`Macular Edema by Steroid and Nonsteroidal
`Anti-inflammatory Eye Drops
`A Systematic Review
`
`Line Kessel, MD, PhD,1,2 Britta Tendal, PhD,2 Karsten Juhl Jørgensen, MD, DrMedSci,2,3 Ditte Erngaard, MD,4
`Per Flesner, MD, PhD,5 Jens Lundgaard Andresen, MD, PhD,6 Jesper Hjortdal, MD, DrMedSci7
`
`Purpose: Favorable outcome after cataract surgery depends on proper control of the inflammatory response
`induced by cataract surgery. Pseudophakic cystoid macular edema is an important cause of visual decline after
`uncomplicated cataract surgery.
`Design: We compared the efficacy of topical steroids with topical nonsteroidal anti-inflammatory drugs
`(NSAIDs) in controlling inflammation and preventing pseudophakic cystoid macular edema (PCME) after un-
`complicated cataract surgery.
`Participants: Patients undergoing uncomplicated surgery for age-related cataract.
`Methods: We performed a systematic literature search in Medline, CINAHL, Cochrane, and EMBASE da-
`tabases to identify randomized trials published from 1996 onward comparing topical steroids with topical NSAIDs
`in controlling inflammation and preventing PCME in patients undergoing phacoemulsification with posterior
`chamber intraocular lens implantation for age-related cataract.
`Main Outcome Measures: Postoperative inflammation and pseudophakic cystoid macular edema.
`Results: Fifteen randomized trials were identified. Postoperative inflammation was less in patients ran-
`domized to NSAIDs. The prevalence of PCME was significantly higher in the steroid group than in the NSAID
`group: 3.8% versus 25.3% of patients, risk ratio 5.35 (95% confidence interval, 2.94e9.76). There was no sta-
`tistically significant difference in the number of adverse events in the 2 treatment groups.
`Conclusions: We found low to moderate quality of evidence that topical NSAIDs are more effective in
`controlling postoperative inflammation after cataract surgery. We found high-quality evidence that topical NSAIDs
`are more effective than topical steroids in preventing PCME. The use of topical NSAIDs was not associated with
`an increased events. We recommend using topical NSAIDs to prevent inflammation and PCME after routine
`cataract surgery. Ophthalmology 2014;121:1915-1924 ª 2014 by the American Academy of Ophthalmology.
`
`Open access under CC BY-NC-ND license.
`
`Supplemental material is available at www.aaojournal.org.
`
`Cataract surgery is one of the most frequently performed
`elective surgical procedures in developed countries. The
`surgical methods have improved significantly over
`the
`years, thus lowering the risk of complications and raising
`patients’ and surgeons’ expectations of a successful visual
`outcome. In patients without other eye diseases, 20/20 vi-
`sual outcome is a realistic expectation.
`Like other types of surgery, cataract surgery induces a
`surgical inflammatory response. Uncontrolled inflammation
`may lead to serious side effects, such as posterior synechia,
`uveitis, and secondary glaucoma. Management of inflam-
`mation is thus a mainstay in modern cataract surgery.
`Currently, 2 drug groups are available to control ocular
`inflammation: steroids and nonsteroidal anti-inflammatory
`drugs (NSAIDs). Steroids are potent anti-inflammatory
`agents that work by acting on a number of intercellular
`
`inflammatory mediators, and NSAIDs work by inhibiting
`the cyclooxygenase enzymes. The cyclooxygenase enzymes
`catalyze the formation of prostaglandins and thromboxanes.
`Prostaglandins mediate inflammatory reactions. Preventing
`the formation of prostaglandins reduces the inflammatory
`process.
`Pseudophakic cystoid macular edema (PCME, also
`termed “IrvineeGass syndrome”) is a swelling of the fovea
`due to fluid accumulation occurring a few weeks to months
`after cataract surgery. It is the most common cause of visual
`decline after cataract surgery. The prevalence of PCME
`varies from study to study depending on how PCME is
`defined. By using fluorescein angiography, a prevalence of
`PCME of up to 20% has been reported,1,2 whereas only 2%
`were diagnosed with PCME when loss of visual acuity was
`required to establish the diagnosis.1,3 Usually, PCME is
`
`Ó 2014 by the American Academy of Ophthalmology
`Published by Elsevier Inc.
`Open access under CC BY-NC-ND license.
