIDs and corticosteroids significantly reduced the odds of developing CME in nondiabetic and diabetic patients, as compared to topical corticosteroids in a single-drug treatment. Based on an indirect treatment comparison, no difference could be found between topical combination treatment and topical NSAIDs in nondiabetic patients. None of the included studies was able to show a statistically significant effect of intravitreal corticosteroid or intravitreal anti-VEGF treatments to reduce the odds of developing CME after cataract surgery in diabetic patients.
Funding/Support: This study was funded by the European Society of Cataract and Refractive Surgeons (ESCRS), Blackrock, Ireland. The ESCRS was not involved in any part of the study design, data analysis and interpretation, or writing of the report, or in the decision to submit the article for publication. Financial disclosures: Rudy M.M.A. Nuijts is a consultant at Alcon (Vilvoorde, Belgium), Asico (Westmont, Illinois), and TheaPharma (Clermont-Ferrand, France) and receives financial support for research from AcuFocus (Irvine, California), Alcon, Bausch & Lomb (Montpellier, France), Gebauer (Neuhausen, Germany), HumanOptics (Erlangen, Germany), Ophthec (Groningen, Netherlands), and PhysIOL (Liège, Belgium). The following authors have no financial disclosures: Laura H.P. Wielders, Verena A. Lambermont, Jan S.A.G. Schouten, Frank J.H.M. van den Biggelaar, Gill Worthy, Rob W.P. Simons, and Bjorn Winkens. All authors attest that they meet the current ICMJE criteria for authorship.
The authors thank Mor M. Dickman and Natalia Y. Makhotkina (University Eye Clinic Maastricht UMC+, Maastricht University Medical Center+, Maastricht, Netherlands) for assessing the Hebrew and Russian articles.
Appendix
