ARTICLE

Influence of trifocal intraocular lenseson standard autorefraction and

aberrometer-based autorefractionQ6 Nuria Garz�on, PhD, María García-Montero, PhD, Esther L�opez-Artero, Msc, Francisco Poyales, MD,

C�esar Albarr�an-Diego, PhD

Purpose: To study the agreement between manifest refractionand objective refraction measured with two autorefractor modelsand an aberrometer in eyes implanted with a trifocal diffractive intra-ocular lens (IOL).

Setting: IOA Madrid Innova Ocular, Madrid, Spain.

Design: Prospective comparative cohort study.

Methods: An autorefractor keratometer (KR-8800), based on aScheiner double pinhole, and a 3-dimensionQ1 wavefronttopography aberrometer system (OPD-Scan III), based on thescanning-slit retinoscopy principle, were used to obtainobjective refraction readings. In addition, lower-order Zernikecoefficients (Z) were used to calculate objective refraction. A setof 7 different results was obtained in power vector notation(spherical equivalent [SE], Jackson cross-cylinder, axes at 180degrees and 90 degrees [J0] and Jackson cross-cylinder, axesat 45 degrees and 135 degrees [J45]) for 7 different methods:manifest refraction, autorefraction obtained with theautorefractor keratometer, WF-P (Z-based objective refractionfor the photopic pupil), WF-M (Z-based objective refraction for

the mesopic pupil), WF-4 (Z-based objective refraction for a4.0 mm pupil), OPD-C (autorefraction measured with the 3-dimension wavefront topography aberrometer system underphotopic conditions), and OPD-M (autorefraction measured withthe 3-dimension wavefront topography aberrometer systemunder mesopic conditions).

Results: The study comprised 102 eyes from 51 cataract patientswho underwent binocular implantation of a diffractive trifocal IOL(FineVision POD F). All 6 objective methods yielded more negativeSE values thanmanifest refraction (P < .001). As for the astigmatismcomponents (J0 and J45), only autorefraction (PZ .003) andOPD-M (P < .001) differed significantly from manifest refraction. The bestand worst correlation for the SE component were intraclasscorrelation coefficient (ICC) Z 0.70 (for WF-M) and ICC Z 0.48(for WF-4).

Conclusion: Objective methods tend to yield more negativesphere values than manifest refraction.

J Cataract Refract Surg 2019; -:-–- Q 2019 ASCRS and ESCRS

Cataract surgery is increasingly becoming a refractivesurgery procedure that seeks to provide patientswith the best possible vision for all viewing dis-

tances. Multifocal intraocular lens (IOL) implantation en-ables the patient to be spectacle-free after cataract surgeryfor both distance and near vision, offering an improvedquality of life; therefore, the number of patients optingfor this type of IOL has increased. These IOLs can be clas-sified based on their design: namely, there are refractiveversus diffractive; monofocal, bifocal, or trifocal; andthey can be either rotationally symmetric or asymmetric.Multifocal IOLs make the most of the brain’s ability to

adapt to far or near vision because different elements of

the lens are used, depending on where the patient isfocusing.In devising the present study, our main goal was to assess

the quality of vision that multifocal IOLs provide using boththe standard approach and quicker alternative measuringtechniques. Subjective or manifest refraction is the gold-standard method to determine the eye’s refractive status,whereas autorefraction is considered a fast and reliablemethod to measure refraction in the general population.However, autorefraction has not yet been able to fullyoust and replace manifest refraction; in fact, it is commonlyused just to provide an optimum starting point for manifestrefraction assessment. The accuracy of autorefraction can

Submitted: December 28, 2018 | Final revision submitted: April 16, 2019 | Accepted: April 17, 2019

From the IOA Madrid Innova Ocular (Garz�on, L�opez-Artero, Poyales), Madrid, Optometry and Vision Department (Garz�on, García-Montero), Faculty of Optics andOptometry, Complutense University of Madrid, Optics (Albarr�an-Diego), Optometry and Vision Science Department. Faculty of Physics, University of Valencia, andClínica Baviera Castell�on (Albarr�an-Diego), Castell�on de la Plana, Spain.

Corresponding author: Nuria Garz�on, PhD, IOA Madrid Innova Ocular, C/Galileo 104, 28003 Madrid, Spain. E-mail: ngarzon@ioamadrid.com.

Q 2019 ASCRS and ESCRSPublished by Elsevier Inc.

0886-3350/$ - see frontmatterhttps://doi.org/10.1016/j.jcrs.2019.04.017

1

123456789

10111213141516171819202122232425262728293031323334353637383940414243444546474849505152535455565758

5960616263646566676869707172737475767778798081828384858687888990919293949596979899

100101102103104105106107108109110111112113114115116

FLA 5.6.0 DTD � JCRS10283_proof � 27 June 2019 � 12:50 am

be compromised by previous corneal refractive surgeries,media opacities, a small pupil size, and the presence ofmultifocal IOLs.1–5

In this context, the purpose of this study was to evaluatethe agreement between manifest refraction and the objec-tive refraction readings provided by the aforementionedautomated devices in eyes that had diffractive trifocal IOLimplantation.

MATERIALS AND METHODSStudy Design and Patient PopulationThis was a prospective comparative cohort study of patients un-dergoing cataract surgery and binocular implantation of a trifocalIOL (FineVision POD F, PhysIOL S.A.) at IOA Madrid InnovaOcular, Madrid, Spain. All patients provided written informedconsent before enrollment. This study was approved by the localethics committee, and it was performed in accordance with theDeclaration of Helsinki.Other inclusion criteria were the desire for spectacle indepen-

dence after surgery with realistic expectations and the availabilityand willingness to comply with all the study visits and eyeexaminations.Exclusion criteria were a history of ocular disease other than

cataract (eg, uveitis, amblyopia, glaucoma), astigmatism above1.25 diopters (D), any acute or chronic condition that would in-crease the risk or confound study results, any capsule or zonularabnormalities that might affect postoperative centration or tilt ofthe IOL, and the presence of pupil abnormalities.

