Ophthalmic NSAIDs’ FDA-approved indications .
Postoperative macular edema is considered one cause of diminished vision after cataract surgery. It was approved that inflammatory mediators especially prostaglandins play a key role in macular edema formation especially in the presence of risk factors that affect blood-retinal barrier such as diabetes, uveitis, tear of posterior capsule, and vitreous loss. So, anti-inflammatory medications like corticosteroids and NSAIDs are the cornerstone of macular edema managements. In spite of using corticosteroids as gold standard for treatment of ocular inflammation, they cannot be used for prolonged period due to associated adverse effects. Lastly, there were many studies about benefits of NSAIDs in management and prevention of macular edema to avoid the side effects of corticosteroids.
- macular edema
- blood-retinal barrier
Macula is an important part of the retina, which is responsible for color vision, contrast sensitivity, sharp vision, communications and interpersonal relationships .
Macular thickening is well-known postoperative complication after cataract surgery, even with uncomplicated small incision phacoemulsification surgery. Subclinical cystoid macular edema (CME) is diagnosed with fluorescein angiography as leakage from perifoveal dilated capillaries without visual acuity affection . Although fluorescein angiography is considered gold standard for diagnosis of macular edema, quantification of fluorescein leakage is difficult. Optical coherence tomography (OCT) nowadays has an upper hand in diagnosis of macular edema because of its advantages as a noninvasive device and can detect macular edema quantitatively and qualitatively .
Clinical CME can be identified on biomicroscopic examination and is associated with decreased visual acuity .
The pathogenesis of CME is disruption of blood-retinal barrier (BRB) by inflammatory mediators generated through several cascades as a result of surgical trauma to iris, ciliary body, or lens epithelial cells. Also, preexisting ocular conditions such as diabetes, hypertension, and uveitis, which affect BRB, can increase risk of CME .
The blood-retinal barrier in diabetic eyes is impaired to a variable degree, which plays a role in development of postoperative CME. CME in diabetic patients is affected by many factors including duration, severity of the disease, presence of retinopathy, and previous treatment with photocoagulation . Total ophthalmic payments were documented to be 47% higher in patients who developed postoperative CME . So, prophylactic prevention or even decreased CME severity is cost savings, particularly among diabetic patients.
2. Macular anatomy and physiology
Being the main part of clear vision, the macula is located in the heart of the retina between upper and lower arcades measured about 4.5–6 mm in diameter. It can be divided into central fovea, surrounded by parafovea and outer perifovea (Figure 1) .
As a whole, the retina blood supply is divided into outer part supplied by choriocapillaris and inner part supplied by retinal vascular plexus branch of posterior ciliary arteries .
3. The blood-retinal barrier
The blood-retinal barrier (BRB) is a barrier that physiologically establishes and maintains specific substrate and ion concentrations to allow proper neural function. BRB regulates flux of substances in retina such as ion, protein, and water and also regulates infiltration of immune competent cells and blood toxins. BRB is formed at two levels, inner and outer BRB. The inner one is composed of tight junctions between retinal vascular endothelial cells. The outer barrier is composed of tight junction between retinal pigment epithelium cells .
Pericytes secrete angiopoietin 1, which induces tight junction protein expression to support endothelial cells barrier .
3.1 Tight junction
Tight junction is mainly apical junctional complex, which has a barrier function against solute flux and movement of proteins and lipids into retinal parenchyma. This junction is showed as transmembrane proteins and junctional adhesion molecule (JAM) .
3.2 Adherens junction
Adherens junction is second barrier beneath the tight junction. This junction is important for development of the barrier as it affects formation of tight junction .
Tight junction: Transmembrane proteins claudins, occludin, and junctional adhesion molecule (JAM) connected with scaffolding protein ZO to actin.
Adherens junction: Vascular endothelial (VE) cadherin connected to actin through complex of β-catenin, α-catenin, and vinculin.
The macula is responsible for central 30 degrees of sharp vision with color vision, interpersonal relationships, communications, and contrast sensitivity . The retina is very sensitive to fluctuation in blood oxygen levels and intraocular changes, as it consumes oxygen more than other tissues, being highly active tissue. Microchanges not felt by patient visual acuity are also not seen by inspection ophthalmoscopy examination .
It has been reported that clinical affection due to CME after uneventful phacoemulsification is between 0 and 9%. Furthermore, clinical affection between 9.1 and 20.4% with angiographic leakage is reported . Interruption of blood-retinal barrier is the most accepted explanation of postoperative macular edema, which causes macular thickening. Surgical trauma disrupts the blood-aqueous barrier, release of prostaglandins, and increase of perifoveal capillaries’ permeability of liquid in extracellular spaces, which cause macular thickening and CME . The pathophysiology of these macular changes may be considered consecutively as follows: (1) release of inflammatory mediators into anterior chamber produced by surgical procedures; (2) removal of normally lens barrier, which separate posterior segment from anterior segment; (3) local effect of inflammatory mediators on macular area; and (4) anterior displacement of vitreous leading to increase traction on macula .
