Efficacy and Safety of Femtosecond Laser-Assisted Cataract ...tazz.ch/wp-content/uploads/2017/01/Laser_vs_Manual_Cataract_Sur… · Efficacy and Safety of Femtosecond Laser-Assisted

Efficacy and Safety of FemtosecondLaser-Assisted Cataract Surgery Comparedwith Manual Cataract Surgery

A Meta-Analysis of 14 567 Eyes

Marko Popovic, MD(C),1 Xavier Campos-Möller, MD,2,3 Matthew B. Schlenker, MD,2,3

Iqbal Ike K. Ahmed, MD, FRCSC2,3,4

Topic: To investigate the efficacy and safety of femtosecond laser-assisted cataract surgery (FLACS) relativeto manual cataract surgery (MCS).

Clinical Relevance: It is unclear whether FLACS is more efficacious and safe relative to MCS.Methods: A literature search of MEDLINE, EMBASE, and Scopus from 2007 to March 2016 was conducted.

Studies containing both FLACS and MCS arms that reported on relevant efficacy and/or safety parameters wereincluded. Weighted mean differences (WMDs) and risk ratios (RRs) with 95% confidence intervals (CIs) werecalculated.

Results: From 2802 screened articles, 14 567 eyes from 15 randomized controlled trials and 22 observationalcohort studies were included. For primary visual and refractive outcomes, no statistically significant differencewas detected between FLACS and MCS in uncorrected distance visual acuity (WMD, �0.02; 95% CI, �0.04 to0.01; P ¼ 0.19), corrected distance visual acuity (WMD, �0.01; 95% CI, �0.02 to 0.01; P ¼ 0.26), and meanabsolute error (WMD, �0.02; 95% CI, �0.07 to 0.04; P ¼ 0.57). In terms of secondary surgical end points, therewas a statistically significant difference in favor of FLACS over MCS for effective phacoemulsificationtime (WMD, �3.03; 95% CI, �3.80 to �2.25; P < 0.001), capsulotomy circularity (WMD, 0.16; 95% CI, 0.11e0.21;P < 0.001), postoperative central corneal thickness (WMD, �6.37; 95% CI, �11.88 to �0.86; P ¼ 0.02), andcorneal endothelial cell reduction (WMD, �55.43; 95% CI, �95.18 to �15.69; P ¼ 0.006). There was no statis-tically significant difference between FLACS and MCS for total surgery time (WMD, 1.25; 95% CI, �0.08 to 2.59;P ¼ 0.07), capsulotomy circularity using a second formula (WMD, 0.05; 95% CI, �0.01 to 0.12; P ¼ 0.10), andcorneal endothelial cell count (WMD, 73.39; 95% CI, �6.28 to 153.07; P ¼ 0.07). As well, there was a significantlyhigher concentration of prostaglandins after FLACS relative to MCS (WMD, 198.34; 95% CI, 129.99e266.69;P < 0.001). Analysis of safety parameters revealed that there were no statistically significant differences in theincidence of overall complications between FLACS and MCS (RR, 2.15; 95% CI, 0.74 to 6.23; P ¼ 0.16); however,posterior capsular tears were significantly more common in FLACS versus MCS (RR, 3.73; 95% CI, 1.50e9.25;P ¼ 0.005).

Conclusions: There were no statistically significant differences detected between FLACS and MCS in termsof patient-important visual and refractive outcomes and overall complications. Although FLACS did show astatistically significant difference for several secondary surgical outcomes, it was associated with higher pros-taglandin concentrations and higher rates of posterior capsular tears. Ophthalmology 2016;123:2113-2126 ª 2016 by the American Academy of Ophthalmology.

Supplemental material is available at www.aaojournal.org.

Today, more than 9.5 million cataract surgeries are performedeach year worldwide.1 Advances in measurement technology,emergence of phacoemulsification, and invention of foldablelens designs have lead to increasingly safer and morepredictable results. These technologies also have allowedcataract surgery to become a refractive procedure withincreasingly precise postoperative refractive results.2

� 2016 by the American Academy of OphthalmologyPublished by Elsevier Inc.

Manual cataract surgery (MCS) involves the creation ofcorneal incisions with a keratome blade, a continuouscurvilinear capsulorrhexis using forceps or a cystotome, andmanual splitting or cracking of the nucleus followed byphacoemulsification and cortical aspiration. Although thecurrent standard of care procedure confers a favorable effi-cacy and safety profile, complication rates vary by surgeon

2113http://dx.doi.org/10.1016/j.ophtha.2016.07.005ISSN 0161-6420/16

Ophthalmology Volume 123, Number 10, October 2016

and setting after MCS, suggesting that a more automatedprocedure may achieve more reproducible results.3,4

Femtosecond laser-assisted cataract surgery (FLACS) isa technology that uses a laser to replace several of themanual steps of cataract surgery with the goal of improvingaccuracy, safety, and refractive outcomes. Femtosecondlaser-assisted cataract surgery uses a femtosecond laser togenerate free electrons and ionized molecules, which in turnproduce photodisruption and photoionization of opticallytransparent tissue through an acoustic shock wave.5 Thefemtosecond laser is unique because of its shorter pulsetime relative to other ophthalmic lasers.6 Theoretically,lasers with shorter pulse times are able to reduce energyoutput significantly for a given effect, thereby reducingcollateral damage to ocular tissues.

Femtosecond lasers have been used in several differentstages of cataract surgery, including clear corneal incisions,capsulotomy, and lens fragmentation. Femtosecond laser-assisted cataract surgery was approved for cataract surgeryby the United States Food and Drug Administration in2010.7 By 2013, more than 120 000 eyes globally hadundergone FLACS.8 A 2014 survey of new FLACSadopters in the United States showed that 30% of cataractpatients choose FLACS over conventional MCS.9

Given the increasing interest in FLACS, evidence of safetyand efficacy of this technology is needed urgently. In 2013,the Department of Veterans Affairs published a systematicreview in the gray literature that concluded that there was nocurrent benefit in the safety and effectiveness of FLACSrelative to MCS.10 Furthermore, they noted that there weresignificant methodologic concerns in the included studies,including low sample sizes, unclear study methods, fewrandomized controlled trials, issues with patient selection,and financial conflicts of interest. More recently, the firstpublished meta-analysis of FLACS compared with MCSwas conducted in 2015 by Chen et al.11 Analyzing a total of989 eyes and 9 randomized controlled trials, the authorsfound a statistically significant improvement for FLACSover MCS in terms of mean phacoemulsification energyand effective phacoemulsification time; however, there wasno difference for surgical complications. There wereconflicting results for visual outcomes, central cornealthickness, and endothelial cell count depending on thelength of follow-up at which outcomes were compared.

An updated and comprehensive meta-analysis of peer-reviewed clinical studies comparing FLACS with MCS isneeded. This synthesis would be useful to clinicians, policymakers, and researchers who are interested in identifying therole of FLACS. Thus, we performed a meta-analysis toinvestigate the comparative efficacy and safety of FLACSrelative to MCS in published clinical studies.

Methods

Search Strategy

Using Ovid MEDLINE (2007eMarch 2016, week 2), MEDLINEIn-Process and Other Non-Indexed Citations (up to March 18,2016), EMBASE (2007e2016, week 12), and Scopus(2007eMarch 2016), a systematic search of the literature was

2114

performed (Appendix 1AeC, available at www.aaojournal.org).Reference lists of included articles and pertinent reviews alsowere searched.