Surgical ProcedureAll surgeries were carried out by the same surgeon (F.P.) undertopical anesthesia. A 2.2 mm corneal incision and a paracentesiswere made with a surgical knife. Anterior capsulotomy and nu-clear fragmentation were performed with a femtosecond laser un-der optical coherence tomography image control (CATALYSPrecision System, Abbott Medical Optics, Inc.), and for lensphacoemulsification, a commercial microsurgical system (Centu-rion Vision System, Alcon Laboratories, Inc.) was employed. Twoophthalmic viscosurgical devices were used throughout the entireprocedure: the cohesive sodium hyaluronate 1.0% (Healon) andthe dispersive sodium hyaluronate 1.2% (Amvisc). The POD FIOL was then implanted into the capsular bag with a single-use in-jection system (Microse, PhysIOL S.A.). In all cases, a capsular ten-sion ring was inserted. All surgeries were supported by thecomputer-assisted cataract surgery system (CALLISTO Eye fromthe Cataract Suite Markerless, Carl Zeiss Meditiec AG).Once the procedure was completed, patients were treated with a

combination of antibiotics, corticosteroids, and antiinflammatoryeyedrops (moxifloxacin, dexamethasone, and bromfenac).

Intraocular LensThe POD F IOL model is a spherical trifocal IOL that combinestwo diffractive structures. This combination provides three foci:C0.00 D for far vision, a C1.75 D addition for intermediatevision, and a C3.50 D addition for near vision. This correspondsto a nominal intermediate addition of approximatelyC1.2 D anda near addition of approximatelyC2.4 D at the corneal plane, de-pending on the particular geometry of the eye. The IOL’s optics isbiconvex aspheric (spherical aberration [SA]�0.11 mm). The IOLhas a diffractive anterior surface that is entirely convoluted. Byvarying the step height of the IOL’s diffractive structure acrossthe pupil, the energy distribution for different distances can becontrolled.6 The amount of energy directed to far vision focus issuperior to that directed to intermediate and near vision focuswith increasing apertures by a gradual decrease in the height of dif-fractive steps from the center to the periphery, which is also the

case for the refractive multifocal IOL. The lens is 26% hydrophilicacrylic and has a ultraviolet and blue-light blocker. It has an optic-body diameter of 6.00 mm and an overall diameter of 11.40 mm,its refractive index is 1.46, and it has a 5-degree angulation.

Postoperative Eye ExaminationsSubjective Refraction Patients were examined 1 day, 1 week, and1 month after surgery, although the data reported in this paperwere taken at the 1-month visit. The decision to analyze the1-month postoperative data was based on the study authors’ pre-vious experience regarding the stability of refractive results at thistime, and not before.All refraction assessments were carried out by the same optom-

etrist (N.G.). Manifest refraction was always performed underphotopic conditions, with the same illumination for all patients,using the Early Treatment Diabetic Retinopathy Study chartwith a trial frame. Objective refraction was measured underboth photopic and mesopic conditions. The best visual acuity sce-nario manifest refraction was then further fine-tuneddbothspherical and cylindrical componentsdwith cross-cylinders insteps of 0.25 D. Other tests that were performed were bio-microscopy, tonometry, and fundus evaluation.

Autorefraction The autorefractor keratometer (KR-8800, TopconCorp.) is a multifunctional device that determines corneal refrac-tive status using a rotary prismmeasuring system. This device pro-vides keratometry measurements for corneal diameters rangingfrom 2.0 mm to 7.7 mm according to the presence of anteriorcorneal astigmatism, with keratometry values and corneal curva-tures obtained over a range from 5.00 mm to 10.00 mm(0.01 mm, step display). The autorefractor keratometer relies onthe Scheiner double-pinhole principle for data capture: two lightsources are imaged onto the pupil plane to simulate the Scheinerpinhole apertures. First, the Badal system is focused onto one me-ridian, and then continuous measurements are taken throughout a180-degree range using a rotating prism system. A “fogging” targetwas used to relax accommodation.7 Automatic capture of 4 mea-sures was repeated twice, and the average values were used for sta-tistical analysis. Measurement accuracy was set to 0.12 D for powerand to 1 degree for axis, as advised by the manufacturer.8,9

Aberrometry The 3-dimension Q2wavefront topography aberrome-ter system (OPD-Scan III, Nidek Co., Ltd.) is an aberrometer–corneal topographer workstation. It combines a wavefrontaberrometer, a topographer, an autorefractor, an autokeratometer,and a pupillometer, all in one device. The autorefractor relies onthe principle of scanning-slit retinoscopy, where the retina isscanned with an infrared slit beam. Measurement light emittedin a grid-like pattern is projected onto the retina, and the light re-flected from the retina is then captured by multiple pairs of pho-todetectors. Refraction of the eye causes time (phase) differencesin the signals sent out by these pairs of photodetectors. The devicecalculates the patient’s refraction (spherical and cylindrical refrac-tive errors, as well as the cylinder axes angle) based on these phasedifferences.10 In addition to providing objective refraction in theform of a spherocylindrical reading, the aberrometer also com-putes the Zernike coefficients for lower-order and higher-orderaberrations. The Zernike coefficients corresponding to lower-order aberrations Z(0,2), Z(2,C2) and Z(2,�2) can be used tocalculate objective refraction in vector notation (SE, J0, and J45)according to the following expressions11: Q3

SEZ4ffiffiffi3

p

r2Z02

J0Z2ffiffiffi6

p

r2Zþ22

2 OBJECTIVE REFRACTION IN EYES WITH DIFFRACTIVE TRIFOCAL IOLS

Volume - Issue - - 2019

117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174

175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232

FLA 5.6.0 DTD � JCRS10283_proof � 27 June 2019 � 12:50 am

J45Z2ffiffiffi6

p

r2Z�22

where r is the pupil radius (or semidiameter) measured by the3-dimension wavefront topography aberrometer system either inphotopic or mesopic conditions, SE is the spherical equivalent,J0 is the vertical Jackson cross-cylinder, axes at 180 degrees and90 degrees and J45 is the oblique Jackson cross-cylinder, axes at45 degrees and 135 degrees.Moreover, these vector components can be turned into the

more common clinical spherocylindrical notation (sphere [S], cyl-inder [C], and axis [A]) using the following expressions11,12:

