LABORATORY SCIENCE

Modulation of cent

ral corneal thicknessby various riboflavin eyedrop compositions

in porcine corneasJan M. Vetter, MD, Stephan Brueckner, Marija Tubic-Grozdanis, PhD, Urs Voßmerb€aumer, MD,

Norbert Pfeiffer, MD, Sabine Kurz, MD

Q 2012 A

Published

SCRS an

by Elsev

PURPOSE: To evaluate the modulatory effect of various riboflavin 0.1% and 0.2% compositions onthe central corneal thickness (CCT) in fresh porcine corneas.

SETTING: Department of Ophthalmology, Johannes Gutenberg University of Mainz, Mainz, Germany.

DESIGN: Experimental study.

METHODS: The CCT in freshly enucleated porcine globes was measured by ultrasound pachymetrybefore and after (if applicable) deepithelialization and every 10 minutes thereafter during 120 min-utes of eyedrop application. In Groups 1 and 2 (controls), no eyedrops were applied. In Groups3 and 4, isotonic riboflavin eyedrops were used. In Groups 5 to 9, hypotonic riboflavin eyedropswere given. In Groups 10 and 11, preparations for transepithelial crosslinking were applied. InGroups 2 to 9, deepithelialization was performed. The final CCT in the groups was compared byanalysis of variance.

RESULTS: One hundred ten freshly enucleated porcine globes were used. The mean final CCT com-pared with preoperative values was 97% G 4% (SD) in Group 1, 91% G 4% in Group 2, 66% G5% in Group 3, 151%G 13% in Group 4, 65%G 2% in Group 5, 105%G 3% in Group 6, 120%G4% in Group 7, 130% G 4% in Group 8, 132% G 4% in Group 9, 114% G 2% in Group 10, and114%G 4% in Group 11. The differences between Group 1 and each of Groups 3, 4, 5, 7, 8, and 9were statistically significant (P<.05).

CONCLUSION: There was considerable variation in the final CCT as a result of varying riboflavin eye-drop compositions.

Financial Disclosure: No author has a financial or proprietary interest in any material or methodmentioned.

J Cataract Refract Surg 2012; 38:525–532 Q 2012 ASCRS and ESCRS

Corneal collagen crosslinking (CXL) using the photo-mediator riboflavin and ultraviolet-A (UVA) irradia-tion is a recently developed treatment for progressivecorneal thinning (ie, keratoconus, keratectasia afterlaser in situ keratomileusis, pellucid marginal degen-eration) and corneal ulcers.1,2 The parameters of thestandard treatment protocol for CXL (irradiance,wavelength, riboflavin concentration, duration ofUVA radiation) have been established with animalmodels.1,3–8 According to this protocol, the cornea iscentrally deepithelialized and then rinsed with a ribo-flavin 0.1%–dextran 20.0% solution until yellow stain-ing of riboflavin is observed in the anterior chamber.The eye is then irradiated by UVA light (3 mW/cm2,

d ESCRS

ier Inc.

370 nm, 30 minutes). This activates the intracornealriboflavin, generating reactive oxygen species andincreasing the biomechanical1,2,7,8 and biochemical6

stability of the cornea by inducing additional cross-links between the collagen fibers. Several prospectivestudies indicate clinical effectiveness of CXL.1,9,10

The anterior 300 mm of the cornea is depopulated ofkeratocytes because of the cytotoxic effect of UVA irra-diation and free radicals.4,5,11,12 As the UVA light is ab-sorbed by the precorneal and intracorneal riboflavin,the UVA intensity is reduced to below the cytotoxicthreshold in deeper parts of the cornea.6 It is thereforerecommended to maintain at least 400 mm of cornealthickness during the treatment to protect the

0886-3350/$ - see front matter 525doi:10.1016/j.jcrs.2011.09.045

526 LABORATORY SCIENCE: MODULATION OF CORNEAL THICKNESS BY RIBOFLAVIN

endothelial cell layer. Using the standard riboflavineyedrop composition with 20.0% dextran, marked in-traoperative corneal thinning is observed.13