Eligibility Criteria

Studies were included if they met the following criteria: (1) ran-domized controlled trials or prospective or retrospective observa-tional cohort studies; (2) studies that included only patients whounderwent cataract surgery; (3) studies that provided safety or ef-ficacy data, or both, for both FLACS and MCS study arms; and (4)studies that accrued more than 5 eyes to each study arm. Thefollowing exclusion criteria were used in the selection of includedstudies: (1) nonpublished articles (e.g., abstracts and conferenceproceedings); (2) articles not published in English; (3) articles withrepeat data; (4) case reports or small (�5 eyes per study arm) caseseries; and (5) literature reviews, letters to the editor, correspon-dence, notes, editorials, and forthcoming journal articles. Giventhat existing studies in the published literature were used for thismeta-analysis, institutional review board approval was not neces-sary. Nonetheless, the study adhered fully to the Declaration ofHelsinki.

Study Selection, Data Collection, and OutcomeMeasures

Two authors (M.P. and X.C.-M.) examined search results to selectpertinent articles for inclusion, first by title and abstract screeningand then by screening full text articles. Uncertainty in inclusionwas resolved through consultation with a third author (M.B.S.).The same 2 authors (M.P. and X.C.-M.) extracted the followingbaseline demographic and clinical data from each study arm: studydesign, country of origin, femtosecond laser type, date of inter-vention, number of included eyes, mean cohort age, gender dis-tribution, mean corrected distance visual acuity (CDVA), and meanaxial length. In addition, a comprehensive list of intraoperative andpostoperative outcomes were extracted from included studies andwere reported using the following headings:

1. Primary visual and refractive outcomes: uncorrected dis-tance visual acuity (UDVA), CDVA, mean absolute error(MAE) of manifest refraction spherical equivalent.

2. Secondary surgical end points, effective phacoemulsifica-tion time, surgery time, balanced salt solution volume,cumulative dissipated energy (CDE), circularity of capsu-lotomy or capsulorrhexis, capsule opening diameter,absolute mean deviation from intended capsule diameter,intraocular lens (IOL) horizontal and vertical decentration,central corneal thickness, corneal endothelial cell count andpreoperative to postoperative reduction, total prostaglandinconcentration, and mean aqueous flare.

3. Safety parameters: overall complications, capsular com-plications, corneal complications, and pupillarycomplications.

In the extraction of data, continuous variables were recorded asmeans � standard deviations, whereas categorical variables werereported as percentages of the total sample. If any included studyprovided acceptable measures of variation that could be convertedto a standard deviation (e.g., standard error), these data also wereextracted. To facilitate the meta-analysis design, complicationswere grouped by anatomic site (Table 1, available atwww.aaojournal.org). Data for all postoperative outcomes werecollected at last follow-up. To ensure balance in the averagelength of follow-up between comparators, outcomes were extractedfrom each included study at the same follow-up period for bothFLACS and MCS eyes. If outcome data were repeated in 2 or more

Popovic et al � Meta-Analysis of FLACS

studies, then data from only 1 study were incorporated into themeta-analysis. Microsoft Excel (Microsoft Corporation, Redmond,WA) was used to manage all identified records and to compileextracted data.

Risk of Bias Assessment

To perform an assessment of study quality, 2 authors (M.P. andX.C.-M.) independently completed the Newcastle-Ottawa QualityAssessment Scale (NOS) for included observational studies andused the guidelines set by the Cochrane Collaboration for ran-domized controlled trials (Appendix 2A-B, available atwww.aaojournal.org).12,13 The NOS is an 8-item scale that eval-uates study quality based on 3 criteria: patient selection, compa-rability between treatment arms, and outcomes. To differentiatebetween high and low risk of bias on the follow-up item of theNOS, a threshold of 3 weeks of follow-up was set for all outcomesexcept intraoperative efficacy and safety parameters. In addition, aconservative estimate of 10% was used for the maximum accept-able loss to follow-up. For randomized controlled trials, 7 aspectsof quality assessment were performed: sequence generation, allo-cation concealment, blinding of participants, personnel, andoutcome assessors, management of incomplete outcome data,completeness of outcome reporting, and other potential threats tovalidity. Studies were excluded if they had a high or unclear risk ofbias in all assessment categories. We also evaluated the rate ofauthorship conflicts of interest and reported funding from industrysponsors. Further, we performed a qualitative synthesis on baselinefactors that may have impacted refractive outcomes significantly.

Data Synthesis and Analysis

Weighted mean differences were reported for continuous variableswith accompanying 95% confidence intervals. For dichotomousvariables, risk ratios and 95% confidence intervals were computed.Using a random effects model in all cases, the inverse variancemethod was used for continuous data and the Mantel-Haenszelapproach was used for dichotomous outcomes. The weightedmean was defined as

X ¼Pn

i¼1wixiPni¼1wi

whereas the weighted standard deviation was represented by

sdw ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiPN

i¼1wiðxi � xwÞ2ðN’�1Þ

PN

i¼1wi

N’

vuut :

In testing for an overall effect, the number of eyes was used as aweighting variable and comparisons that had a P value of less than0.05 were deemed statistically significant. Statistical heterogeneitywas assessed by computing a chi-square statistic; for this test, onlya result of P < 0.05 was considered heterogeneous and was re-ported in the text. Additionally, an I2 measure was computed toinvestigate the percentage of variance in the meta-analysis that maybe attributed to heterogeneity. Sources of clinical and methodo-logic heterogeneity across the included studies also were examinedand described qualitatively. Meta-analysis was performed for anoutcome only if there were appropriate data (i.e., percentage forcategorical outcome, mean � standard deviation or standard errorfor continuous variable) for at least 2 study arms in eachcomparator. After study selection, meta-analysis was performedregardless of study design. One could argue that only randomizedcontrolled trials should be included because of the high internalvalidity of this study design. However, most randomized controlledtrials in the FLACS literature are unmasked and funded by in-dustry, and as such show many of the same biases as observational

cohort studies. Furthermore, the significantly increased statisticalpower achieved by including observational studies in the meta-analysis may outweigh the bias from confounding, especially forsafety outcomes, which are underpowered frequently in random-ized controlled trials. Given that a previous report found similareffect sizes based on meta-analyses that included observationalstudies when compared with meta-analyses of randomizedcontrolled trials, this meta-analysis included both randomizedcontrolled trials and observational cohort studies.14 ReviewManager version 5.3 (The Nordic Cochrane Centre, TheCochrane Collaboration, Copenhagen, Denmark) was used for allstatistical analyses. No protocol amendments were made for thisstudy.

Results

Study Inclusions and Demographics

Two thousand eight hundred and two records underwent title andabstract screening. After 2716 exclusions, 86 full texts werescreened. Overall, 37 articles were included in the meta-analysiswith 7127 eyes undergoing FLACS and 7440 eyes undergoingMCS (Fig 1; Table 2).2,15e50 Mean baseline age ranged from 58.5to 75 years in the FLACS cohort and 56.5 to 74.3 years in the MCScohort (n ¼ 25 studies). In the 17 studies reporting on baselinegender distribution, 1890 of 2901 eyes (65.1%) were those ofwomen in the FLACS cohort and 1694 of 3099 eyes (54.7%) werethose of women in the MCS cohort. Fourteen studies reported onbaseline cohort axial length. Across these articles, mean axiallength ranged from 23.33 to 25.09 mm in the FLACS cohort andfrom 23.08 to 26.94 mm in the MCS cohort. A complete list ofbaseline demographic and clinical information is provided onTable 3 (available at www.aaojournal.org).