C Z � 2ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiJ20 þ J245

q

S Z SE�C2

AZ12tan�1J45

J0

Statistical AnalysisThemanifest refraction and objective refraction values obtained inclinical spherocylindrical notation were converted into power-vector notation for comparison purposes, by means of thefollowing expressions11:

SE Z SþC2

J0 Z �C2cos 2 A

J45 Z �C2sin 2 A

A set of three objective refraction calculations of the Zernike co-efficients were performed: one for photopic pupil, one for mesopicpupil, and one for a “standard” pupil fixed at 4.0 mm because thiswas the value yielding the best agreement for SE in a previousstudy by Campbell.13 Given that the IOL implanted in this studyhas an aspherical profile (SA Z �0.11 mm) not designed to fullycompensate for the average corneal SA of the human eye(w0.27 mm), a decision was made to consider automated objectiverefraction values for more than one pupil diameter to checkwhether one of those objective measurements was statistically bet-ter correlated to subjective refraction than the others.For each eye included in the study, 7 result sets (one for each

assessment method) were collected: manifest refraction, autore-fractionmeasured with the autorefractor keratometer, WF-P (Zer-nike-coefficients-based objective refraction, photopic pupil size),WF-M (Zernike-coefficients-based objective refraction, mesopicpupil size), WF-4 (Zernike-coefficients-based objective refraction,4.0 mm pupil), OPD-C (autorefraction measured with the3-dimension wavefront topography aberrometer system in thecentral pupil/photopic conditions), and OPD-M (autorefractionmeasured with the 3-dimension wavefront topography aberrome-ter system under mesopic conditions).SigmaPlot software for Windows (version 12, Systat Software,

Inc.) was used for statistical analysis and graphic plotting. TheFriedman repeated-measurements analysis of variance on rankswas used to look for differences across the 7 assessmentmethods for each of the refraction vector components. Whendifferences were found, pairwise multiple-comparison testingwas applied by the Tukey test to identify those differences.Agreement was evaluated with Bland-Altman plots, and ICCswere calculated with Medcalc Statistical software for Windows(version 12.5, MedCalc Software bvba) to study the strength

of the agreement between methods.14 Statistical significancewas set at a Z 0.05.Considering a repeated-measurements design, with 0.25 diop-

ters (D) as the minimum clinically relevant difference in refrac-tion, and estimating an expected standard deviation of thedifferences two times this mean value (0.50 D, based on previousexploratory measurements), the sample size estimated with theSigmaPlot software to reach a proper statistical power(1 � b) Z 0.80, resulted in n Z 34. A decision was then madeto recruit as many patients as possible above this amount.

RESULTSThe study comprised 102 eyes from 51 patients. The meanage was 67.2 years G 8.5 (SD). The mean pupil size was3.31 G 0.62 mm and 4.71 G 0.84 mm under photopicand mesopic conditions, respectively.Table 1 shows a summary of the surgical outcomes in

terms of refraction and visual acuity. The average refractiveresult was very close to emmetropia, with a mean SE of�0.08 D and both astigmatic components being below0.05 D.Figure 1 shows a boxplot illustrating the differences be-

tweenmanifest refraction outcomes and each of the 6 objec-tive refraction measuring approaches under evaluation, forsphere, SE, and the astigmatism components (J0 and J45).Table 2 shows meansG SD and range for all the objective

refraction methods considered.As Figure 1 shows, all 6 objective methods produced

more negativedor less positivedsphere and SE outcomesthan manifest (subjective) refraction. In particular, WF-Pyielded the biggest average difference for SE (mean differ-ence with manifest refraction: �0.73 G 0.69 D), whereasthe closest results to manifest refraction were obtainedwith OPD-C (mean difference with manifest refraction:�0.27G 0.34 D). Regarding the astigmatism components,for J0,WF-Mwas closest to the manifest refraction readings(mean difference: 0.00 G 0.20 D), whereas autorefractionyielded the biggestdalthough minorddifferences withmanifest refraction (mean difference: �0.07 G 0.19 D).In contrast, for the J45 component, the lowest differenceswere obtained with autorefraction (0.01G 0.14 D), whereas

Table 1. Descriptive statistics obtained after surgery forrefraction and visual acuity.

Parameter Mean ± SD Range

Refraction (D)

SE �0.08 G 0.27 �1.00, 0.50

J0 �0.03 G 0.13 �0.48, 0.25

J45 0.01 G 0.11 �0.35, 0.43

Sph 0.01 G 0.25 �0.75, 0.75

Cyl �0.17 G 0.31 �1.25, 0.00

Visual acuity (logMAR)

UDVA 0.04 G 0.07 0.40, �0.10

CDVA 0.00 G 0.03 0.14, �0.10

CDVA Z corrected distance visual acuity; Cyl Z cylinder; J0 Z verticalJackson cross-cylinder, axes at 180 degrees and 90 degrees;J45 Z oblique Jackson cross-cylinder, axes at 45 degrees and 135 de-grees; logMAR Z logarithm of the minimum angle of resolution;SEZ spherical equivalent; Sph Z sphere; UDVA Z uncorrected distancevisual acuity

3Q5OBJECTIVE REFRACTION IN EYES WITH DIFFRACTIVE TRIFOCAL IOLS

Volume - Issue - - 2019

233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290

291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348

FLA 5.6.0 DTD � JCRS10283_proof � 27 June 2019 � 12:50 am

OPD-M yielded the largest differences withmanifest refrac-tion (0.06 G 0.24 D).Using the Friedman repeated-measures analysis of vari-

ance on ranks, statistically significant differences werefound between manifest refraction and the objective refrac-tion approaches, both for the sphere (P ! .001) and SE(P ! .001) values. In particular, the post hoc testing forsphere revealed differences between manifest refractionand each of the objective methods, except for WF-M. Asfor SE, the Turkey post hoc testing revealed differences be-tween manifest refraction and each of the objectivemethods. Regarding astigmatism, the Friedman repeated-measures analysis of variance on ranks also revealed statis-tically significant differences between methods, both for J0(P Z .003) and J45 (P ! .001). More specifically, post hoctesting identified significant differences for J0 only betweenmanifest refraction and autorefraction (but not for the re-maining manifest refraction–objective refraction pairwisecomparison), whereas for J45, it was manifest refractionversus OPD-M, which was the only pairwise comparisonthat yielded significant differences from the 6 objectiverefraction approaches.Figures 2 through 5 show the Bland-Altman plots for

sphere and each component of the power vector. In eachplot, the vertical axis represents the difference found be-tween each objective method and the subjective manifestrefraction outcomes, whereas the horizontal axis indicatesthe corresponding manifest refraction value. This absolutemanifest refraction value was plotted, rather than plottingthe average across all methods, because subjective refrac-tion is considered the gold standard technique for refractivestatus determination.