Corneal thickness can alter with changes in hydra-tion, pH, and ionic composition.14,15 An intact epithe-lium and the ion pumping endothelium is crucial formaintaining physiologic corneal dehydration. Deepi-thelialization and subsequent application of liquidscan increase hydration and thickness due to cornealswelling pressure.14,16 Application of the polyglucosebiopolymer dextran T500 (molecular weight 500 kDa)at high concentrations may cause oncotic pressure-related dehydration of the cornea and reduce cornealthickness.14,17 Dextran possesses a high affinity forwater because of its abundant hydrophilic hydroxylgroups. Unsurprisingly, marked intraoperative thin-ning of the deepithelialized cornea has been describedduring application of riboflavin 0.1%–dextran 20.0%solution.13

In this study, we evaluated the effect of variousriboflavin eyedrop compositions on corneal thicknessto find a composition with a moderate swelling effectthat could be used continuously during the entireCXL treatment process in cases of advancedkeratectasia.

MATERIALS AND METHODS

Experimental Setup

One hundred ten freshly enucleated porcine globes wereobtained from a local slaughterhouse and the experimentswere performed within the facility of the abattoir. In anagreement with the local staff, the eyes could be removed be-fore the pigs’ scalding treatment to prevent corneal damageby heat or scratching. This allowed for a very short death-to-experiment time (mean 19.5 minutes G 5.2 [SD]). The eyeswere divided into 11 groups (Table 1). Eyes with

Submitted: June 9, 2011.Final revision submitted: September 2, 2011.Accepted: September 5, 2011.

From the Department of Ophthalmology (Vetter, Brueckner,Voßmerb€aumer, Pfeiffer, Kurz) and the Department of Pharmacy(Tubic-Grozdanis), University Medical Center, Johannes GutenbergUniversity Mainz, Mainz, Germany.

Melissa Faust, PhD, Bettina Sprang, Irina Vachtel, and ThomasMeyer contributed to the construction of the measurement appara-tus. Hannah Walz-Jung assisted in the pharmaceutical productionof the riboflavin eyedrops. This manuscript contains parts of theyet unpublished doctoral thesis of Stephan Br€uckner.

Corresponding author: Jan M. Vetter, MD, Department of Ophthal-mology, University Medical Center of the Johannes GutenbergUniversity Mainz, Langenbeckstrasse 1, 55131 Mainz, Germany.E-mail: jan.vetter@unimedizin-mainz.de.

J CATARACT REFRACT SURG -

macroscopic epithelial damage or corneal haziness were ex-cluded from the study. During each experiment, 8 eyes wereplaced in holders on a rotating platform (Figure 1). The plat-form was fitted with a latching position at the side of eachholder so that each globe could be rotated to an exact mea-suring position. This allowed repeated measurements atthe exact same spot on each cornea. At the side of the plat-form, the measuring head of a pachymeter (Tomey AL2000, Tomey Corp.) was fixed to a flexible arm that couldbe lowered onto a globe located in the measuring position.The 0.9% sodium chloride infusion set was positioned 27.2cm higher than the platform (from cornea to liquid level).It was connected via Y-tubes and 21-gauge needles into thevitreous cavity of each globe, maintaining an intraocularpressure of 20 mm Hg. A thermohygrometer was used(TFD 128 Temperatur Feuchte Datenlogger USB, ELV Elek-tronik AG) for parallel measurements of air temperatureand humidity.

Treatment

The globes were placed into the holders and fixed. Cannu-lation needles were then inserted into the vitreous cavities atthe height of the equator with a slight outflow of saline solu-tion to prevent blockage of the needles by the vitreous body.During a measurement, the central corneal thickness (CCT)was obtained 5 times by ultrasound pachymetry with theplatform being fixed by the latch; the mean pachymetryvalue was used. The platform was then rotated to the nextlatching position until all 8 globesweremeasured. This seriesof measurements was performed before and after deepitheli-alization and every 10 minutes for 120 minutes during theperiod of eyedrop application. The duration of 2 hours waschosen to reach a plateau in the development of CCT andto obtain the maximum swelling or deswelling effect.

Groups

Each treatment group consisted of 10 eyes (Table 1).Groups 1 and 2 served as controls; no eyedrops were ap-plied, and deepithelialization was performed in Group 2but not in Group 1. In Group 3, Medio Cross eyedropswere applied. A solution of hydroxypropyl methylcellulose(HPMC)–based isoosmolar eyedrops in use at the UniversityHospital Dresden was tested in Group 4 (HPMC

Figure 1. Experimental setup for parallel pachymetric measure-ments of the CCT in 8 porcine globes. The thermohygrometer isshown at the bottom left part of the photograph (A).