`
`http://dx.doi.org/10.1016/j.ophtha.2014.04.035
`ISSN 0161-6420/14
`
`1915
`
`SENJU EXHIBIT 2202
`LUPIN v. SENJU
`IPR2015-01097
`
`Page 1 of 10
`
`
`
`Ophthalmology Volume 121, Number 10, October 2014
`
`subclinical and self-limiting, but in a few patients it may
`become chronic, resulting in permanent visual loss.
`The cause of PCME is thought
`to be an increased
`vascular permeability induced by inflammatory mediators
`such as prostaglandins. Some reports have found an
`increased risk of PCME in patients using prostaglandin
`analogs to control glaucoma.4,5 There is a tendency toward a
`higher prevalence of PCME in patients with increased
`postoperative inflammation.2 The relationship between
`inflammation and PCME is further supported by the 3-
`fold increase in the risk of PCME in patients with a his-
`tory of uveitis.6 Macular thickness is greater in patients with
`complicated cataract surgery compared with uncomplicated
`surgery.7 Increased surgical trauma such as iatrogenic iris
`lesion increases the risk of PCME.1 Furthermore, the risk
`of PCME is increased in patients with a history of retinal
`venous occlusion or an epiretinal membrane,3 whereas
`posterior vitreous detachment seems to protect against
`PCME.1
`Deciding which anti-inflammatory agent to use as stan-
`dard in patients undergoing cataract surgery is important to
`ensure a favorable outcome. The present systematic review
`compares the efficacy of topical steroids with that of topical
`NSAIDs in reducing postoperative inflammation and pre-
`venting PCME. The study was initiated by the Danish
`Health and Medicines Authorities to formulate evidence-
`based national guidelines on the management of age-
`related cataract.
`
`Sources and Methods of Literature Search
`
`We performed this systematic review and subsequent meta-
`analyses on the basis of the principles described in the Grades of
`Recommendation, Assessment, Development, and Evaluation
`(GRADE) approach.8 We first defined the topic of the systematic
`review using the Patient,
`Intervention, Comparison,
`and
`Outcome approach.9 We compared the efficacy of steroid eye
`drops (Intervention) with NSAID eye drops (Comparison) in
`preventing inflammation (Outcome) and PCME (Outcome) after
`uncomplicated cataract surgery by phacoemulsification with
`posterior chamber intraocular lens implantation in patients with
`age-related cataract (Patients). We included only randomized
`controlled trials in the meta-analysis. We excluded references
`comparing other types of interventions or surgical methods. We
`did not compare the additive effects of steroids plus NSAIDs
`versus steroids or NSAIDs alone because a Cochrane protocol
`covers this topic.10 We included all types of topical steroids and
`topical NSAIDs in the review.
`For outcomes, we analyzed the number of cells and flare as
`inflammation markers measured by laser flare-cell photometry or
`slit-lamp evaluation, PCME as defined in the included studies
`(fluorescein angiograms or optical coherence tomography [OCT]),
`and best-corrected distance visual acuity at last follow-up after
`cataract surgery. The time point for evaluation of inflammation
`was at 2 to 8 days post-surgery. The time point for evaluation of
`PCME was as chosen by the included studies. Risks and adverse
`events associated with the use of topical eye drops were also
`quantified using the number of complications as defined in the
`included studies and the intraocular pressure (IOP) after the
`treatment period.
`We performed a systematic literature search in April 2013 in the
`EMBASE, Medline (Ovid), Cochrane Library, and CINAHL
`
`1916
`
`databases. An example of the search strategy for the EMBASE
`database
`is
`provided
`in Appendix
`1
`(available
`at
`www.aaojournal.org). Similar search strategies were used for the
`other databases. The search was limited to references published
`from 1996 and onward in the English or Scandinavian languages.
`The year limitation was chosen to ensure that only studies using
`surgical methods that were comparable to modern date methods
`were included. The literature search was performed by a trained
`information specialist (Birgitte Holm Pedersen). We did not
`search trial registries for unpublished trials. According to Danish
`law, no institutional review board approval was required for the
`study.