Recently, being noninvasive, OCT has been established to be the main method for examining retinal architecture . OCT can measure microhistological retinal changes in difference to fluorescein angiography, which detects it as a leakage that cannot be detected by biomicroscopy . It was reported that subclinical increase in retinal thickness and volume can be found in the early course postoperatively at 4 weeks after phacoemulsification .
5. Nonsteroidal anti-inflammatory drugs
Surgical trauma stimulates arachidonic acid cascade, which stimulates phospholipase A2 enzyme to release arachidonic acid from membrane phospholipids and produces inflammatory mediators including prostaglandins (PGs) and leukotriene by activation of cyclooxygenase (COX) enzymes. COX-1 and COX-2 isoforms are believed to be the primary mediator of ocular inflammation. PG is an important mediator of postoperative complications, associated with symptoms including pain, ciliary injection, cystoid macular edema, impaired vision, and intraoperative miosis .
So, treatment of ocular inflammation depends mainly on stopping of arachidonic cascades by corticosteroids and nonsteroidal anti-inflammatory drugs (NSAIDs). Corticosteroids inhibit the activity of phospholipase A2 enzyme, but NSAIDs inhibit PGs synthesis irreversibly and nonspecifically by direct inhibition of COX-1 and COX-2 activity  (Figure 3).
In spite of using corticosteroids as gold standard for treatment of ocular inflammation, they cannot be used for prolonged period due to associated adverse effects. Side effects of corticosteroids include increased susceptibility to infections as a result of suppression of host immune response, retardation of corneal wound healing, and increased intraocular pressure (IOP) .
5.2 Nonsteroidal anti-inflammatory drugs
NSAIDs are considered as safety option used for treatment of ocular inflammation. NSAIDs inhibit COX activity patently by several chemically heterogeneous classes .
Currently, the uses of topical NSAIDs in ophthalmology to reduce pain and discomfort after cataract and refractive surgery prevent intraoperative miosis during cataract surgery and manage postoperative inflammation (Table 1) and are also reported to have a role in prevention of CME after cataract surgery [25, 27].
|Diclofenac sodium||Management of postoperative inflammation in patients after cataract extraction|
Decrease of pain and photophobia in patients after corneal refractive surgery
|Flurbiprofen sodium||Maintenance of intraoperative pupil dilatation|
|Ketorolac tromethamine||(0.5%)||Temporary relief of ocular itching associated with seasonal allergic conjunctivitis|
Management of postoperative inflammation in patients after cataract extraction
|(0.4%)||Decrease of ocular pain and burning/stinging in patients undergoing corneal refractive surgery|
|(0.45%)(0.5%)||Reduction of ocular pain and inflammation associated with/following cataract surgery|
|Bromfenac sodium||Management of postoperative inflammation and pain after cataract surgery|
|Nepafenac||Management of inflammation and pain associated with cataract surgery|
NSAIDs have beneficial effects over corticosteroids including analgesia effect, maintaining pupillary dilatation if used preoperatively (Figure 4), and also reduce the risk of secondary infections and increased IOP .
Diclofenac: Plasma levels reached (10 ng/mL) during a 4 hours period after instillation of two drops in each eye .
Flurbiprofen: No information about systemic absorption was approved.
Ketorolac: The use of ketorolac in 26 patients with one drop in each eye three times daily resulted in 5/26 (= 19.2%) with detectable plasma level of ketorolac (10.7–22.5 ng/mL) after 10 days.
Bromfenac: After topical administration, plasma level is expected to be below the detectable limit (50 ng/mL).
Nepafenac: It is an inactive prodrug, which penetrates corneal epithelium and is converted by ocular tissue hydrolases to active form amfenac. Plasma levels of bromfenac and amfenac were detected 2–3 hours in most patients after installation of nepafenac in both eyes but with low plasma level.
5.4. Limitation of NSAIDs
Topical NSAIDs contain some warnings, which include the following:
Prolonged bleeding time .
Cross-sensitivity with phenyl acetic acid derivatives, acetylsalicylic acid, and other NSAIDs .
Delay wound healing mainly if used in combination with corticosteroids .
Topical NSAIDs may cause ocular surface toxicity especially in patients having dry eye syndrome, corneal denervation, corneal epithelial defects, or diabetes .