VOL 38, MARCH 2012

Table 1. Treatment groups with details on deepithelialization, compositions used, drop frequency, duration of the treatment phases, andaim of measurements.

Treatment (Min)

Group DE Composition Name Drop Frequency Start End Simulating

1 No d d d d Control2 Yes d d d d Control3 Yes Medio Cross 1x/3 min 0 120 Tx w/ isoosmolar riboflavin4 Yes HPMC isoosmolar 1x/3 min 0 120 Tx w/ isoosmolar riboflavin5 Yes Medio Cross 1x/3 min 0 30 Tx w/ hypoosmolar riboflavin5 Yes Medio Cross sine 3x/1 min 30 40 Tx w/ hypoosmolar riboflavin5 Yes Medio Cross 1x/3 min 40 120 Tx w/ hypoosmolar riboflavin6 Yes Riboflavin dextran 10.0% 1x/3 min 0 120 Tx w/ hypoosmolar riboflavin7 Yes HPMC hypoosmolar 1x/3 min 0 120 Tx w/ hypoosmolar riboflavin8 Yes Sodium chloride 0.9% 3x/1 min 0 120 Tx w/ hypoosmolar riboflavin9 Yes Riboflavin 0.1% 3x/1 min 0 120 Tx w/ hypoosmolar riboflavin10 No Proparakain-POS 0.5% 3x/1 min 0 30 Transepithelial crosslinking10 No Proparakain-POS 0.5% C Medio Cross 1x/3 min 30 120 Transepithelial crosslinking11 No Ricrolin TE 1x/3 min 0 120 Transepithelial crosslinking

DE Z deepithelialization; HPMC Z of hydroxypropyl methylcellulose; Tx Z Treatment ; w/ Z with

527LABORATORY SCIENCE: MODULATION OF CORNEAL THICKNESS BY RIBOFLAVIN

isoosmolar). The Medio Cross eyedrops were again appliedin Group 5, interrupted by a 10-minute interval of MedioCross sine eyedrops. This design was chosen to simulatean alternative treatment protocol for thin corneas.18 InGroup6, a riboflavin 0.1%–dextran 10.0%mixture was tested. A hy-poosmolar composition (HPMC hypoosmolar) also in use atthe University Hospital Dresden was used in Group 7. InGroup 8, sodium chloride 0.9% was used. In Group 9, ribo-flavin 0.1% eyedrops without dextran were applied. Prepa-rations for transepithelial crosslinking were tested inGroups 10 and 11 using benzalkonium chloride (Group 10)and sodium edetate (Ricrolin TE, Group 11) to increaseepithelial permeability. Tables 1 and 2 show the details ondeepithelialization, drop frequency, and active and inactiveingredients.

Osmolality and pH

The pH of the eyedrops was measured using a pH meter(Microprocessor pH meter 210, Hanna Instruments), and os-molality was determined on a 50 mL sample in a test tube byfreezing-point depression (Osmomat 030, Gonotec GmbH).Measurements were performed on 3 samples of each solu-tion and averaged.

Statistical Analysis

At each time point, the relative CCT was expressed asa percentage of the CCT before deepithelialization. Thisvalue combines the effect of deepithelialization and swell-ing/deswelling on the CCT. It provides the surgeon witha factor that can be multiplied with the thinnest cornealthickness value of the preoperative pachymetry map to ob-tain the thickness at the thinnest point during treatment,which is difficult to measure accurately by ultrasound pa-chymetry. For each group, a curve was generated using themeans of these values and their standard deviation. The low-est CCT and the highest CCT during a presumed

J CATARACT REFRACT SURG -

crosslinking treatment (between 30 minutes and 60 minutesafter the start of riboflavin application) and the final CCTwere obtained again as percentage values. Themeans of finalrelative CCT (%) after 120 minutes in each group were com-pared with those in Group 1 using analysis of variance anda Bonferroni-Holm post hoc test for multiple comparisons.The final relative CCT was correlated with osmolality onone hand and the concentration of dextran on the otherhand using regression analysis. A P value below 0.05 wasconsidered statistically significant.