Quality Assessment

Most of the included eyes (12 967 eyes [89.0%]) came fromobservational studies (22 studies [59.5%]). There were authorshipconflicts of interest in 11 of 22 (50.0%) observational studies andin 12 of 15 (80.0%) included randomized controlled trials. Therewas direct funding from industry sponsors in 1 of 22 (4.5%)observational studies and in 2 of 15 (13.3%) randomized controlledtrials (Table 3, available at www.aaojournal.org). Assessment ofstudy quality revealed that no included studies had a high orunclear risk of bias in all assessment categories (Table 4A, B,available at www.aaojournal.org). However, only 2 of 22 (9.1%)observational studies and 1 of 15 (6.7%) randomized controlledtrials had a low risk of bias in all categories.25,27,49 For random-ized controlled trials, omissions in the description of randomization(7/15 [46.7%]), allocation concealment (11/15 [73.3%]), blindingof participants and personnel (10/15 [66.7%]), and blinding ofoutcome assessment (12/15 [80.0%]) may have introduced biasinto the findings. Analysis of observational studies using the NOSrevealed that all articles were of either medium or high qualitybecause no study was awarded fewer than 5 stars (range, 5e9stars). Here, omissions in comparator comparability, outcomeassessment, and completeness of follow-up may have introducedbias in terms of patient selection, group comparability, andoutcome assessment. For instance, only 5 of 22 (22.7%) observa-tional studies attempted to demonstrate comparability of cohorts atbaseline and then attempted to adjust for confounding variables(Table 4B, available at www.aaojournal.org). In general, theseadjustments were limited by low sample sizes and a lack ofimportant baseline characteristics that are known to influencevisual and refractive outcomes.

2115

Figure 1. Flow diagram illustrating the course of Meta-Analyses and Systematic Reviews of Observational Studies (MOOSE) guidelines.

Ophthalmology Volume 123, Number 10, October 2016

It was also found that certain efficacy end points like CDVA,effective phacoemulsification time, and CDE showed considerablebetween-study heterogeneity, thus limiting the interpretability ofthe overall effect. There were many causes of heterogeneity,

Table 2. Baseline Clinical and

Baseline Parameter

Type of FLACS procedureCCIs, fragmentation, capsulotomy15,20,21,24,26,31,35,49,50

Only CCIs, capsulotomy2,16,34,41

Only capsulotomy, fragmentation23,25,27,28,30,33,37,38,40,43,45e48

Only capsulotomy17e19,22,36,44

Type of FLACS machineLenSx (Alcon Inc, Hünenberg, Switzerland)2,16,19e22,24,26,31,34,41,49,50

LENSAR (LENSAR, Inc, Orlando, FL)28,32,41,44,48

Catalys (Abbott Laboratories, Inc, Abbott Park, IL)15,17,18,29,30,33,35,36,38e40,4

Victus (Bausch & Lomb, Inc, Bridgewater, NJ)23,25,37,43

Type of phacoemulsification machineAccurus (Alcon, Inc)16,19,22,34

Constellation (Alcon, Inc)31,41

Infiniti (Alcon, Inc)2,20,21,24,26,28,32,49,50

Legacy (Alcon, Inc)17

Megatron (Geuder Group, Heidelberg, Germany)15,18,40,42,47

Stellaris (Bausch & Lomb, Inc)27,30,33,36e38,43,46,48

Last follow-up (mos)<137,40,49,50

1e32,16e18,20,21,23,25e28,33,35,48

>319,22,31,32,34,36,38,41,42,46,47

CCI ¼ clear corneal incision; FLACS ¼ femtosecond laser-assisted cataract su

2116

including variability in the steps of FLACS, differences in FLACSand MCS platforms, surgical technique and equipment, and in-consistencies in measurement techniques, length of follow-up,patient populations, and study design (Table 2).

Demographic Information

No. of Eyes (%) No. of Studies (%)

6390 (100.0) 33 (100.0)2415 (37.8) 9 (27.3)357 (5.6) 4 (12.1)

3273 (51.2) 14 (42.4)345 (5.4) 6 (18.2)

7054 (100.0) 37 (100.0)929 (13.2) 13 (35.1)321 (4.6) 5 (13.5)

2,45e47 4458 (63.2) 15 (40.5)1346 (19.1) 4 (10.8)6614 (100.0) 30 (100.0)153 (2.3) 4 (13.3)60 (0.9) 2 (6.6)

1953 (29.5) 9 (30.0)24 (0.4) 1 (3.3)

3458 (52.3) 5 (16.7)966 (14.6) 9 (30.0)

8185 (100.0) 29 (100.0)408 (5.0) 4 (13.8)

4115 (50.3) 14 (48.3)3662 (44.7) 11 (37.9)

rgery.

Popovic et al � Meta-Analysis of FLACS

Qualitative synthesis revealed that a variety of potential con-founding factors may have impacted the results. The followingfactors were not always measured, reported, compared, or adjustedfor: clinical parameters (e.g., preoperative vision, refraction, cata-ract density), biometry measurements (e.g., differences in axiallength, keratometry, anterior chamber depth, lens thickness), andsurgical planning (e.g., type of IOL, approach to astigmatismcorrection).

Visual and Refractive Outcomes

To perform a meta-analysis of visual acuity data from varyingstudies, the formulae provided by Khoshnood et al51 were used toconvert from decimal to logarithm of minimum angle of resolutionvisual acuity. There was no statistically significant differencebetween FLACS and MCS in terms of postoperative UDVA(P ¼ 0.19; heterogeneity, P ¼ 0.004; I2 ¼ 69%; Fig 2A) orCDVA (P ¼ 0.26; heterogeneity, P ¼ 0.004; I2 ¼ 66%; Fig 2B).The postoperative MAE also was nonsignificantly differentbetween comparator arms (P ¼ 0.57; heterogeneity, P < 0.001;I2 ¼ 75%; Fig 2C).

Procedure Time and Energy

On average, meta-analysis revealed that effective phacoemulsi-fication time was more than 3 seconds longer for MCS eyescompared with eyes undergoing FLACS (P < 0.001; Fig 2D);however, total surgery time was nonsignificantly differentbetween comparators (P ¼ 0.07; Fig 2E). Both analysesshowed considerable statistical heterogeneity (P < 0.001; I2 ¼98% and 97%, respectively). There were no significantdifferences between comparators in terms of balanced saltsolution volume (P ¼ 0.71; Fig 2F) and total CDE (P ¼ 0.21;Fig 2G).

Capsulotomy and Capsulorrhexis Parameters

Depending on the study, circularity of the removed capsule wasmeasured in 1 of 2 ways: first, as the normalized ratio of the area ofthe capsule to the area of a hypothetical disc with a diameter equal tothe greatest linear dimension of the capsule,17,35 and second, byusing the following formula: circularity ¼ 4p(area/perim-eter2).34,41,45 In both cases, the ratio is equal to 1 for an ideal circle.Studies using the first formula found that FLACS extracted asignificantly more circular capsule by 0.16 units (P < 0.001;Fig 2H). However, this result was not detected in studies using thesecond formula because there was no significant differencebetween comparators (P ¼ 0.10; heterogeneity, P < 0.001; I2 ¼99%; Fig 2I).

There was no statistically significant difference in terms ofcapsule opening diameter between FLACS and MCS study arms(P ¼ 0.40; Fig 2J); however, this end point also exhibitedsignificant heterogeneity (P < 0.001; I2 ¼ 80%). Conversely,analysis of absolute mean deviation from intended diameterrevealed that FLACS produced capsulotomies that weresignificantly closer to the intended diameter (P ¼ 0.007; Fig2K). Again, this end point showed heterogeneity (P < 0.001;I2 ¼ 93%).