Table 3 shows the resulting ICCs for each measuringmethod and for each refractive component. Q4The strongestcorrelation with manifest refraction for sphere values wasfound for WF-M, whereas the weakest was for WF-P. Forthe SE, the strongest and weakest correlations with manifestrefraction were for WF-M and WF-4, respectively. Autore-fraction showed the strongest correlation with manifestrefraction for astigmatism, whereas WF-P showed theweakest correlation.

DISCUSSIONThis prospective study compared clinically obtained mani-fest refraction versus objective refraction after trifocal dif-fractive IOL implantation.Because of the multifocal nature of trifocal diffractive

IOLs, their depth of focus is larger than that of standardmonofocal IOLs. The absence of a unique focal plane makesit more difficult for us to determine unambiguously andaccurately our patients’ objective or subjective manifestrefraction.15

Several methods are available to estimate refractive er-ror after lens extraction and IOL implantation; theseinclude keratometry, retinoscopy, and autorefraction;however, manifest refraction is still considered the goldstandard. Autorefraction’s accuracy has been found todecrease in the presence of a multifocal IOL,2,4,16 whereasretinoscopy becomes more complicated to perform withsome of these lensesdsuch as refractive sectorial IOLsdbecause of the presence of two opposite retinoscopyshadows.1

Several authors have compared objective and subjectiverefraction values in refractive and bifocal diffractive IOL

Figure 1. Differences between objectiverefraction and subjective (manifest)refraction for sphere, SE, and J0 andJ45 (vector components of astigmatism),and for each of the 6 objective refractionscenarios under assessment. The aster-isks (*) indicate statistically significantdifferences (AR Z autorefraction;J0 Z vertical Jackson cross-cylinder,axes at 180 degrees and 90 degrees;J45 Z oblique Jackson cross-cylinder,axes at 45 degrees and 135 degrees;OPD-C Z autorefraction measuredwith the 3-dimension wavefront topog-raphy aberrometer system in thecentral pupil/photopic conditions;OPD-M Z autorefraction measured withthe 3-dimension wavefront topographyaberrometer system under mesopic con-ditions; SE Z spherical equivalent;Sph Z sphere; WF-4 Z wavefront4.0 mm; WF-M Z wavefront mesopic;WF-P Z wavefront photopic).

4 OBJECTIVE REFRACTION IN EYES WITH DIFFRACTIVE TRIFOCAL IOLS

Volume - Issue - - 2019

349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406

407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464

FLA 5.6.0 DTD � JCRS10283_proof � 27 June 2019 � 12:50 am

wearers; however, to our knowledge, this is the first attemptto perform such a comparison with the more complextrifocal IOLs, which have three foci (for far, intermediate,and near distances).

Our results suggest that objective methods for postsur-gical refraction evaluation (objective refraction) tend toyield more negative sphere values than manifest refraction.A similar tendency has been found for bifocal IOLs to a

Table 2. Descriptive statistics for the objective refractions obtained with all the evaluated methods.

Parameter AR WF-P WF-M WFL4 OPD-C OPD-M

Sph

Mean G SD �0.13 G 0.41 �0.30 G 0.87 �0.15 G 0.43 �0.25 G 0.62 �0.17 G 0.44 �0.25 G 0.43

Range �0.75, 1.00 �2.85, 2.52 �1.50, 1.12 �1.86, 1.49 �1.50, 1.25 �1.75, 0.50

Cyl

Mean G SD �0.48 G 0.35 �1.00 G 0.66 �0.48 G 0.30 �0.66 G 0.42 �0.35 G 0.31 �0.50 G 0.58

Range �1.50, 0.00 �3.47, �0.03 �1.54, �0.02 �1.98, �0.02 �1.50, 0.00 �2.25, 1.50

SE

Mean G SD �0.37 G 0.40 �0.80 G 0.80 �0.39 G 0.39 �0.58 G 0.60 �0.35 G 0.41 �0.50 G 0.51

Range �1.13, 0.88 �2.87, 1.10 �1.51, 0.57 �2.18, 1.07 �1.50, 0.63 �1.88, 0.88

J0

Mean G SD �0.11 G 0.22 �0.08 G 0.46 �0.04 G 0.22 �0.05 G 0.29 �0.04 G 0.17 �0.01 G 0.29

Range �0.73, 0.37 �1.70, 1.18 �0.76, 0.58 �0.82, 0.77 �0.74, 0.49 �0.74, 1.12

J45

Mean G SD 0.02 G 0.17 �0.02 G 0.38 �0.01 G 0.18 �0.02 G 0.26 �0.02 G 0.15 0.07 G 0.24

Range �0.38, 0.65 �1.21, 0.97 �0.54, 0.44 �0.93, 0.55 �0.61, 0.37 �0.86, 0.79

ARZ autorefraction; CylZ cylinder; J0Z vertical Jackson cross-cylinder, axes at 180 degrees and 90 degrees; J45Z oblique Jackson cross-cylinder, axesat 45 degrees and 135 degrees; OPD-CZ autorefraction measured with the 3-dimension wavefront topography aberrometer system in the central pupil/phot-opic conditions; OPD-M Z autorefraction measured with the 3-dimension wavefront topography aberrometer system under mesopic conditions;SE Z spherical equivalent; Sph Z sphere; WF-4 Z wavefront 4.0 mm; WF-M Z wavefront mesopic; WF-P Z wavefront photopic

print&

web4C=FPO

Figure 2. Bland-Altman plots for thesphere, showing the agreement betweensubjective (manifest) refraction andeach of the 6 objective refractionapproaches (AR Z autorefraction; OPD-CZ autorefraction measured with the3-dimension wavefront topography aberr-ometer system in the central pupil/phot-opic conditions; OPD-M Z autorefractionmeasured with the 3-dimension wavefronttopography aberrometer system undermesopic conditions; WF-4 Z wavefront4.0 mm; WF-M Z wavefront mesopic;WF-P Z wavefront photopic).