RESULTS

Figure 2 and Table 3 show the mean CCT as the per-centage of the CCT before treatment. In the controlgroups (Groups 1 and 2), the final relative CCT wasstable during the eyedrop application. In Groups 3and 5 (Medio Cross; Medio Cross C Medio Crosssine), marked thinning was observed. The thickeningeffect of the 10-minute Medio Cross sine treatment inGroup 5 seemed to be reversed 10 minutes later(Figure 2). The highest increase in CCT was in Group4. The eyedrops using sodium chloride 0.9% as an in-active ingredient caused similar swelling of the corneain Group 8 and Group 9). The riboflavin 0.1%–dextran10.0% composition in Group 6 led to a mild increase inCCT. The corneas in Group 7 (HPMC hypoosmolar)hadmarked swelling. Both transepithelial crosslinkinggroups (10 and 11) had amoderate increase in the finalrelative CCT. The changes in the final relative CCT inGroups 3, 4, 5, 7, 8, and 9were statistically significantlydifferent from those in Group 1 (Figure 3).

The mean surrounding temperature during sono-graphic pachymetry measurements was 18.7� G

VOL 38, MARCH 2012

Table 2. Active and inactive ingredients of the compositions used and their pH and osmolality.

Composition Name Group Active ingredients Inactive Ingredients Mean pHMean Osmolality(mOsmol/L)

Medio Cross 3, 5, 10 Riboflavin phosphate 0.1% Dextran T 500 20%, disodiumhydrogen phosphate, sodiumdihydrogen phosphate,purified water q.s.

6.67 233

HPMC isoosmolar 4 Riboflavin phosphatesodium 0.127%

Methyl hydroxypropylcellulose 1.5 g, sodiumchloride 1.1 g, aqua adiniectabilia q.s. ad 100.0g

7.49 307

Medio Cross sine 5 Riboflavin phosphate O0.1% Disodium hydrogenphosphate, sodiumdihydrogen phosphate,purified water q.s.

6.95 54

Riboflavin dextran 10.0% 6 Riboflavin phosphatesodium 0.127%

Dextran T500 10.0 g, sodiumchloride 0.9 g, aqua adiniectabilia q.s. ad 100.0 g

5.72 337

HPMC hypoosmolar 7 Riboflavin phosphatesodium 0.254%

Methyl hydroxypropylcellulose 0.5 g, sodium chloride0.7 g, aqua ad iniectabilia q.s.ad 100.0 g

6.44 275

NaCl 0.9% 8 – Sodium chloride 0.9% 5.51 308Riboflavin 0.1% 9 Riboflavin phosphate

sodium 0.127%Sodium chloride 0.9% 5.94 311

Proparakain-POS 0.5% 10 Proxymetacainehydrochloride 5.0 mg/mL

Benzalkonium chloride,sodium chloride, sodiumedetate, sodium metabisulfite,purified water q.s.

3.94 297

Ricrolin TE 11 Riboflavin phosphate 0.1% Dextran T500 15.0 g, sodiumedetate, tromethamine,bihydrate sodium phosphatemonobasic, bihydrate sodiumphosphate dibasic, aqua adiniectabilia q.s. ad 100.0 g

6.97 160

aqua ad iniectabilia Z water for injection; q.s. Z quantum sufficit (sufficient quantity)

528 LABORATORY SCIENCE: MODULATION OF CORNEAL THICKNESS BY RIBOFLAVIN

3.5�C and the mean humidity, 61.5% G 8.7%. Theosmolality of each composition ranged from 54 to337 mOsmol/L and the pH, from 3.94 to 7.49(Table 2). No correlation between the osmolality ofthe composition and the swelling behavior of thetreated corneas was observed (Figure 4) (Pearson co-efficient (r) Z 0.456, PZ.256). An inverted correla-tion was verified between the dextran concentrationand the swelling effect (Figure 5) (r Z �0.950,PZ.004).

DISCUSSION

To our knowledge, this is the first experimental studyto compare different compositions of riboflavin eye-drop solutions in their swelling effect on fresh post-mortem porcine corneas.

J CATARACT REFRACT SURG -

In the present study, we found a large variation inCCT regulation by the different eyedrop compositions(between 65% and 151% final relative CCT). In Group6 (riboflavin 0.1%–dextran 10.0%) and Groups 10 and11 (both transepithelial crosslinking), only a mild in-crease in CCT was observed. The mild swelling char-acteristics, in particular, of the eyedrops in Group 6,seemed to be promising because the intraoperativemean CCT appeared to be close to the preoperativemean CCT. Furthermore, transepithelial crosslinkingcompositions led to moderate swelling of the cornea.It is unclear, however, whether they penetrated intothe cornea because a decrease in corneal thicknesswould be expected due to the 15.0% (Ricrolin TE)and 20.0% (Medio Cross) dextran concentrations. Fur-ther studies are needed to evaluate the clinical effec-tiveness of CXL using these eyedrop compositions.