In the last few years, there were many studies using NSAIDs to prevent postoperative CME in different situations. In Table 2, some studies focus on the effect of NSAIDs in prevention of postoperative CME in diabetic patients. All of them approved significant low macular thickening in diabetic patients receiving NSAID bromfenac only , classical postoperative steroid regimen in addition of nepafenac  or either nepafenac or ketorolac with classical postoperative steroid regimen .
|Endo et al. ||62 (26 patients with nonproliferative diabetic retinopathy)||OCT|
|The perifoveal values were significantly lower in the bromfenac group|
|Singh et al. ||263 patients with nonproliferative diabetic retinopathy||OCT|
|Significantly lower percentage of patients in the nepafenac group developed macular edema|
|Alnagdy et al. ||80 without diabetic retinopathy||OCT|
|Perioperative NSAIDs reduce the frequency and severity of CME in diabetic eyes following cataract surgery|
Other studies documented the effect of NSAIDs in a mixed population (nondiabetic and diabetic patients) without cystoid macular edema preoperatively and with no predisposing factors for developing cystoid macular edema (Table 3). Some studies documented that topical steroid medication may not be absolutely essential after uneventful cataract surgery [37, 44]. Topical nonsteroidal started 1 day before surgery and continued for 1 month by Almeida et al. , while Nishino et al.  started medication postoperatively and continued for 1 month. On the other hand, many studies approved the significant importance of topical NSAIDs in prevention of macular edema [36, 38, 39, 40, 41, 42, 43, 45, 46]. topical ketorolac had more effect in decreasing macular edema than prednisolone only, if started 2 days before surgery and continued for 1 month postoperatively in combination with prednisolone had more effect in decreasing macular edema in comparison to postoperative topical prednisolone alone . Wittpenn et al. approved that starting ketorolac 1 hour before surgery had significant effects in preventing macular edema . Cervantes-Coste et al. documented that diclofenac effectively maintains mydriasis and decreases macular thickness . Also, diclofenac eye drops in Rossetti et al. effectively reduced incidence of angiographic CME and ocular inflammation after cataract surgery , while Miyake et al. suggested that diclofenac effectively decreases CME in comparison to fluorometholone  and also approved its effect in preventing chronological change in choroidal blood flow and disruption of the blood-aqueous barrier . In 2011, there was another study by Miyake et al. which approved that nepafenac was more effective than fluorometholone in preventing cystoid macular edema . In other hand, some studies compared the effect of different NSAIDs similar to the study by Cable et al. who suggested that bromfenac is more effective than nepafenac  and Weber et al. who suggested that there was no change from baseline in retinal thickness between indomethacin group and ketorolac group.
|Almeida et al. ||106||OCT|
|After cataract surgery, ketorolac 0.5% effectively decreases postoperative macular edema|
|Almeida et al. ||193||OCT|
|There was no difference in macular volume between the placebo, ketorolac, and nepafenac|
|Cable et al. ||20 diabetic patients without edema could be enrolled||OCT|
|Postoperative bromfenac showed a trend toward improved vision, less retinal thickening, and more stable macular volumes overall|
|Cervantes-Coste et al. ||60 excluding proliferative DR, macular edema, uncontrolled DM||OCT|
|Nepafenac has a rule in reducing macular edema after cataract surgery and in maintaining intraoperative mydriasis|
|Rossetti et al. 1996 ||88||Ocular inflammation and FFA and|
|Diclofenac eye drops effectively reduced incidence of angiographic CME and ocular inflammation after cataract surgery|
|Miyake et al. ||118 without complications from diabetes mellitus||Fundus fluorescein angiography (FFA)|
|Diclofenac was effective in prevention of CME following cataract surgery|
|Miyake et al. ||62||FFA|
|Diclofenac was more effective than fluorometholone in preventing cystoid macular edema|
|Miyake et al. ||60||OCT|
|Nepafenac was more effective than fluorometholone in preventing cystoid macular edema|
|Nishino et al. ||21||FFA|
|Topical steroid medication may not be absolutely essential after uneventful phacoemulsification|
|Weber et al. ||123||OCT|
|Indomethacin 0.1% had the same effect as ketorolac 0.5% at first day postoperative and more effective than ketorolac 0.5% at first week in treating ocular inflammation after uncomplicated cataract surgery. There was no change from baseline in retinal thickness in two groups|
|Wittpenn et al. ||546||OCT|
|Perioperative ketorolac to postoperative prednisolone significantly reduces the macular thickening after cataract surgery|
Nondiabetic patients who have undergone cataract surgery without preoperative macular edema and with no predisposing factors for developing cystoid macular edema were enrolled in studies in Table 4. In glaucoma patients after cataract surgery, Miyake et al. approved that diclofenac seems to prevent macular edema formation enhanced by latanoprost therapy , and timolol and its preservative benzalkonium chloride .