`We assessed the risk of bias of each included study using the
`Cochrane risk of bias tool11 in the Review Manager Software
`(Review Manager
`[RevMan] version 5.2. Copenhagen: The
`Nordic Cochrane Centre, The Cochrane Collaboration,
`2012, available at: http://tech.cochrane.org/revman/download,
`Accessed April 2013). In short, the Cochrane risk of bias tool
`assesses risk of bias associated with the selection of patients
`(randomization or patient
`allocation and concealment of
`allocation),
`study performance
`(blinding of patients
`and
`personnel), measurement of outcomes (blinding of outcome
`assessment), attrition of data (e.g., missing patients or dropouts),
`reporting of study findings (selective outcome reporting), or other
`types of bias related to the study design that could affect the
`internal validity. This part of the systematic review was done
`independently by 2 reviewers (BT and KJJ). Disagreement was
`resolved through discussion and consensus.
`We evaluated the quality of the evidence for each prespecified
`outcome across the included studies using the GRADE system in the
`Grade Profiler Software (version 3.6, 2011, available at: http://
`tech.cochrane.org/revman/other-resources/gradepro/download, Acc-
`essed April 2013). We analyzed each outcome for study limitations
`that could affect the outcome (i.e., risk of bias),12 inconsistency
`(different results between studies),13 indirectness (was the study
`population and intervention comparable to the patient population
`and intervention that is relevant to users [external validity], use of
`surrogate measures),14 imprecision (large confidence intervals [CIs]
`or the lack of statistical strength),15 and risk of publication bias
`(small number of studies or included patients, lack of reporting of
`negative findings).16 We upgraded or downgraded the quality of the
`evidence for each of the prespecified outcomes on the basis of the
`assessment of each of the limitations mentioned earlier.
`We analyzed continuous outcome data using mean difference
`and dichotomous outcome data using risk ratios. We used the
`Review Manager 5 Software to calculate estimates of overall
`treatment effects and random-effects models to calculate pooled
`estimates of effects.
`
`Summary of Evidence
`
`Our systematic literature search returned 352 titles and ab-
`stracts, and 82 references were identified by other sources.
`Titles and abstracts were reviewed by 1 reviewer (LK), and
`115 references were judged to be of potential interest by the
`reviewer. These were collected in full text, and 15 ran-
`domized controlled clinical
`trials met our
`inclusion
`criteria.17e31 All included studies excluded patients with
`ocular diseases (e.g., glaucoma, uveitis, previous surgery, or
`trauma), which might affect
`the outcome after surgery.
`Seven of the included trials compared the prophylactic ef-
`fect of topical steroids and NSAIDs on the occurrence of
`cystoid macular edema after cataract surgery.17,25e28,31
`
`Page 2 of 10
`
`
`
`Kessel et al
`
` Topical Steroids versus NSAIDs after Cataract Surgery
`
`Table 1. Overview of Interventions in Included Studies
`
`Study ID
`Asano et al 200817
`
`Steroid
`
`NSAID
`
`Betamethasone sodium 0.1%
`
`Diclofenac sodium 0.1%
`
`Demco et al 199718
`El-Harazi et al 199819
`
`Prednisolone acetate 1.0%
`Prednisolone acetate 1%
`
`Diclofenac sodium 0.1%
`Diclofenac sodium 0.1%
`
`El-Harazi 1998 (steriod B)19
`
`Prednisolone acetate 1%
`
`Ketorolac tromethamine 0.5%
`
`Endo et al 201020
`
`Hirneiss et al 200521
`
`Betamethasone sodium
`phosphate for 1 wk
`and fluorometholone
`0.1% for 5 wks
`Prednisolone acetate 1%
`
`Bromfenac
`
`Ketorolac tromethamine 0.5%
`
`Hirneiss et al 2005 B21
`
`Rimexolone 1%
`
`Ketorolac tromethamine 0.5%
`
`Holzer et al 200222
`
`Loteprednol etabonate 0.5%
`
`Ketorolac tromethamine 0.5%
`
`Laurell and Zetterstrom 200223
`Missotten et al 200124
`
`Dexamethasone phosphate 0.1%
`Dexamethasone 0.1%
`
`Diclofenac sodium 0.1%
`Indomethacin 0.1%
`
`Miyake et al 200028
`
`Miyake et al 200727
`
`Miyake et al 201126
`
`Miyanaga et al 200925
`
`Roberts and Brennan 199529
`Solomon et al 200130
`Wang et al 201331
`
`Fluorometholone 0.1%
`
`Diclofenac 0.1%
`
`Fluorometholone 0.1%
`
`Diclofenac 0.1%
`
`Fluorometholone 0.1%
`
`Nepafenac 0.1%
`
`Betamethasone 0.1%
`for 1 mo, then
`fluorometholone 0.1%
`for 1 mo
`Prednisolone acetate 1%
`Rimexolone 1%
`Fluorometholone 0.1%
`
`Bromfenac 0.1%
`
`Diclofenac 0.1%
`Ketorolac tromethamine 0.5%
`Bromfenac sodium 0.1%
`
`Wang et al 2013 B31
`
`Dexamethasone 0.1%
`
`Bromfenac sodium 0.1%
`
`NSAID ¼ nonsteroidal anti-inflammatory drug.