VOL 38, MARCH 2012

Figure 2. Mean CCT in percentages of the CCT before treatmentin 110 porcine globes after deepithelialization and application of var-ious eyedrops.Nodeepithelializationwasperformed inGroups1, 10,and 11, and no eyedrops were applied in Groups 1 and 2. Values areG1 SD (CCTZ central corneal thickness; DEZ deepithelialization).

529LABORATORY SCIENCE: MODULATION OF CORNEAL THICKNESS BY RIBOFLAVIN

Endothelial damage is a serious complication afterCXL. It can be caused by high UVA irradiation inten-sity, biochemical stress by free radicals, or both.6 Thethreshold UVA intensity for endothelial damage hasbeen estimated as 0.36 mW/cm2, which is only 12%of the superficial UVA intensity (3 mW/cm2).6 Afterthe standard protocol of CXL (riboflavin 0.1%, surfaceUVA dose 3 mW/cm2) and using the known absorp-tion coefficient of 53 cm�1,19 this threshold wasreached at a depth of 330 mm into the cornea. It hasbeen repeatedly emphasized that a minimum CCT of400 mm should be maintained during UVA treatment.However, Kymionis et al.,13 Wollensak et al.,20 andHersh et al.A found marked thinning of the cornea

Table 3. Mean of absolute CCT before and after deepithelialization, meaCCT after 120 minutes and P values (10 in each group).

Group

Mean Absolute CCT (mm) Mean Minim

Pre DE Post DE Between

1 785 G 55 d

2 789 G 55 731 G 573 781 G 36 718 G 384 778 G 54 713 G 555 731 G 55 672 G 566 799 G 39 732 G 387 765 G 45 700 G 368 817 G 64 747 G 729 746 G 29 679 G 32

10 762 G 30 d

11 756 G 44 d

CCT Z central corneal thickness; DE Z deepithelializationMeans G SD*Proposed ultraviolet treatment†Compared with Group 1

J CATARACT REFRACT SURG -

after deepithelialization using the standard riboflavin0.1%–dextran 20.0% eyedrops. In Kymionis et al.’sstudy,13 corneal thickness reached a mean of 325 mm(�22% compared with CCT after deepithelialization),theoretically putting the endothelium at considerablerisk. Fortunately, no mean endothelial damage oc-curred, although no single values of preoperativeand postoperative endothelial cell count in eyes withthe thinnest corneas were given. Taking the above-mentioned 22% corneal thinning into account, theCCTwould have to be above 511 mmafter and roughly560 mm before deepithelialization to abide fully withthe widely accepted 400 mm intraoperative minimumCCTwhen using the standard riboflavin 0.1%–dextran20.0% eyedrops. This, however, would exclude mostkeratoconus patients from treatment.

Several alternative protocols for treatment of thincorneas have been proposed. In a comment on Kymio-nis et al.’ study,13 Tahzib et al.21 recommended closureof the eyelids during the 30 minutes of riboflavinapplication and reported no corneal thinning in 3 pa-tients as a result of reduced evaporation. However, itremained unclear which exact riboflavin eyedroppreparation they used and whether the riboflavinfilm was wiped away by the eyelids and diluted bytear production. Furthermore, with the viscous ribofla-vin 0.1%–dextran 20.0% composition, evaporation ofcorneal fluid into the air is improbable because 1 dropcovers the corneal surface for more than 20 minutes.22

Recently, a modified treatment protocol was proposedfor cases of advanced keratoconus with thin corneasusing a hypoosmolar formula of the riboflavin 0.1%eyedrops.18 In this protocol, the CCT is measured

n relative CCT between 30 minutes and 60 minutes, mean relative

um Relative CCT (%) Mean Relative CCT (%)

30 Min and 60 Min* After 120 Min P Value†

98 G 3 97 G 4 d

91 G 3 91 G 4 1.00070 G 5 66 G 5 O.001

119 G 7 151 G 13 O.00170 G 5 65 G 2 O.001

102 G 2 105 G 3 1.000115 G 4 120 G 4 .003110 G 3 130 G 4 O.001119 G 2 132 G 4 O.001103 G 2 114 G 2 1.000107 G 2 114 G 4 .113

VOL 38, MARCH 2012

Figure 3.Mean CCT after 120minutes of riboflavin eyedrop applica-tion in percentages of the initial CCT (before deepithelialization,where applicable) (10 per group) (* Z statistically significant com-pared with Group 1 [P!.05]; CCT Z central corneal thickness; DEZ deepithelialization; HPMC Z hydroxypropyl methylcellulose).