5OBJECTIVE REFRACTION IN EYES WITH DIFFRACTIVE TRIFOCAL IOLS

Volume - Issue - - 2019

465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522

523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567568569570571572573574575576577578579580

FLA 5.6.0 DTD � JCRS10283_proof � 27 June 2019 � 12:50 am

greater or lesser extent and, more specifically, a strongertrend toward myopia has been observed for refractivebifocal IOLs than for diffractive ones.4,17

As for sectorial refractive IOLs, van der Linden et al.2 andAlbarr�an-Diego et al.1 found a systematic shift toward morenegative values for autorefraction, between 1.00 D and 1.25D, compared with subjective refraction, with poor correla-tion for cylinder values. Bissen-Miyajima et al.3 reportedsimilar issues with refractive IOLs, hypothesizing that pupilsize might confound autorefraction outcomes; actually, thegeometry of the lens could be the main factor causing mea-surement inaccuracy.Other studies, which compared a concentric refractive

IOL model with two bifocal diffractive ones,4,17 concludedthat the presence in the IOL of several concentric refractivezonesdwhich results in the overlapping of two images atthe retinal planedcould cause an undesirable scatteringof the autorefractor’s infrared beam, thus leading to inaccu-rate results.In contrast, with diffractive bifocal IOL wearers, the

autorefractor proved useful as a starting point to esti-mate manifest refractiondboth its spherical and itsastigmatic componentsdthus highlighting the note-worthy differences between refractive and diffractiveIOLs.4 In their study, Mu~noz et al.4 found that the

mean spherical power difference between autorefractionand subjective refraction was near zero for the bifocaldiffractive IOL models considered (0.03 G 0.09 D forthe Restor model [Alcon Laboratories, Inc.] and�0.05 G 0.11 D for the Tecnis model [Johnson & John-son Vision Care, Inc.]).As for our study, encompassing trifocal diffractive IOL

wearers only, we assessed two autorefractor modelsdonethat was based on the Scheiner double-pinhole principle(autorefractor keratometer) for data capture and anotherone that relies on the principle of scanning-slit retinoscopy(3-dimension wavefront topography system, used undereither photopic or mesopic conditions)dtogether with anaberrometer that can measure the aberration pattern fordifferent pupil sizes. Similar results were obtained for theautorefractor keratometer (mean difference between objec-tive refraction and manifest refraction: �0.29 G 0.39 D)and the 3-dimension wavefront topography aberrometersystem under photopic conditions (mean difference be-tween objective refraction and manifest refraction:�0.27 G 0.34 D). The results did not correlate as wellwhen using the 3-dimension wavefront topographyaberrometer system under mesopic conditions (meandifference between objective refraction and manifestrefraction: �0.42 G 0.47 D).

print&

web4C=FPO

Figure 3. Bland-Altman plots for SE,showing the agreement between sub-jective (manifest) refraction andeach of the 6 objective refraction ap-proaches (AR Z autorefraction;OPD-C Z autorefraction measuredwith the 3-dimension wavefront topog-raphy aberrometer system in thecentral pupil/photopic conditions;OPD-M Z autorefraction measuredwith the 3-dimension wavefront topog-raphy aberrometer system under mes-opic conditions; SE Z sphericalequivalent; WF-4 Z wavefront4.0 mm; WF-M Z wavefront mesopic;WF-P Z wavefront photopic).

6 OBJECTIVE REFRACTION IN EYES WITH DIFFRACTIVE TRIFOCAL IOLS

Volume - Issue - - 2019

581582583584585586587588589590591592

593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630

631632633634635636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672673674675676677678679680681682683684685686687688

FLA 5.6.0 DTD � JCRS10283_proof � 27 June 2019 � 12:50 am

In our study we found a more myopic shift in objectiverefraction values with trifocal IOLs than that found in thestudies by Mu~noz et al.4,17 with bifocal IOLs. The additionof a third focus could be one of the reasons, but pupil sizeand SA could also be a couple of factors explaining thisbehaviour. In our sample, the mean value for spherical po-wer in subjective refraction was near zero (�0.08 G 0.27),whereas objective measurements yielded average valuesranging from �0.13 G 0.41 (for autorefraction) to�0.30 G 0.87 (for the WF-P). The implanted lens has anSA of �0.11 mm, which is not enough to compensate forthe average corneal SA. In fact, total SA in our cohort post-operatively, resulted in a mean value of 0.24 G 0.13 mm.This could explain why 3-dimension wavefront topographyaberrometer system measurements resulted in moremyopic objective refractions for mesopic than photopic pu-pil. Also, this could explain the myopic shift whencompared with subjective refraction, in which the StilesCrawford effect could account for the bigger height of thecentral than the peripheral rays incoming the pupil, andthus, diluting the effect of SA.Concerning the use of aberrometry to assess multifocal

IOL wearers, Charman et al.18 concluded that aHartmann-Shack aberrometer might not provide reliable

information on the wavefront aberration associated witheither the distance or the near components of diffractiveIOLs because the results could depend on factors such asthe power of the diffractive addition and the relative ampli-tudes of the distance and near wavefronts.In this same context, Campbell19 used a Hartmann-