`
`Dosing
`1 drop 3 hrs, 2 hrs, 1 hr, and 1/2
`hr preoperatively and
`then 3/day for 8 wks
`4/day from the first postoperative day
`4/day from the first postoperative
`day for 1 wk, then 2/day for 3 wks
`4/day from the first postoperative
`day for 1 wk, then 2/day for 3 wks
`Steroid group: 4/day for 5 wks
`NSAID group: 2/day for 5 wks
`
`6 drops/day on days 1e3,
`5 drops/day on days 4e10,
`4 drops/day on days 11e14,
`3 drops/day on days 15e18,
`2 drops/day on days 19e21,
`1 drop/day on days 22e28
`6 drops/day on days 1e3,
`5 drops/day on days 4e10,
`4 drops/day on days 11e14,
`3 drops/day on days 15e18,
`2 drops/day on days 19e21,
`1 drop/day on days 22e28
`1 drop 4/day the first week
`after surgery, then 1 drop 2/day for
`the remainder of the study
`4/day the first week, then 2/day for 3 wks
`4/day beginning the day before surgery
`and for 30 days postoperatively
`1 drop 3 hrs, 2 hrs, 1 hr, and 1/2 hr before
`surgery, then 3/day for 8 wks
`1 drop 3 hrs, 2 hrs, 1 hr, and 1/2 hr before
`surgery, then 3/day for 5 wks
`3/day starting the day before surgery
`until 5 wks postoperatively
`Steroid group: 4/day for 8 wks
`NSAID group: 2/day for 8 wks
`
`4/day for 1 wk, then 2/day for 3 wks
`4/day beginning immediately after surgery
`Steroid group: 3/day for 1 mo
`NSAID group: 2/day for 1e2 mos
`Steroid group: 3/day for 1 mo
`NSAID group: 2/day for 1e2 mos
`
`Characteristics and risk of bias assessments of the included
`studies
`are provided in Appendix 2 (available
`at
`www.aaojournal.org). A list of excluded studies with
`reasons for exclusion is provided in Appendix 3 (available
`at www.aaojournal.org).
`The included studies compared different types of steroids
`with different
`types of NSAIDs. Table 1 provides an
`overview of the included interventions and comparisons.
`
`Prevention of Inflammation
`
`The anti-inflammatory effect of topical NSAIDs and steroid
`eye drops after cataract surgery was evaluated by examining
`signs of intraocular inflammation: cells and flare. Some
`
`studies used laser cell-flare photometry, and others used a
`slit-lamp to identify inflammatory signs. Those studies that
`used a slit-lamp did not consistently use comparable grading
`systems, which made their inclusion in a meta-analysis
`difficult. For this reason, we chose to include only studies
`evaluating inflammation by laser cell-flare photometry in
`our meta-analysis. All included studies used a study design
`in which patients with a history of ocular inflammation
`(iritis or uveitis) had been excluded from the study.
`
`Inflammation Measured as Number of Cells
`
`Only 4 of the included studies reported on the number of
`cells as evaluated by laser cell-flare photometry. We did not
`
`1917
`
`Page 3 of 10
`
`
`
`Ophthalmology Volume 121, Number 10, October 2014
`
`Figure 1. Forest plot comparing the effect of topical steroid versus nonsteroidal anti-inflammatory drug (NSAID) eye drops on inflammation quantified as
`the number of cells detected by laser cell-flare photometry (photons/ms) at 1 week postoperatively. CI ¼ confidence interval; df ¼ degrees of freedom; IV ¼
`inverse variance; SD ¼ standard deviation.