|Donnenfeld et al. ||100||OCT|
|The preoperative use of ketorolac tromethamine 0.4% for 3 days followed by 1-day preoperatively provided optimum efficacy and superior outcomes relative to 1 hour preoperatively used|
|Mathys et al. ||84||OCT|
|The increase in postoperative macular thickness was small in both the control and treatment groups|
|Miyake et al. ||80 eyes with primary open-angle glaucoma, normal tension glaucoma, or ocular hypertension||FFA|
|Administration of nonsteroidal eye drops such as diclofenac seems to prevent macular edema formation enhanced by latanoprost therapy|
|Miyake et al. ||60 eyes with primary open-angle glaucoma, normal-tension glaucoma, or ocular hypertension||FFA||Diclofenac prevents cystoid macular edema caused by timolol and its preservative, benzalkonium chloride|
|Miyanaga et al. ||72||OCT|
|There were no significant differences in anti-inflammatory effects among the three groups|
|Moschos et al. ||79||OCT|
|The addition of diclofenac did not seem to offer any additional benefit after uneventful phacoemulsification|
|Ticly et al. ||91||FFA|
|The addition of ketorolac tromethamine 0.4% did not seem to offer any additional benefit after uneventful phacoemulsification|
|Wang et al. ||240||OCT|
|Bromfenac sodium was more effective and safer than fluorometholone and dexamethasone as an anti-inflammatory|
|Yavas et al. ||189||FFA|
|Preoperative nonsteroidal anti-inflammatory drugs decreased the incidence of CME more than postoperative use only|
|Capote et al. ||243||OCT|
|Bromfenac is more effective than diclofenac and nepafenac in reducing macular thickness after phacoemulsification|
|McCafferty et al. ||1000||OCT|
|Topical nepafenac 0.3% reduces macular edema in patients with preoperative risk compared to placebo, but there were no differences in patients without risk factors|
|Stock et al. ||77||OCT|
|There were no significant differences in macular thickness between the three groups|
|Milla et al. ||38 glaucoma patients||OCT||Nepafenac has a prophylactic effect aganist postoperative macular edema|
|Tzelikis et al. ||224||Bilateral cataract was included in this study|
Each patient was assigned randomly to receive nepafenac 0.3% drops in one eye and a placebo in the fellow eye
|Nepafenac 0.3% was effective in reducing macular thickness compared with a placebo 5 weeks postoperatively|
Donnenfeld et al. documented that early started ketorolac 3 days or 1 day preoperatively provided superior outcomes over ketorolac startes only 1 hour preoperatively . Also, preoperative indomethacin drugs decreased the incidence of CME more than postoperative use only in the study of Yavas et al. .
In comparative study between different NSAIDs, Capote et al. approved that bromfenac is more effective than diclofenac and nepafenac in reducing macular thickness after phacoemulsification . In other study, bromfenac was more effective and safer in comparison to topical steroid; inspite of using oral prednisones for all patients in the study .
In other studies, NSAIDs did not seem to offer any additional benefit after uneventful phacoemulsification of diclofenac in the study of Moschos et al.  and ketorolac in the study of Ticly et al. . Miyanaga et al. documented that 2 months' use of topical NSAIDs, different topical steroids, or alternating steroids and NSAIDs had no significant differences . Stock et al. suggested no differences between nepafenac, control, and ketorolac through 45-day follow-up .
Mathys et al. told that routine use of preoperative nepafenac may be necessary to achieve excellent visual recovery if continued for 3 weeks postoperatively or not .
In the study of McCafferty et al., postoperative topical nepafenac reduces macular edema in patients with preoperative risk (diabetic retinopathy, contralateral CME, or prostaglandin use) compared to placebo, but there were no differences in patients without risk factors . Nepafenac was effective in reducing macular thickness compared with a placebo in fellow eye 5 weeks postoperatively in patients who had bilateral phacoemulsification enrolled in the study of Tzelikis et al. .
The most important line of management of postoperative macular edema is by prevention. NSAIDs have large effects in prevention of postoperative macular edema with minimal side effects. Furthermore, some studies have suggested that NSAIDs may have a greater effect in re-establishment of the blood-aqueous barrier than corticosteroids. The claim about synergistic effects of NSAIDs and corticosteroids is made by several authors. Although several studies may have favored starting NSAID treatment preoperatively, there are no comparison studies about starting corticosteroids preoperatively.
Conflict of interest
The authors declare that they have no conflict of interest.