Shack WaveScan aberrometer (Abbott Medical Optics,Inc.) to evaluate two different IOLs: a refractive model (Re-Zoom, Abbott Medical Optics, Inc.) versus a diffractive one(Tecnis multifocal C4.00). The study was performed in anartificial eye, and in this scenario, he was not able to reliablymeasure distance refraction and higher-order aberrationswith a clear tendency towardmyopia (the sphere was higherthan �1.25 D, and it was dependent on pupil size). Incontrast, the Tecnis IOL could be measured reliably; thevalues obtained were close to the set value of 0.00 D ofsphere.Jendritza et al.20 found a similar trend in vivo for a dif-

fractive IOL (Tecnis multifocal C4.00) when using thesame aberrometer as Campbell19 to compare its readingswith manifest refraction values (sphere C0.45D, cylinder�0.14D). When assessing wearers of a diffractive apo-dized IOL (Restor C4.00), aberrometer results and man-ifest refraction values were very comparable; however, the

print&web4C=FPO

Figure 4. Bland-Altman plots for J0(vector component of astigmatism),showing the agreement between sub-jective (manifest) refraction and eachof the 6 objective refraction methods(AR Z autorefraction; J0 Z verticalJackson cross-cylinder, axes at 180degrees and 90 degrees;OPD-C Z autorefraction measuredwith the 3-dimension wavefront topog-raphy aberrometer system in the cen-tral pupil/photopic conditions;OPD-M Z autorefraction measuredwith the 3-dimension wavefront topog-raphy aberrometer system under mes-opic conditions; WF-4 Z wavefront4.0 mm; WF-M Z wavefront mesopic;WF-P Z wavefront photopic).

7OBJECTIVE REFRACTION IN EYES WITH DIFFRACTIVE TRIFOCAL IOLS

Volume - Issue - - 2019

689690691692693694695696697698699700701702703704705706707708709710711712713714715716717718719720721722723724725726727728729730731732733734735736737738739740741742743744745746

747748749750751752753754755756757758759760761762763764765766767768769770771772773774775776777778779780781782783784785786787788789790791792793794795796797798799800801802803804

FLA 5.6.0 DTD � JCRS10283_proof � 27 June 2019 � 12:50 am

former showed greater variability than the latter and ten-dency toward myopia (sphere �0.34 D, cylinder �0.20D). Our findings, which also reveal a slight trend towardmyopia, are in good agreement with these studies.As for our assessment of the 3-dimension wavefront

topography aberrometer system, which is based on thescanning-slit retinoscopy principle, it is worth highlightingthat the SE readings that were closest to manifest refractionwere obtained with this aberrometer under mesopic condi-tions (�0.32 G 0.32 D), whereas the readings that differedthe most were the ones yielded by this same aberrometer

but under photopic conditions (�0.73 G 0.69 D).Compared with the Jendritza et al.20 results for a 4.0 mmpupil, our results showed more myopia drift (sphere�0.26 D versus C0.45 D respectively).Regarding the astigmatic components of the power

vector, the resulting J0 and J45 were similar for theaberrometersdclose to 0.00 Ddirrespective of pupilsize. We would like to emphasize the importance of un-derstanding the distinction between “statistical” and“clinical” differences. These astigmatic componentsshowed very low values in our patient population

print&web4C=FPO

Figure 5. Bland-Altman plots for J45(vector component of astigmatism),showing the agreement between sub-jective (manifest) refraction and eachof the 6 objective refraction ap-proaches (AR Z autorefraction;J45 Z oblique Jackson cross-cylinder, axes at 45 degrees and 135degrees; OPD-C Z autorefractionmeasured with the 3-dimension wave-front topography aberrometer systemin the central pupil/photopic condi-tions; OPD-M Z autorefractionmeasured with the 3-dimension wave-front topography aberrometersystem under mesopic conditions;WF-4 Z wavefront 4.0 mm;WF-M Z wavefront mesopic;WF-P Z wavefront photopic).

Table 3. ICCs for each of the methods and for each of the refractive components.

Parameter

ICC

AR WF-P WF-M WF-4 OPD-C OPD-M

Sph 0.45 0.38 0.57 0.43 0.52 0.49

SE 0.51 0.50 0.70 0.48 0.69 0.52

J0 0.60 0.39 0.58 0.49 0.62 0.40

J45 0.72 0.37 0.56 0.41 0.58 0.37

AR Z autorefraction; ICC Z intraclass correlation coefficient; J0 Z vertical Jackson cross-cylinder, axes at 180 degrees and 90 degrees; J45 Z obliqueJackson cross-cylinder, axes at 45 degrees and 135 degrees; OPD-C Z autorefraction measured with the 3-dimension wavefront topography aberrometersystem in the central pupil/photopic conditions; OPD-M Z autorefraction measured with the 3-dimension wavefront topography aberrometer system undermesopic conditions; SE Z spherical equivalent; Sph Z sphere; WF-4 Z wavefront 4.0 mm; WF-M Z wavefront mesopic; WF-P Z wavefront photopic

8 OBJECTIVE REFRACTION IN EYES WITH DIFFRACTIVE TRIFOCAL IOLS

Volume - Issue - - 2019

805806807808809810811812813814815

816817818819820821822823824825826827828829830831832833834835836837838839840841842843844845846847848849850851852853

854855856857858859860861862863864865866867868869870871872873874875876877878879880881882883884885886887888889890891892893894895896897898899900901902903904905906907908909910911

FLA 5.6.0 DTD � JCRS10283_proof � 27 June 2019 � 12:50 am

(resulting from the low-astigmatism inclusion criterion);therefore, even though statistical testing did reveal signif-icant differences (see autorefraction for J0 and OPD-Mfor J45 in Figure 1), those differences are not clinicallyrelevant because the magnitude was below 0.1 D (seeTable 2).It is well known that diffractive efficiency falls and add

power increases as wavelength increases21,22; therefore, itmight be that the Hartmann-Shack aberrometers that uselonger infrared-light wavelengths are more likely to pro-duce wavefront results that correspond to the wavefrontproduced by the distance power of a diffractive IOL. Thisproblem should not arise when using the OPD-Scan IIIaberrometer because it relies on a different principle (ie,retinoscopy); however, as mentioned above, relevant differ-ences are observed depending on whether the measure-ments are performed under either mesopic or photopicconditions.Discrepancies between objective aberrometer-based

refraction and subjective refraction have been attributedto many variables, including the eye’s longitudinal chro-matic aberration,23 accommodative status during theobjective measurement,24 the different retinal referenceplanes chosen by each technique,25 the Stiles-Crawfordeffect,26 image noise in the objective measurements,24

and the merit function that is chosen to determine bestfocus.27

We hypothesize that the use of narrow beams might leadto erroneous results because the diffractive behavior re-quires that the area of the lens illuminated is sufficientlylarge for adequate summation of secondary wavelets tooccur.The sphere component (S) can be calculated in terms of