Intraocular lens decentration was calculated by using the dis-tance between the pupillary axis and the IOL center.19,34 Thefindings were mixed when decentration parameters were consid-ered: FLACS produced significantly more horizontally centeredIOLs by an average of 128.84 mm (P < 0.001; Fig 2L); however,vertical decentration was nonsignificantly different betweencomparators (P ¼ 0.90; Fig 2M).

Central Corneal Thickness and Endothelial CellCount Reduction

There was a significant difference in favor of FLACS over MCS inaverage central corneal thickness: after surgery, FLACS corneaswere thinner by an average of 6.37 mm (P ¼ 0.02; Fig 2N). Forendothelial cell count, no significant difference was found in thenumber of postoperative cells per square millimeter for FLACSeyes relative to MCS eyes (P ¼ 0.07; Fig 2O). At the sametime, there was a significantly greater reduction of 55.43endothelial cells/mm2 for the MCS comparator before versusafter surgery (P ¼ 0.006; Fig 2P).

Prostaglandin Concentration and Mean AqueousFlare

Across 2 studies by the same research team, total prostaglandinconcentration was greater for eyes receiving FLACS relative toMCS (P < 0.001; Fig 2Q). In addition, between-study heteroge-neity was noted for this outcome (P < 0.001; I2 ¼ 96%). For meanaqueous flare, meta-analysis revealed a nonsignificant differencebetween comparators (P ¼ 0.28; Fig 2R).

Safety Analysis

Analysis of the overall incidence of complications showed thatthere was no statistically significant difference between compara-tors (P ¼ 0.16; Fig 3A); however, there was significantheterogeneity between studies (P < 0.001; I2 ¼ 95%). The sameresult was maintained for other end points: capsularcomplications except for posterior capsular tears (P ¼ 0.14;heterogeneity, P < 0.001; I2 ¼ 88%; Fig 3B), cornealcomplications (P ¼ 0.27; heterogeneity, P < 0.001; I2 ¼ 85%;Fig 3D), and pupillary complications (P ¼ 0.10; heterogeneity,P ¼ 0.006; I2 ¼ 81%; Fig 3E). Eyes that underwent MCS had asignificantly lower incidence of posterior capsular tears whencompared with those that underwent FLACS (P ¼ 0.005; Fig 3C).

Discussion

The purported efficacy and safety benefits of FLACS rela-tive to MCS are based on its ability to produce more ac-curate, reproducible capsulotomies and clear cornealincisions, as well as to reduce the ultrasound energy andintraocular manipulation required for lens fragmentation andremoval.52e57

From an efficacy standpoint, we were unable to detect adifference between MCS and FLACS for UDVA, CDVA,and MAE. In reviewing the literature, visual and refractiveoutcomes are the most patient-important end points from aclinical perspective. However, we do note that our analysiswas limited to the outcomes that were reported. For futureresearch, an editorial by Hoffer et al58 reminds authors tozero the mean arithmetic error for their study populations,to compare median not mean absolute errors, to reportcategorical outcomes of patients within reasonablerefractive targets, to report manifest refractions only forpatients with vision of 20/40 or better, to account forcorrelation between eyes, and to report the instrumentsused to obtain various study measurements. Furthermore,we welcome novel studies addressing other measures ofvisual quality, including higher-order aberrations, contrastsensitivity, and dysphotopsia. For instance, Mihaltz et al22

2117

Figure 2. Meta-analysis of the efficacy of femtosecond laser-assisted cataract surgery compared with manual cataract surgery. A, Uncorrected distance visualacuity (logarithm of the minimum angle of resolution [logMAR]). B, Corrected distance visual acuity (logMAR). C, Mean absolute error of manifestrefraction spherical equivalent. D, Effective phacoemulsification time. E, Total surgery time. F, Balanced salt solution volume. G, Cumulative dissipatedenergy. H, Circularity using formula 1. I, Circularity using formula 2. J, Capsule opening diameter. K, Absolute mean deviation from intended capsulediameter. L, Horizontal decentration. M, Vertical decentration. N, Central corneal thickness. O, Corneal endothelial cell count. P, Corneal endothelialcell loss. Q, Total prostaglandin concentration. R, Mean aqueous flare. CI ¼ confidence interval; FLACS ¼ femtosecond laser-assisted cataract surgery;IV ¼ inverse variance; MCS ¼ manual cataract surgery; Random ¼ random effects model; SD ¼ standard deviation.

Ophthalmology Volume 123, Number 10, October 2016

found some evidence that the Strehl ratio and modulationtransfer function were higher in eyes undergoing FLACSrelative to MCS.

Mixed results were found when examining the ability ofFLACS to produce more circular capsulotomies. Given the 2definitions that currently exist, it is recommended that futurestudies standardize the definition of circularity to facilitatebetter interstudy comparison. When comparing mean devia-tion from the intended diameter, FLACS produced capsu-lotomies that were significantly closer to the intendeddiameter relative to MCS. This difference may be attributedto variability in surgical technique and the effect of cornealmagnification when performing manual capsulorrhexis,

2118

which does not apply to the FLACS capsulotomy.59 Further,FLACS capsulotomies may not be uniform between patients;Dick et al60 demonstrated an age-dependent variability incapsulotomy size with pediatric FLACS procedures.

Our meta-analysis also showed that IOLs implanted inFLACS patients had significantly better horizontal centra-tion, presumably because of a more centered and circularcapsulotomy. Previous authors have asserted that FLACSdelivers greater accuracy and precision of the capsulotomy,which may make for a more predictable effective lens posi-tion, thus improving visual and refractive outcomes.53

Theoretically, capsule overlap throughout the entirecircumference of the IOL optic may prevent pea-podding,

Figure 2. (continued).

Popovic et al � Meta-Analysis of FLACS

reducing the risk of myopic shift or astigmatism resultingfrom anterior optic displacement and tilt.61e63 AlthoughReddy et al37 found that capsulorrhexis centration wasimproved significantly for FLACS eyes relative to MCSeyes, so far it is only possible to center the capsulotomy on

the pupil. Ideally, the visual axis should be used forcentration, which can be performed with MCS by usingcorneal markers that take advantage of Purkinje images tocenter the capsulorrhexis on the visual axis. In addition, thenotion of circularity and centration materially affecting

2119

Figure 2. (continued).

Ophthalmology Volume 123, Number 10, October 2016

refractive outcomes is still being debated. For instance, in anobservational study, Okada et al64 did not detect a significantcorrelation between either circularity or centration with targetspherical equivalent and cylinder 1 month and 1 year aftersurgery. Nonsignificant findings in refractive outcomesmay be attributable to numerous sources of errorin refractive predictability, including preoperativemeasurement, choice of IOL formula, and methods usedfor prediction error assessment, which could hide any true

2120

difference resulting from a consistent capsulotomy.Furthermore, complete capsuleeoptic overlap also hasbeen shown to reduce the rate of posterior capsularopacification and dysphotopsia.62,65,66 However, only 1published study has investigated differences in the rate ofposterior capsular opacification between FLACS andMCS.48 We encourage future research in this area.