`
`find a significant difference in the number of cells detected
`by laser cell-flare photometry at 1 week postoperatively
`between patients randomized to steroid or NSAID eye
`drops. The mean difference was 1.01 (95% CI, 0.78 to
`2.81; I2 29%). All 4 studies used steroid eye drops of low to
`medium potency: prednisolone,19,29
`loteprednol,22 or
`fluorometholone.28 The meta-analysis is shown in Figure 1.
`
`Inflammation Measured as Flare
`
`We found that topical NSAIDs were more effective than
`steroid eye drops in reducing postoperative inflammation
`measured as the amount of flare by laser flare photometry at
`1 week postoperatively. The mean difference was 6.88 (95%
`CI, 3.26e10.50; I2 89%). However, steroids of medium to
`high potency (betamethasone, dexamethasone, loteprednol,
`and prednisolone) were not significantly different from
`
`NSAIDs in controlling inflammation, whereas steroids of
`low potency (fluorometholone) were significantly less
`effective in controlling inflammation (Fig 2).
`
`Pseudophakic Cystoid Macular Edema
`
`We identified 7 randomized clinical trials that compared the
`prevalence
`of
`PCME
`after
`topical
`steroid
`or
`NSAID.17,20,25e28,31 One of the 7 studies reported foveal
`thickness measured by OCT in patients with diabetes mel-
`litus and was excluded from the analysis of PCME.20 Thus,
`all 6 studies included in this meta-analysis used a study
`design in which patients with a history of uveitis, diabetes,
`or diabetic retinopathy were excluded from participation.
`Four studies evaluated the presence of PCME by fluorescein
`angiography 5 weeks after cataract surgery.17,26e28 The
`remaining 2 studies evaluated the presence of PCME by
`
`Figure 2. Topical steroid versus nonsteroidal anti-inflammatory drug (NSAID) eye drops on preventing postoperative inflammation quantified by laser flare
`photometry (photons/ms) at 1 week after cataract surgery. CI ¼ confidence interval; df ¼ degrees of freedom; IV ¼ inverse variance; SD ¼ standard
`deviation.
`
`1918
`
`Page 4 of 10
`
`
`
`Kessel et al
`
` Topical Steroids versus NSAIDs after Cataract Surgery
`
`Figure 3. Topical steroid versus nonsteroidal anti-inflammatory drug (NSAID) for preventing cystoid macular edema at 1 month after cataract surgery. CI ¼
`confidence interval; df ¼ degrees of freedom; M-H ¼ ManteleHaenszel.
`
`OCT 1 month after cataract surgery.25,31 Some of the pa-
`tients received highly potent steroids (betamethasone or
`dexamethasone),17,25,31 whereas others received a less
`potent steroid (fluorometholone).26e28 In the steroid group,
`25.3% of patients had PCME at 1 month versus 3.8% in the
`NSAID group (risk ratio, 5.35; 95% CI, 2.94e9.76; I2 0%).
`Potent and weaker steroids were both less effective than
`NSAIDs, and there was no indication that potent steroids
`were more effective than weaker steroids (P ¼ 0.74, test for
`subgroup difference) (Fig 3).
`
`Visual Acuity after Cataract Surgery
`
`Four studies reported the visual acuity at the longest follow-
`up 6 to 8 weeks after cataract surgery.17,20,25,31 Best-
`corrected distance visual acuity was on average 0.02
`logarithm of the minimum angle of resolution (95% CI, 0.01
`to 0.05; I2 72%) better in the NSAID group compared with the
`steroid group. This corresponds to 1 letter on the Early
`Treatment of Diabetic Retinopathy Study chart. The differ-
`ence was not statistically significant (P ¼ 0.19) (Fig 4).
`
`Risks and Adverse Events
`
`Both topical steroids and topical NSAIDs can be associated
`with harms. Twelve of the included studies reported the
`number of harms in both treatment groups.17e19,21e24,26,28e31
`Harms ranged from bitter taste to uveitis with hypopyon, but
`the majority of harms were simply reported as “complica-
`tions” without further description. We evaluated the number
`of harms as reported in the included studies in addition to
`study withdrawals due to harms of the treatment. The overall
`prevalence of harms was 5.5% in the steroid group and 6.6%
`in the NSAID group. The difference was not significant (risk
`ratio, 0.76; 95% CI, 0.50e1.15; I2 0%) (Fig 5).