C(2,0), which is the Zernike expansion’s defocus coefficientthat the aberrometer yields:

S Z � 4ffiffiffiffiffiffi3c02

pR2

!

where R is the pupil radius. Because for a given aberrationpattern, the resulting S is inversely proportional to Rsquared, the impact of defocus will be greater under phot-opic conditions (ie, small pupil sizes) than under mesopicones. Moreover, we have to bear in mind that under phot-opic conditions and when no mydriatic agent has beenapplied (ie, nondilated pupil), the depth of focus could in-crease; our hypothesis is that this scenario can make wave-front measurement more difficult, thus resulting inrefraction/power calculation errors.Another parameter that could lead to variability is the

wavelength used by the measuring system. It might alsocontribute to the inaccuracy and imprecision of objectivewavefront refraction.28 The optical performance of diffrac-tive multifocal IOLs, measured under either visible or near-infrared illumination differs considerably: namely, theseIOLs show two distinct (near and far) foci under visible

light, whereas under near-infrared illumination, their per-formance outcomes are clearly biased in favor of their farfocus. These results might help prevent a misleading useof near-infrared-based clinical instruments for the assess-ment of eyes implanted with diffractive multifocal IOLs.29

The longer the near-infrared wavelength, the weaker thenear focus,18,30 and thus, reported wavefront measurementsperformed with aberrometers that rely on longer wave-lengths (808 nm and 850 nm)31,32 in patients who have dif-fractive multifocal IOL implants would produce even morebiased results, and the properties of the near focus would behard to discern because of the much stronger presence ofthe far focus. The OPD-Scan III uses an 808 nm lightsource, which could thus lead to the aforementionedmeasuring errors.Based on the outcomes of our study about a trifocal IOL,

we could conclude that no objective measuring technique isas accurate as the subjective method with which the patientattains the best visual acuity possible. Among the objectivemethods under assessment, the aberrometer (for mesopicpupil sizes) and the autorefractor keratometer are theones that achieved the best outcomes.The results obtained in this work have important impli-

cations in those patients with refractive surprise aftertrifocal IOL implantation programmed for laser enhance-ment of the residual refraction. The proper measurementof that residual refraction is mandatory to achieve thebest result and patient satisfaction. Given that theseIOLs have three focus parameters and a greater depthof focus, careful attention must be paid to properly mea-sure the refractive status through the far focus of the IOL,and not through intermediate or near foci. For thisreason, a proper starting point for subjective refractionis mandatory, and this point will be achieved properlyby taking into account that autorefraction measurementsmust be reinterpreted by adding nearly 0.25 to 0.50 D tothe result. Then refraction can be performed from thisstarting point, and it should be guided by defocus curvemeasurement in case of doubt.Our study has certain limitations that must be taken

into consideration: First, the manifest refraction valueswe have dealt with are close to emmetropia. For surgeriesin which a diffractive trifocal IOL is implanted, qualityoutcomes are measured in terms of postoperative residualrefractive error (among other variables); that is why themean residual cylinder (J0 and J45) and the mean SEwere as low as 0.08 D and 0.01 D, respectively. Thislow-aberration scenario sets important limitations tothis type of comparison across measuring techniques(manifest refraction, autorefraction, wavefront, and 3-dimension wavefront topography) to assess the eye’spostoperative refractive power.Additional studies encompassing higher refractive error

cases and larger samples covering a wider range of refrac-tive error values would be required to confirm the findingsshown in the present paper.

9OBJECTIVE REFRACTION IN EYES WITH DIFFRACTIVE TRIFOCAL IOLS

Volume - Issue - - 2019

912913914915916917918919920921922923924925926927928929930931932933934935936937938939940941942943944945946947948949950951952953954955956957958959960961962963964965966967968969

970971972973974975976977978979980981982983984985986987988989990991992993994995996997998999

1000100110021003100410051006100710081009101010111012101310141015101610171018101910201021102210231024102510261027

FLA 5.6.0 DTD � JCRS10283_proof � 27 June 2019 � 12:50 am

REFERENCES1. Albarr�an-Diego C, Mu~noz G, Rohrweck S, García-L�azaro S, Albero JR. Val-

idity of automated refraction after segmented refractive multifocal intraoc-ular lens implantation. Int J Ophthalmol 2017; 10:1728–1733

2. van der Linden JW, Vrijman V, Al-Saady R, van der Meulen IJ,Mourits MP, Lapid-Gortzak R. Autorefraction versus subjective refrac-tion in a radially asymmetric multifocal intraocular lens. Acta Ophthalmol2014; 92:764–768

3. Bissen-Miyajima H, Minami K, Yoshino M, Nishimura M, Oki S. Autorefrac-tion after implantation of diffractive multifocal intraocular lenses. J. CataractRefract Surg 2010; 36:553–556

4. Mu~noz G, Albarr�an-Diego C, Sakla HF. Autorefraction after multifocal IOLs.Ophthalmology 2007; 114:2100

5. Goes FJ. Visual results following implantation of a refractive multifocal IOL inone eye and a diffractive multifocal IOL in the contralateral eye. J RefractSurg 2008; 24:300–305

6. Gatinel D, Pagnoulle C, Houbrechts Y, Gobin L. Design and qualification ofa diffractive trifocal optical profile for intraocular lenses. J Cataract RefractSurg 2011; 37:2060–2067

7. Pesudovs K, Weisinger HS. A comparison of autorefractor performance.Optom Vis Sci 2004; 81:554–558

8. Ogbuehi KC, Almaliki WH, AlQarni A, Osuagwu UL. Reliability and repro-ducibility of a handheld videorefractor. Optom Vis Sci 2015; 92:632–641

9. Wang X, Dong J,WuQ. Comparison of anterior corneal curvaturemeasure-ments using a Galilei dual Scheimpflug analyzer and Topcon auto kerato-refractometer. J Ophthalmol 2014; 2014:140628

10. McGinnigle S, Naroo SA, Eperjesi F. Evaluation of the auto-refraction func-tion of the Nidek OPD-Scan III. Clin Exp Optom 2014; 97:160–163