Although average surgical time was nonsignificantlydifferent between comparators, FLACS produced a shorter

Figure 3. Meta-analysis of the safety of femtosecond laser-assisted cataract surgery compared with manual cataract surgery. A, Overall incidence ofcomplications. B, Incidence of capsular complications excluding posterior capsular tears. C, Incidence of posterior capsular tears. D, Incidence of cornealcomplications. E, Incidence of pupillary complications. CI ¼ confidence interval; FLACS ¼ femtosecond laser-assisted cataract surgery; IV ¼ inversevariance; M-H¼Mantel-Haenszel approach; MCS¼manual cataract surgery; Random¼ random effects model; SD ¼ standard deviation. *For the Chen etal article, the outlier data from surgeon 5 was excluded.

Popovic et al � Meta-Analysis of FLACS

effective phacoemulsification time, which signifies a lowertotal amount of energy delivered to the eye.33 Betweenstudies, differences in surgical equipment, surgeon skill,patient selection, and definitions of terms may have led tosignificant heterogeneity in the total procedure time andeffective phacoemulsification time outcomes. At the sametime, the pooled treatment effect of studies reporting onCDE showed no significant difference between FLACS andMCS (Fig 2G). The contradictory results between effectivephacoemulsification time and CDE could be explained bymethodologic variation in the included studies, bydifferences in surgical techniques, or by the differentialnumber of included studies, which may have introducedreporting bias into the findings. Another confounding factoris between-study variability in energy delivery parameters,which affects the calculation for CDE differently. Specif-ically, the formula for calculating CDE assigns only 40% ofthe actual torsional effective phacoemulsification time to thesum, whereas the effective phacoemulsification time forlongitudinal ultrasound remains the same.67

The meta-analysis demonstrated that there was a statis-tically significant difference in favor of FLACS over MCS insurgery-induced corneal endothelial cell loss. However, it isuncertain whether this mean difference of 55.43 cells/mm2

carries any clinical significance. For instance, this differencemay be attributed to instrument bias, given the poorrepeatability of specular microscopy. Notwithstanding, cor-neas were significantly thinner in the early postoperativeperiod in patients undergoing FLACS, which may suggestless surgically-induced corneal stress. Studies evaluatingcorneal endothelial changes after femtosecond LASIK sug-gest that this procedure is safe for the endothelium.68e70 Incontrast, a study by Abell et al42 found that eyes undergoingFLACS with laser-automated corneal incisions had a greaterendothelial cell loss at 6 months than eyes undergoingFLACS, but with manual corneal incisions (P< 0.001); eyeswith 0 effective phacoemulsification time (EPT) and manualincisions had the least endothelial cell loss (P < 0.001).Differences in findings between studies comparing femto-second LASIK and FLACS may be explained by the deeper

2121

Figure 3. (continued).

Ophthalmology Volume 123, Number 10, October 2016

energy penetration at the level of the endothelium whenperforming clear corneal incisions compared with LASIKflaps.71

The safety analysis showed no significant differences be-tween FLACS and MCS for overall, capsular, pupillary, andcorneal complications. However, there was a significantlygreater incidence of posterior capsular tears after FLACSrelative toMCS.Given thatmany of the included studies werepublished early after the introduction of FLACS, the surgeonlearning curve may have influenced these results. Nonethe-less, posterior capsular tears are associated with complica-tions like retinal detachment, endophthalmitis, and cystoidmacular edema. Beyond the evaluated outcomes, there arealso complications unique to femtosecond lasers that havebeen noted in past case reports, including interface cornealstromal irregularities, corneal perforation, and incompletelaser-assisted capsulotomy and fragmentation resulting fromsilicone oil in the anterior chamber.72e75 Based on the liter-ature to date, the overall safety profile of FLACS is unfa-vorable relative to MCS. However, this result may change asthe femtosecond laser technology and surgeon skill withFLACS continues to evolve.

Femtosecond laser-assisted cataract surgery was shownto be associated with a significantly greater concentration ofintraocular prostaglandins relative to MCS (P < 0.001).29,39

Schultz et al39 note that, despite the significantconcentrations of prostaglandins synthesized by the irisand ciliary body, the specific location of prostaglandinrelease currently is uncertain. They hypothesize that themicroplasma of gas and water generated by the focusedFLACS laser spot may trigger the release ofprostaglandins. Prostaglandins have been shown to beassociated with inflammation-induced miosis and may bea causative factor in the development of cystoid macularedema and uveitis after cataract surgery.39

Comparing the present analysis with the only previousmeta-analysis in the published literature, we found that theprevious study screened only 297 articles (present meta-

2122

analysis, 2802 articles).11 Our broader screening strategyled to the inclusion of 6 additional randomized controlledtrials. These randomized controlled trials, in addition tothe inclusion of 22 observational studies, led to anincrease in sample size from 989 to 14 567. Our designalso included additional outcomes not reported previously:mean absolute error, total surgery time, balanced saltsolution volume, capsule opening diameter, absolute meandeviation from intended diameter, decentration, totalprostaglandin concentration, mean aqueous flare, andvarious complication parameters.

Beyond efficacy and safety, the cost effectiveness ofFLACS is an important consideration that is not addressedin the present meta-analysis. In their comparative cost-effectiveness analysis of FLACS and MCS, Abell andVote76 reviewed complication rates in the literature. In theirhypothetical cohort of patients between 6 months to 1 yearafter cataract surgery, there was a quality-adjusted life-yeargain of 0.06 units for FLACS compared with MCS. How-ever, this equated to a cost per quality-adjusted life-year of$102 691, which was not viewed as cost effective.

Recently, the updated European Registry of QualityOutcomes for Cataract and Refractive Surgery (EUREQUO)case control data concerning the efficacy and safety ofFLACS was presented (Barry P. ESCRS femto laser assistedcataract surgery (FLACS): case control study. Paper pre-sented at: 33rd Congress of the ESCRS, September 5e9,2015; Barcelona). Although the study remains unpublished,it is notable for reporting on a large number of eyes (2814vs. 4987 eyes after FLACS and MCS, respectively) derivedfrom a registry-based database. Their findings corroboratethe results of the current meta-analysis: higher complicationrates (3.4% vs. 2.3%) and a similar proportion of eyes withan improvement in CDVA (86.0% vs. 89.3%) after FLACSrelative to MCS. We look forward to the publication of theirresults.

This meta-analysis is the most comprehensive review ofthe published literature investigating the efficacy and safety

Popovic et al � Meta-Analysis of FLACS

of FLACS relative to MCS. As such, the analysis benefitsfrom a large sample size (n ¼ 14 567) and a high number ofpublished studies (n ¼ 37). Further, all included studiescontained an MCS arm; based on descriptive statistics,preoperative parameters were comparable between theFLACS and MCS cohorts.

Despite these advantages, the study is subject to certainlimitations. In terms of limitations related to the study de-signs of extracted articles, many observational studies wereincluded (22 of 37 studies [59.5%]; 12 967 of 14 567 eyes[89.0%]). This may have introduced confounding by indi-cation, information bias, and selection bias into the findings.Adjusted effect estimates could not be extracted because ofthe limitations in the reporting of individual observationalstudies, which may have introduced confounding. However,one may hypothesize that the combination of confoundingand information bias would skew the results in favor ofFLACS, a finding that largely was not present in this meta-analysis. We found that 15 of 37 (40.5%) included studiesdisclosed that at least some patients contributed 2 eyes to theindividual analysis. Of these 15, only 8 (53.3%) noted thatpatients who contributed 2 eyes were treated independently.The other 7 (46.7%) studies should have taken into accountthat there is some within-patient correlation between eyes ofthe same patient. For our visual outcome analysis, we re-ported data as they were analyzed in the existing literature,understanding the limitations of this parametric analysis.77

We encourage future studies to provide broader detail ontheir visual outcomes and to consider using clinicallysignificant cutoffs in their reporting.58 In terms oflimitations related to the selection of included studies, thismeta-analysis considered only published data to ensurethat the rigors of peer review were met for each includedarticle. We recognize that a consequence of this approach isthe potential for publication bias (i.e., not including un-published negative studies). Also, there might have beenlanguage bias because the meta-analysis considered onlystudies published in English. Finally, in terms of limitationsrelated to variability in clinical reporting, it was difficult tocontrol for different technologies and surgeon experiencebecause of a large between-study variance and lack ofreporting of these parameters. There was significant vari-ability in the duration of reported follow-up in the includedstudies; as such, we used the last available follow-up foranalysis of postoperative outcomes. This decision wassupported by the work of Conrad-Hengerer et al,46 whoshowed that there was no significant change in the meanrefractive spherical equivalent between 1 week and 1month after FLACS and between 1, 2, 3, and 6 monthsafter either FLACS or MCS. Our results also showed thatthere was significant heterogeneity across numerousoutcomes and that the effect sizes of certain end points(e.g., corneal thickness) were smaller than the knownvariability in the accuracy of measurement.