`Nonsteroidal anti-inflammatory drugs have been associ-
`ated with corneal melts, and although all patients had an
`anterior segment slit-lamp examination postoperatively,
`none of the studies specifically reported melts; thus, we
`could not perform a meta-analysis for complications spe-
`cifically related to NSAID use.
`Steroids are known to be associated with a risk of
`increased IOP. As shown in Figure 6, patients who were
`
`Figure 4. Final visual acuity (logarithm of the minimum angle of resolution [logMAR]) at the last follow-up 6 or 8 weeks after cataract surgery in patients
`randomized to topical steroids or topical nonsteroidal anti-inflammatory drug (NSAIDs). CI ¼ confidence interval; df ¼ degrees of freedom; IV ¼ inverse
`variance; SD ¼ standard deviation.
`
`1919
`
`Page 5 of 10
`
`
`
`Ophthalmology Volume 121, Number 10, October 2014
`
`Figure 5. Number of complications as defined in the included studies. CI ¼ confidence interval; df ¼ degrees of freedom; M-H ¼ ManteleHaenszel;
`NSAID ¼ nonsteroidal anti-inflammatory drug.
`
`randomized to topical steroids had a statistically significant
`higher IOP at the end of the treatment period than patients
`randomized to topical NSAIDs. The mean difference was
`0.50 mmHg (95% CI, 0.05e0.96; I2 51%). The treatment
`period ranged from 28 days to 2 months. The IOP was
`highest in the group receiving the most potent steroids
`and lowest in the group receiving the least potent steroid,
`
`but the difference between the groups was not statistically
`(P ¼ 0.42 for subgroup difference). Two
`significant
`studies reported the number of patients with a marked
`increase in IOP.21,23 One study21 identified 1 steroid
`responder, who was excluded from the rest of
`the
`analysis. The other
`study did not find any steroid
`responders.23
`
`Figure 6. The intraocular pressure (IOP) at the end of the treatment period (28 days to 8 weeks duration) in patients randomized to topical steroid versus
`topical nonsteroidal anti-inflammatory drug (NSAID) after cataract surgery. CI ¼ confidence interval; df ¼ degrees of freedom; IV ¼ inverse variance.
`
`1920
`
`Page 6 of 10
`
`
`
`Kessel et al
`
` Topical Steroids versus NSAIDs after Cataract Surgery
`
`Table 2. Summary of Findings and Assessment of the Quality of the Evidence
`
`Illustrative Comparative Risks* (95% CI)
`Assumed Risk
`Corresponding Risk
`NSAIDs
`Steroids
`
`Outcomes
`
`Cells 1 wk postoperatively
`by laser cell photometry
`
`Flare 1 wk postoperatively
`by laser photometry
`
`PCME
`Visual acuity at last follow-up,
`logMAR
`
`38/1000
`
`Adverse events as defined by
`study
`IOP at the end of treatment
`
`66/1000
`
`Relative Effect
`(95% CI)
`
`No of Participants
`(Studies)
`
`Quality of the
`Evidence (Grade)
`
`y
`269 (4 studies) 4442 Moderate
`
`yz
`931 (11 studies) 4422 Low
`
`RR 5.35 (2.94e9.76)
`
`yx
`521 (6 studies) 4444 High
`yk
`344 (4 studies) 4422 Low
`
`RR 0.76 (0.50e1.15)
`
`y
`1207 (12 studies) 4442 Moderate
`y{
`969 (12 studies) 4442 Moderate
`
`Mean cells 1 wk postoperatively
`by laser cell photometry in the
`intervention groups were
`1.01 higher (0.78
`lower to 2.81 higher)
`Mean flare 1 wk postoperatively
`by laser photometry in the
`intervention groups was 6.88
`higher (3.26 to 10.5 higher)
`201/1000 (110e366)
`Mean visual acuity at last follow-
`up in the intervention groups
`was 0.02 higher (0.01 lower to
`0.05 higher)
`50/1000 (33e76)
`
`Mean IOP at the end of treatment
`in the intervention groups was
`0.50 higher (0.05 to 0.96 higher)
`
`CI ¼ confidence interval; IOP ¼ intraocular pressure; logMAR ¼ logarithm of the minimum angle of resolution; PCME ¼ pseudophakic cystoid macular
`edema; RR ¼ risk ratio.
`GRADE Working Group grades of evidence.