11. Mic�o V, Albarr�an-Diego C, Thibos L. Power vectors for the management ofastigmatism: from theoretical to clinical applications. In: Buckley R, ed,Astigmatism: Types, Diagnosis and Treatment Options. Hauppauge, NY,Nova Science Publishers, Inc., 2014; ISBN 10: 163321978X ISBN 13:9781633219786

12. Mu~noz G, Albarr�an-Diego C, Ferrer-Blasco T, García-L�azaro S. Power vec-tor analysis as an aid to correct a rotated Artiflex toric phakic intraocularlens. J Emmetropia 2010; 1:213–217

13. Campbell CE. Determining spherocylindrical correction using four differentwavefront error analysis methods: comparison to manifest refraction.J Refract Surg 2010; 26:881–890

14. McAlinden C, Khadka J, Pesudovs K. Statistical methods for conductingagreement (comparison of clinical tests) and precision (repeatability orreproducibility) studies in optometry and ophthalmology. Ophthalmic Phys-iol Opt 2011; 31:330–338

15. Kretz FT, Linz K, Mueller M, Gerl M, Koss MJ, Gerl RH, Auffarth GU. Rich-tiges Refraktionieren nach Implantation von Multifokal- und presbyopiekor-rigierenden Intraokularlinsen [Refraction after Implantation of Multifocal andPresbyopia-Correcting Intraocular Lenses]. Klinische Monatsblatter furAugenheilkunde 2015; 232:953–956

16. Albarr�an-Diego C, Mu~noz G, Ferrer-Blasco T. Subjective refraction beforeLASIK enhancement in bioptics procedures with refractive multifocal intra-ocular lenses. J Refract Surg 2011; 27:556–557

17. Mu~noz G, Albarr�an-Diego C, Sakla HF. Validity of autorefraction after cata-ract surgery with multifocal ReZoom intraocular lens implantation.J Cataract Refract Surg 2007; 33:1573–1578

18. Charman WN, Mont�es-Mic�o R, Radhakrishnan H. Problems in the mea-surement of wavefront aberration for eyes implanted with diffractive bifocaland multifocal intraocular lenses. J Refract Surg 2008; 24:280–286

19. Campbell CE. Wavefront measurements of diffractive and refractive multi-focal intraocular lenses in an artificial eye. J Refract Surg 2008; 24:308–311

20. Jendritza BB, Knorz MC, Morton S. Wavefront-guided excimer laservision correction after multifocal IOL implantation. J Refract Surg 2008;24:274–279

21. Siedlecki D, Ginis HS. On the longitudinal chromatic aberration of the intra-ocular lenses. Optom Vis Sci 2007; 84:984–989

22. Var�onC, Gil MA, Alba-Bueno F, CardonaG, Vega F,Mill�anMS, Buil JA. Ste-reo-acuity in patients implanted with multifocal intraocular lenses: is thechoice of stereotest relevant? Curr Eye Res 2014; 39:711–719

23. Charman WN, Jennings JA. Objective measurements of the longitudinalchromatic aberration of the human eye. Vision Res 1976; 16:999–1005

24. Strang NC, Gray LS, Winn B, Pugh JR. Clinical evaluation of patient toler-ance to autorefractor prescriptions. Clin Exp Optom 1998; 8:112–118

25. Delori FC, Pflibsen KP. Spectral reflectance of the human ocular fundus.Appl Opt 1989; 28:1061–1077

26. He JC, Marcos S, Burns SA. Comparison of cone directionality determinedby psychophysical and reflectometric techniques. J Opt Soc Am A OptImage Sci Vis 1999; 16:2363–2369

27. Martin J, Vasudevan B, Himebaugh N, Bradley A, Thibos L. Unbiasedestimation of refractive state of aberrated eyes. Vision Res 2011;51:1932–1940

28. Teel DF, Jacobs RJ, Copland J, Neal DR, Thibos LN. Differences betweenwavefront and subjective refraction for infrared light. Optom Vis Sci 2014;91:1158–1166

29. Vega F, Mill�an MS, Vila-Terricabras N, Alba-Bueno F. Visible versus near-infrared optical performance of diffractive multifocal intraocular lenses.Invest Ophthalmol Vis Sci 2015; 56:7345–7351

30. Schwiegerling J, DeHoog E. Problems testing diffractive intraocular lenseswith Shack-Hartmann sensors. Appl Opt 2010; 49:D62–D68

31. Mojzis P, Pe~na-García P, Liehneova I, Ziak P, Ali�o JL. Outcomes of a newdiffractive trifocal intraocular lens. J Cataract Refract Surg 2014; 40:60–69

32. Toto L, Carpineto P, Falconio G, Agnifili L, Di Nicola M, Mastropasqua A,Mastropasqua L. Comparative study of Acrysof ReSTORmultifocal intraoc-ular lensesC4.00 D andC3.00 D: visual performance and wavefront error.Clin Exp Optom 2013; 96:295–302

Disclosures: None of the authors has a financial or proprietary in-terest in any material or methods mentioned.

WHAT WAS KNOWN� Multifocal intraocular lenses (IOLs) can induce errors inobjective refraction resulting from their optical design.

� An accurate subjective refraction assessment is mandatoryto properly determine the refractive status of an eye im-planted with a trifocal IOL.

WHAT THIS PAPER ADDS� To our knowledge, this was the first comparison of objectiveand manifest subjective refraction in eyes with trifocal IOLs.

� Aberrometry-based objective refraction assessment was notmore accurate than traditional autorefractors in the presenceof a trifocal IOL.

10 OBJECTIVE REFRACTION IN EYES WITH DIFFRACTIVE TRIFOCAL IOLS

Volume - Issue - - 2019

102810291030103110321033103410351036103710381039104010411042104310441045104610471048104910501051105210531054105510561057105810591060106110621063106410651066106710681069107010711072107310741075107610771078107910801081108210831084108510861087108810891090109110921093109410951096

109710981099110011011102110311041105110611071108110911101111111211131114111511161117111811191120112111221123112411251126112711281129113011311132113311341135113611371138113911401141114211431144114511461147114811491150115111521153115411551156115711581159116011611162116311641165

FLA 5.6.0 DTD � JCRS10283_proof � 27 June 2019 � 12:50 am