In summary, this meta-analysis found that there were nosignificant differences between FLACS and MCS in termsof key postoperative visual and refractive outcomes, spe-cifically UDVA, CDVA, and MAE. There were mixed re-sults regarding secondary surgical end points, in which anonsignificant difference between comparators was found

for certain parameters like CDE, capsule opening diameter,and vertical IOL centration, whereas there was a statisticallysignificant difference in favor of FLACS for effectivephacoemulsification time, absolute mean deviation fromintended capsule diameter, horizontal IOL centration, andpostoperative central corneal thickness. There was a sig-nificant difference in favor of MCS in terms of prosta-glandin concentration. Safety analysis revealed that FLACSand MCS were nonsignificantly different in the incidence ofoverall, capsular, corneal, and pupillary complications;however, there was a significant difference in favor of MCSover FLACS in the incidence of posterior capsular tears. Ingeneral, it is important to consider the clinical significanceof the measured differences when interpreting thesefindings.

There may be certain clinical scenarios, such as cases inwhich a manual capsulorrhexis is harder to perform (e.g.,subluxated lens), in which FLACS may have specific ad-vantages.78 Furthermore, there may be applications andmodifications of the IOL technology in the future thatmay favor FLACS over MCS. Because of the continualevolution of the femtosecond laser technology, it is likelythat there will be continued head-to-head comparisons be-tween these 2 techniques. We await this evidence andrecommend that a subsequent re-evaluation be performedafter a significant number of well-designed randomized tri-als are introduced into the literature. Through this process,the authors hope a more definitive conclusion can bereached regarding the role of femtosecond lasers in cataractsurgery.

Acknowledgments. The authors thank Austin Pereira, MichelleEfrosman, and Thomas Berk for their collaboration.

References

1. Foster A. Vision 2020: the cataract challenge. Community EyeHealth 2000;13:17–9.

2. Bali SJ, Hodge C, Lawless M, et al. Early experience with thefemtosecond laser for cataract surgery. Ophthalmology2012;119:891–9.

3. Cullen KA, Hall MJ, Golosinskiy A. Ambulatory Surgery inthe United States, 2006. Hyattsville, MD: National Center forHealth Statistics; 2009:11.

4. Bell CM, Hatch WV, Cernat G, Urbach DR. Surgeon volumesand selected patient outcomes in cataract surgery. Ophthal-mology 2007;114:405–10.

5. Soong HK, Malta JB. Femtosecond lasers in ophthalmology.Am J Ophthalmol 2009;147:189–97.

6. He L, Sheehy K, Culbertson W. Femtosecond laser-assistedcataract surgery. Curr Opin Ophthalmol 2011;22:43–52.

7. Baig NB, Cheng GPM, Lam JKM, et al. Intraocular pressureprofiles during femtosecond laser-assisted cataract surgery.J Cataract Refract Surg 2014;40:1784–9.

8. Chen S. Advances in cataract surgery. Available at: http://www.optometry.org.au/media/585285/1.9_sat_advances_in_cataract_surgery_chen.pdf. Accessed July 19, 2015.

9. Berdahl JP, Jensen MP. The business of refractive laserassisted cataract surgery (ReLACS). Curr Opin Ophthalmol2014;25:62–70.

10. Quinones A, Gleitsmann K, Freeman M, et al. Benefits andharms of femtosecond laser assisted cataract surgery: a

2123

Ophthalmology Volume 123, Number 10, October 2016

systematic review. Portland, OR: Evidence-Based SynthesisProgram Center; 2013:1–48. VA-ESP Project no. 05e225.

11. Chen X, Xiao W, Ye S, et al. Efficacy and safety of femto-second laser-assisted cataract surgery versus conventionalphacoemulsification for cataract: a meta-analysis of random-ized controlled trials. Sci Rep 2015;5:13123.

12. Wells G, Shea B, O’Connell D, et al. The Newcastle-OttawaScale (NOS) for assessing the quality of nonrandomised studiesin meta-analyses. Available at: http://www.ohri.ca/programs/clinical_epidemiology/oxford.asp. Accessed July 9, 2015.

13. Higgins JP, Green S, eds. Cochrane Handbook for SystematicReviews of Interventions. Version 5.1.0. London: CochraneCollaboration; 2011.

14. Shrier I, Boivin JF, Steele RJ, et al. Should meta-analyses ofinterventions include observational studies in addition to ran-domized controlled trials? A critical examination of underlyingprinciples. Am J Epidemiol 2007;166:1203–9.

15. Abell RG, Darian-Smith E, Kan JB, et al. Femtosecond laser-assisted cataract surgery versus standard phacoemulsificationcataract surgery: outcomes and safety in more than 4000cases at a single center. J Cataract Refract Surg 2015;41:47–52.

16. Filkorn T, Kovacs I, Takacs A, et al. Comparison of IOLpower calculation and refractive outcome after laser refractivecataract surgery with a femtosecond laser versus conventionalphacoemulsification. J Refract Surg 2012;28:540–4.

17. Friedman NJ, Palanker DV, Schuele G, et al. Femtosecondlaser capsulotomy. J Cataract Refract Surg 2011;37:1189–98.

18. Abell RG, Allen PL, Vote BJ. Anterior chamber flare afterfemtosecond laser-assisted cataract surgery. J Cataract RefractSurg 2013;39:1321–6.

19. Kranitz K, Mihaltz K, Sandor GL, et al. Intraocular lens tiltand decentration measured by Scheimpflug camera followingmanual or femtosecond laser-created continuous circular cap-sulotomy. J Refract Surg 2012;28:259–63.

20. Lawless M, Bali SJ, Hodge C, et al. Outcomes of femtosecondlaser cataract surgery with a diffractive multifocal intraocularlens. J Refract Surg 2012;28:859–64.

21. Mayer WJ, Klaproth OK, Hengerer FH, et al. Impact ofcrystalline lens opacification on effective phacoemulsificationtime in femtosecond laser-assisted cataract surgery. Am JOphthalmol 2014;157:426–32.

22. Mihaltz K, Knorz MC, Alio JL, et al. Internal aberrations andoptical quality after femtosecond laser anterior capsulotomy incataract surgery. J Refract Surg 2011;27:711–6.

23. Khandekar R, Behrens A, Towerki AEA, et al. Determinantsof visual outcomes in femtosecond laser assisted cataractsurgery and phacoemulsification: a nested case control study.Middle East Afr J Ophthalmol 2015;22:356–61.