`High quality: Further research is very unlikely to change our confidence in the estimate of effect.
`Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
`Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
`Very low quality: We are very uncertain about the estimate.
`*The basis for the assumed risk (e.g., the median control group risk across studies) is provided in the footnotes. The corresponding risk (and its 95% CI) is
`based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).
`y
`Risk of selection bias.
`I2 ¼ 89%.
`z
`x
`Risk ratio 6.
`I2 ¼ 72%.
`k
`{
`An effect cannot be ruled out.
`
`Quality of the Evidence
`
`The quality of the evidence for each of the outcomes
`described (number of cells, flare, PCME, visual acuity,
`adverse events, and IOP) was assessed according to the
`criteria defined in the GRADE system.32 A summary of our
`findings and the quality of the evidence are presented in
`Table 2.
`Inflammation, evaluated as the number of cells and flare
`by laser photometry, was less pronounced in the NSAID
`group after 1 week of treatment. The quality of the evidence
`was low to moderate. We downgraded the quality of the
`evidence because of the risks of selection bias and hetero-
`geneity between studies.
`Pseudophakic cystoid macular edema was approximately
`7 times as prevalent in the steroid group compared with the
`NSAID group. The quality of the evidence was high. We
`first downgraded the quality of the evidence because of risk
`of selection bias, and then we upgraded because of the large
`difference in the prevalences.
`There was no significant difference in visual acuity at the
`end of the treatment period in the groups randomized to
`topical steroid or NSAIDs. The quality of the evidence was
`
`low. We downgraded the quality because of risk of selection
`bias in the included studies and large heterogeneity between
`study results.
`There was no difference in the number of adverse events
`as defined in the included studies. We downgraded the
`quality of the evidence to moderate because of risk of se-
`lection bias. The IOP was higher in the steroid group at the
`end of the treatment period. The quality of the evidence was
`downgraded to moderate because of risk of selection bias in
`the included studies.
`
`Discussion
`
`We performed a systematic review and meta-analyses to
`compare the effect of topical steroids with topical NSAIDs
`in controlling inflammation and preventing PCME after
`cataract surgery. We found that topical NSAIDs were more
`effective than even potent topical steroids. Our conclusion
`concerning control of inflammation is based on 931 patients
`randomized to topical
`steroids or NSAIDs, and our
`conclusion concerning PCME is based on 521 randomized
`
`1921
`
`Page 7 of 10
`
`
`
`Ophthalmology Volume 121, Number 10, October 2014
`
`patients. Thus, a large number of patients needs to be
`included in future studies to change our conclusion.
`We did not find evidence for an increased risk of adverse
`events with the use of NSAIDs, but previous reports have
`indicated that prolonged use of topical NSAIDs may be
`associated with a risk of corneal melts33 and impaired
`corneal wound healing.34
`We found high-quality evidence that topical NSAIDs
`are more effective in preventing PCME than topical ste-
`roids. Pseudophakic cystoid macular edema was 6 to 7
`times more prevalent in patients randomized to topical
`steroids compared with topical NSAIDs when evaluated by
`fluorescein angiography or OCT at 4 to 5 weeks after
`cataract surgery. Macular thickness, as assessed by OCT in
`patients without PCME, peaks at approximately 4 to 6
`weeks postoperatively.35e37 Thus, it is not likely that many
`cases of PCME were missed in the included studies. Our
`finding is supported by earlier fluorophotometric findings
`of an earlier reestablishment of the bloodeaqueous barrier
`in NSAID-treated patients compared with steroid-treated
`patients.38
`The quality of the evidence concerning prevention of
`PCME was high, although it may be considered a weakness
`in the generalizability of results that all included studies
`came from Asia; 1 study came from China,31 4 studies came
`from the same Japanese group,17,26e28 and the last study
`came from a second Japanese group.25 Although there is no
`reason to suspect a racial difference in the postoperative
`inflammatory response,
`it would be appreciated if
`the
`findings could be reproduced in a non-Asian population.
`Currently, a multicenter study comparing the effect of topical
`bromfenac with dexamethasone for the prevention of PCME
`is being conducted in cooperation with the European Society
`of Cataract and Refractive Surgeons (available at: http://
`www.clinicaltrialsregister.eu/ctr-search/trial/2012-004873-
`14/NL, Accessed July 2013).