24. Chen M, Swinney C, Chen M. Comparing the intraoperativecomplication rate of femtosecond laser-assisted cataract sur-gery to traditional phacoemulsification. Int J Ophthalmol2015;8:201–3.

25. Chee SP, Yang Y, Ti SE. Clinical outcomes in the first twoyears of femtosecond laser-assisted cataract surgery. Am JOphthalmol 2015;159:714–9.

26. Takacs AI, Kovacs I, Mihaltz K, et al. Central corneal volumeand endothelial cell count following femtosecond laser-assisted refractive cataract surgery compared to conventionalphacoemulsification. J Refract Surg 2012;28:387–91.

27. Conrad-Hengerer I, Al Juburi M, Schultz T, et al. Cornealendothelial cell loss and corneal thickness in conventionalcompared with femtosecond laser-assisted cataract surgery:three-month follow-up. J Cataract Refract Surg 2013;39:1307–13.

2124

28. Krarup T, Holm LM, la Cour M, Kjaerbo H. Endothelial cellloss and refractive predictability in femtosecond laser-assistedcataract surgery compared with conventional cataract surgery.Acta Ophthalmol 2014;92:617–22.

29. Schultz T, Joachim SC, Stellbogen M, Dick HB. Prostaglandinrelease during femtosecond laser-assisted cataract surgery:main inducer. J Refract Surg 2015;31:78–81.

30. Conrad-Hengerer I, Schultz T, Jones JJ, et al. Cortex removalafter laser cataract surgery and standard phacoemulsification: acritical analysis of 800 consecutive cases. J Refract Surg2014;30:516–20.

31. Mastropasqua L, Toto L, Mastropasqua A, et al. Femtosecondlaser versus manual clear corneal incision in cataract surgery.J Refract Surg 2014;30:27–33.

32. Chang JSM, Chen IN, Chan WM, et al. Initial evaluation of afemtosecond laser system in cataract surgery. J CataractRefract Surg 2014;40:29–36.

33. Conrad-Hengerer I, Hengerer FH, Schultz T, Dick HB. Effect offemtosecond laser fragmentation on effective phacoemulsifica-tion time in cataract surgery. J Refract Surg 2012;28:879–83.

34. Kranitz K, Takacs A, Mihaltz K, et al. Femtosecond lasercapsulotomy and manual continuous curvilinear capsulorrhexisparameters and their effects on intraocular lens centration.J Refract Surg 2011;27:558–63.

35. Palanker DV, Blumenkranz MS, Andersen D, et al. Femto-second laser-assisted cataract surgery with integrated opticalcoherence tomography. Sci Transl Med 2010;2:58ra85.

36. Conrad-Hengerer I, Hengerer FH, Al Juburi M, et al. Femto-second laser-induced macular changes and anterior segmentinflammation in cataract surgery. J Refract Surg 2014;30:222–6.

37. Reddy KP, Kandulla J, Auffarth GU. Effectiveness and safetyof femtosecond laser-assisted lens fragmentation and anteriorcapsulotomy versus the manual technique in cataract surgery.J Cataract Refract Surg 2013;39:1297–306.

38. Schargus M, Suckert N, Schultz T, et al. Femtosecond laser-assisted cataract surgery without OVD: a prospective intra-individual comparison. J Refract Surg 2015;31:146–52.

39. Schultz T, Joachim SC, Kuehn M, Dick HB. Changes inprostaglandin levels in patients undergoing femtosecond laser-assisted cataract surgery. J Refract Surg 2013;29:742–7.

40. Abell RG, Kerr NM, Vote BJ. Toward zero effective phaco-emulsification time using femtosecond laser pretreatment.Ophthalmology 2013;120:942–8.

41. Mastropasqua L, Toto L, Mattei PA, et al. Optical coherencetomography and 3-dimensional confocal structured imagingsystem-guided femtosecond laser capsulotomy versus manualcontinuous curvilinear capsulorrhexis. J Cataract Refract Surg2014;40:2035–43.

42. Abell RG, Kerr NM, Howie AR, et al. Effect of femtosecondlaser-assisted cataract surgery on the corneal endothelium.J Cataract Refract Surg 2014;40:1777–83.

43. Daya SM, Nanavaty MA, Espinosa-Lagana MM. Trans-leticular hydrodissection, lens fragmentation, and influence onultrasound power in femtosecond laser-assisted cataract sur-gery and refractive lens exchange. J Cataract Refract Surg2014;40:37–43.

44. Tackman RN, Kuri JV, Nichamin LD, Edwards K. Anteriorcapsulotomy with an ultrashort-pulse laser. J Cataract RefractSurg 2011;37:819–24.

45. Schultz T, Joachim SC, Tischoff I, Dick HB. Histologicevaluation of in vivo femtosecond laser-generated capsuloto-mies reveals a potential cause for radial capsular tears. Eur JOphthalmol 2015;25:112–8.

46. Conrad-Hengerer I, Al Sheikh M, Hengerer FH, et al. Com-parison of visual recovery and refractive stability between

Popovic et al � Meta-Analysis of FLACS

femtosecond laser-assisted cataract surgery and standardphacoemulsification: six-month follow-up. J Cataract RefractSurg 2015;41:1356–64.

47. Ewe SYP, Abell RG, Oakley CL, et al. A comparative cohortstudy of visual outcomes in femtosecond laser-assisted versusphacoemulsification cataract surgery. Ophthalmology 2016;123:178–82.

48. Yu AY, Ni LY, Wang QM, et al. Preliminary clinical inves-tigation of cataract surgery with a noncontact femtosecondlaser system. Lasers Surg Med 2015;47:698–703.

49. Chen H, Lin H, Zheng D, et al. Expression of cytokines,chemokines and growth factors in patients undergoing cataractsurgery with femtosecond laser pretreatment. PLoS One2015;10:e0137227.

50. Chen H, Lin H, Chen W, et al. Femtosecond laser combinedwith non-chopping rotation phacoemulsification technique forsoft-nucleus cataract surgery: a prospective study. Sci Rep2015;6:18684.

51. Khoshnood B, Mesbah M, Jeanbat V, et al. Transformingscales of measurement of visual acuity at the group level.Ophthalmol Physiol Opt 2010;30:816–23.

52. Nagy ZZ, Mastropasqua L, Knorz MC. The use of femtosecondlasers in cataract surgery: review of the published results withthe LenSx system. J Refract Surg 2014;30:730–40.

53. Roberts TV, Lawless M, Bali SJ. Surgical outcomes andsafety of femtosecond laser cataract surgery: a prospectivestudy of 1500 consecutive cases. Ophthalmology 2013;120:227–33.

54. Murano N, Ischizaki M, Sato S, et al. Corneal endothelial celldamage by free radicals associated with ultrasound oscillation.Arch Ophthalmol 2008;126:816–21.

55. Shin YJ, Nishi Y, Engler C, et al. The effect of phacoemul-sification energy on the redox state of cultured human cornealendothelial cells. Arch Ophthalmol 2009;127:435–41.

56. Storr-Paulsen A, Norregaard JC, Ahmed S, et al. Endothelialcell damage after cataract surgery: divide-and-conquer versusphaco-chop technique. J Cataract Refract Surg 2008;34:996–1000.

57. Richard J, Hoffart L, Chavane F, et al. Corneal endothelial cellloss after cataract extraction by using ultrasound phacoemul-sification versus a fluid-based system. Cornea 2008;27:17–21.

58. Hoffer KJ, Aramberri J, Haigis W, et al. Protocols for studiesof intraocular lens formula accuracy. Am J Ophthalmol2015;160:403–5.