`The studies included in our meta-analyses compared
`different types of topical steroids with different types of
`NSAIDs. Steroids are known to be of different potency,
`with betamethasone and dexamethasone being the most
`potent and fluorometholone and rimexolone being the least
`potent. Difluprednate is a new and possibly more potent
`steroid, but its effect in managing inflammation or pre-
`venting PCME after cataract surgery has not been compared
`with NSAIDs. We grouped our meta-analyses according to
`the strength of the steroids but did not find that the most
`potent steroids were significantly more effective in con-
`trolling inflammation or reducing PCME than the weak
`steroids.
`Five different NSAIDs were used in the included studies.
`Diclofenac was used in 7 studies,17e19,23,27e29 ketorolac
`was used in 4 studies,19,21,22,30 bromfenac was used in 3
`studies,20,25,31 nepafenac was used in 1 study,26 and
`indomethacin was used in 1 study.24 Our meta-analyses
`were not designed to determine which NSAID is most
`effective. Other studies have compared the effect of different
`NSAIDs. Diclofenac has been reported to be more effective
`than flurbiprofen and indomethacin in controlling inflam-
`mation,39 whereas no difference was found for diclofenac
`versus ketorolac.40,41 Ketorolac
`and nepafenac
`seem
`
`1922
`
`equally effective in controlling intraocular inflammation42
`and preventing PCME.43 Ketorolac 0.4% reaches higher
`aqueous humor concentration and lower prostaglandin
`level
`than bromfenac 0.09% in patients with cataract
`randomized to either regimen.44 Thus, we do not have
`evidence to recommend 1 type of NSAID over any other
`type of NSAID.
`Our study did not evaluate when the prophylactic treat-
`ment should be initiated. A few studies have compared
`starting NSAIDs 1 to 3 days before surgery versus on the
`day of surgery or
`the day after surgery. Preoperative
`ketorolac45
`diclofenac46 was
`administration
`of
`and
`significantly more effective in controlling inflammation
`than administration starting the day of surgery or the day
`after surgery. Furthermore, the risk of PCME was lower if
`NSAIDs were administered before surgery.45,47 Thus,
`it
`seems advisable to start NSAIDs 1 to 3 days before planned
`surgery.
`Patients with diabetes mellitus comprise a subgroup of
`patients in whom special attention should be paid to reduce
`the risk of macular edema after cataract surgery. A study
`found that the foveal thickness increased more in patients
`with worse diabetic retinopathy and that 22% of patients
`had PCME.48 Our study was not aimed at evaluating PCME
`in patients with diabetic mellitus,
`and no specific
`recommendations can be given concerning the use of
`steroids or NSAIDs in patients with diabetes mellitus.
`Although control of postoperative inflammation and
`prophylaxis of PCME are important in ensuring a successful
`outcome after cataract surgery, current guidelines49,50 do not
`provide specific recommendations concerning the post-
`operative management of inflammation and prevention of
`cystoid macular edema.
`
`Clinical Recommendation
`
`Topical NSAIDs are more effective than topical steroids in
`preventing inflammation and reducing the prevalence of
`PCME after uncomplicated phacoemulsification with pos-
`terior chamber intraocular lens implantation. We did not
`find any indication that the use of topical NSAIDs was
`associated with a higher risk of adverse events than topical
`steroids nor was there any difference in the visual outcome.
`The IOP was higher in patients randomized to topical
`steroids. We recommend using topical NSAIDs after
`cataract surgery to prevent
`inflammation and macular
`edema.
`Acknowledgment. The authors thank informationist specialist
`Birgitte Holm Pedersen at the Danish Health and Medicines Au-
`thorities for assistance in the literature search.
`
`References
`
`1. Gulkilik G, Kocabora S, Taskapili M, Engin G. Cystoid
`macular edema after phacoemulsification: risk factors and ef-
`fect on visual acuity. Can J Ophthalmol 2006;41:699–703.
`2. Ursell PG, Spalton DJ, Whitcup SM, Nussenblatt RB. Cystoid
`macular edema after phacoemulsification:
`relationship to
`
`Page 8 of 10
`
`
`
`Kessel et al
`
` Topical Steroids versus NSAIDs after Cataract Surgery
`
`blood-aqueous barrier damage and visual acuity. J Cataract
`Refract Surg 1999;25:1492–7.
`3. Henderson BA, Kim JY, Ament CS, et al. Clinical