59. Marques FF, Marques DM V, Osher RH. New technique todemonstrate corneal magnification using trypan blue in cata-ract surgery. Rev Bras Oftalmol 2011;70:235–7.

60. Dick HB, Schelenz D, Schultz T. Femtosecond laser-assistedpediatric cataract surgery: Bochum formula. J CataractRefract Surg 2015;41:821–6.

61. Walkow T, Anders N, Pham DT, Wollensak J. Causes of se-vere decentration and subluxation of intraocular lenses.Graefes Arch Clin Exp Ophthalmol 1998;236:9–12.

62. Ravalico G, Tognetto D, Palomba M, et al. Capsulorrhexis sizeand posterior capsule opacification. J Cataract Refract Surg1996;22:98–103.

63. Korynta J, Bok J, Cendelin J. Changes in refraction induced bychange in intraocular lens position. J Refract Corneal Surg1994;10:556–64.

64. Okada M, Hersh D, Paul E, van der Straaten D. Effect of cen-tration and circularity of manual capsulorrhexis on cataract sur-gery refractive outcomes. Ophthalmology 2014;121:763–70.

65. Holladay JT, Zhao H, Reisin CR. Negative dysphotopsia: theenigmatic penumbra. J Cataract Refract Surg 2012;38:1251–65.

66. Prakash G, Prakash DR, Agarwal A, et al. Predictive factorand kappa angle analysis for visual satisfactions in patientswith multifocal IOL implantation. Eye (Lond) 2011;25:1187–93.

67. Alcon Laboratories, Inc. Infiniti Vision System� Operator’sManual. Available at: http://globexbd.com//model_file/INFINITI-Vision-System-Directions-for-Use.pdf. AccessedJuly 21, 2015.

68. Munoz G, Albarran-Diego C, Sakla HF, et al. Effects of LA-SIK on corneal endothelium using the 15-kHz IntraLasefemtosecond laser. J Refract Surg 2011;27:672–7.

69. Smith RT, Waring GO IV, Durrie DS, et al. Corneal endo-thelial cell density after femtosecond thin-flap LASIK andPRK for myopia: a contralateral eye study. J Refract Surg2009;25:1098–102.

70. Klingler KN, McLaren JW, Bourne WM, Patel SV. Cornealendothelial cell changes 5 years after laser in situ keratomil-eusis: femtosecond laser versus mechanical microkeratome.J Cataract Refract Surg 2012;38:2125–30.

71. Liu T, Zhang J, Sun D, et al. Comparative study of cornealendothelial cell damage after femtosecond laser assisteddeep stromal dissection. Biomed Res Int 2014;2014:731565.

72. Grewal DS, Singh Grewal SP, Basti S. Incomplete femto-second laser-assisted capsulotomy and lens fragmentation dueto emulsified silicone oil in the anterior chamber. J CataractRefract Surg 2014;40:2143–7.

73. Alder BD, Donaldson KE. Comparison of 2 techniques for man-aging posterior polar cataracts: traditional phacoemulsificationversus femtosecond laser-assisted cataract surgery. J CataractRefract Surg 2014;40:2148–51.

74. Kymionis GD, Kounis GA, Grentzelos MA, et al. Interfacecorneal stromal irregularities after flap creation using femto-second laser. Eur J Ophthalmol 2011;21:207–9.

75. Cherfan DG, Melki SA. Corneal perforation by an astigmatickeratometry performed with an optical coherence tomography-guided femtosecond laser. J Cataract Refract Surg 2014;40:1224–7.

76. Abell RG, Vote BJ. Cost-effectiveness of femtosecond laser-assisted cataract surgery versus phacoemulsification cataractsurgery. Ophthalmology 2014;121:10–6.

77. Kaiser PK. Prospective evaluation of visual acuity assessment: acomparison of Snellen versus ETDRS charts in clinical practice(an AOS thesis). Trans Am Ophthalmol Soc 2009;107:311–24.

78. Crema AS, Walsh A, Yarmane IS, et al. Femtosecond laser-assisted cataract surgery in patients with Marfan syndromeand subluxated lens. J Refract Surg 2015;31:338–41.

Footnotes and Financial Disclosures

Originally received: March 21, 2016.Final revision: June 15, 2016.Accepted: July 1, 2016.Available online: August 15, 2016. Manuscript no. 2016-573.1 Faculty of Medicine, University of Toronto, Toronto, Canada.

2 Department of Ophthalmology and Vision Sciences, University of Tor-onto, Toronto, Canada.3 Prism Eye Institute, Mississauga, Canada.4 Department of Ophthalmology, Trillium Health Partners, Mississauga,Canada.

2125

Ophthalmology Volume 123, Number 10, October 2016

Presented at: American Society of Cataract and Refractive Surgery/Amer-ican Society of Ophthalmic Administrators Symposium & Congress, May2016, New Orleans, Louisiana; and Canadian Ophthalmological SocietyAnnual Meeting & Exhibition, June 2016, Ottawa, Canada.

Financial Disclosure(s):The author(s) have made the following disclosure(s): I.I.K.A.: Financialsupport e Alcon (Fort Worth, TX); Abbott Medical Optics (Santa Ana,CA); Bausch & Lomb, Inc (Rochester, NY); Carl Zeiss AG (Oberkochen,Germany).

Author Contributions:

Conception and design: Popovic, Campos-Möller, Ahmed

Analysis and interpretation: Popovic, Campos-Möller, Schlenker, Ahmed

Data collection: Popovic, Campos-Möller, Schlenker, Ahmed

Obtained funding: none

2126

Overall responsibility: Popovic, Campos-Möller, Schlenker, Ahmed

Abbreviations and Acronyms:CCI ¼ clear corneal incision; CDE ¼ cumulative dissipated energy;CDVA ¼ corrected distance visual acuity; CI ¼ confidence interval;EUREQUO ¼ European Registry of Quality Outcomes for Cataract andRefractive Surgery; FLACS ¼ femtosecond laser-assisted cataract surgery;IOL ¼ intraocular lens; IOP ¼ intraocular pressure; logMAR ¼ logarithmof the minimum angle of resolution; MAE ¼ mean absolute error;MCS ¼ manual cataract surgery; NOS ¼ Newcastle-Ottawa QualityAssessment Scale; RR ¼ relative risk; UDVA ¼ uncorrected distance vi-sual acuity; WMD ¼ weighted mean difference.

Correspondence:Iqbal Ike K. Ahmed, MD, FRCSC, Prism Eye Institute, 3200 Erin MillsParkway, Unit 1, Mississauga, Ontario L5L 1W8, Canada. E-mail: [email protected].

Pictures & Perspectives

Merkel Cell Carcinoma of the EyelidAn 81-year-old immunocompetent man presented with a

5-week history of a rapidly growing left upper eyelid lesion(Fig 1, arrow). Eight years prior, he had a Merkel cell car-cinoma (MCC) of the left cheek that was treated with repeatwide local excision, limited neck dissection with negativesentinel lymph nodes, and postoperative radiation therapy.Histopathology revealed small blue cells (Fig 2), numerousmitotic figures, large oval nuclei, prominent nucleoli, and saltand pepper dense chromatin (Fig 3). Merkel cell carcinoma ofthe eyelids display an aggressive clinical course with a highrate of local recurrence (14%), regional lymph node invasion(20%), and metastasis (5%).

MEISHA L. RAVEN, DOPAUL D. SELID, MDMARK J. LUCARELLI, MD

Department of Ophthalmology, University ofWisconsineMadison,Madison, Wisconsin