Figure 4. Correlation between the osmolality of various riboflavineyedrop compositions and the change in CCT after 120 minutes ofeyedrop application (CCT Z central corneal thickness; DE Zdeepithelialization).

530 LABORATORY SCIENCE: MODULATION OF CORNEAL THICKNESS BY RIBOFLAVIN

after 30 minutes of eyedrop application. If it is below400 mm, the CCT is increased by the use of dextran-free riboflavin eyedrops. However, our results showthat this effect is reversed after 10minutes of riboflavin0.1%–dextran 20.0% eyedrop application. Therefore,continuous CCT monitoring seems necessary whenusing the modified treatment protocol. Raiskup andSpoerl23 propose the continuous use of a hypoosmolarriboflavin solution in eyes with keratoconus and thincorneas. Unfortunately, they also did not specify theexact composition of the eyedrops they used. Our

Figure 5. Correlation between the concentration of dextran in ribo-flavin eyedrop compositions and the change in CCT after 120 min-utes of eyedrop application (CCT Z central corneal thickness; DEZ deepithelialization).

J CATARACT REFRACT SURG -

data and their data support the need for continuoususe of moderately swelling eyedrops during the entiretreatment. However, we did not find a correlation be-tween osmolality and swelling behavior. It is thereforeimportant to specify all the ingredients of a testedcomposition.

Although our experiments were performed in pigeyes, the results are comparable to data obtainedfrom human eyes in vivo.When applyingMedio Crosseyedrops, we found a marked decrease in cornealthickness by amean 28%G 6% (CCT after 120minutescompared with CCT at 0 minutes). Kymionis et al.13

found a 22% reduction in corneal thickness in humaneyes. Wollensak et al.22 report a 9.1% decrease in cor-neal thickness in human postmortem eyes anda 7.3% decrease in pig eyes using riboflavin 0.1%–dex-tran 20.0% comparing pachymetry values after de-bridement and after 30 minutes of eyedropapplication. We found a mean CCT decrease of12.7% G 5.0% using a riboflavin 0.1%–dextran 20.0%solution in porcine eyes.

It is known that the cornea has an inert swellingpressure.14 This means that the corneal stroma hasthe tendency to increase its volume in an isooncotic en-vironment. In vivo, the active transport of water out ofthe cornea by the endothelial cell layer works againstthis swelling pressure while the intact epithelium sealsoff the outer side. The integrity of both layers is there-fore important for maintaining corneal thickness andtransparency. If deepithelialization is performed, thecornea can lose fluid at the surface by drying if ex-posed to air. It can also dehydrate if exposed to solu-tions with an oncotic pressure higher than the

VOL 38, MARCH 2012

531LABORATORY SCIENCE: MODULATION OF CORNEAL THICKNESS BY RIBOFLAVIN

swelling pressure. Hypertonicity and hypotonicity ofsolutions (using the semipermeable cell membrane asa barrier) do not have a significant effect on stromalthickness because only 2% to 3% of the corneal stromavolume consists of cells.24 Thus, we did not find a cor-relation between osmolality and corneal thickness af-ter eyedrop application. Dextran T500 is a 500 kDapolyglucose biopolymer with a high affinity to waterbecause of its abundant hydrophilic hydroxyl groups.Its oncotic effect is used in a 5.0% concentration in or-gan culture media to deswell corneal transplants toa thickness close to physiologic values.17 It is there-fore logical that an inverted correlation between dex-tran concentration and corneal thickness exists, asshown in Figure 5, and that a 20.0% dextran concen-tration leads to deswelling beyond the physiologicallevel, as we and others have observed. This effect,however, has to be critically reflected in the clinicalcontext of CXL for progressive corneal thinning inthin corneas.

The present study has limitations. The measure-ments were performed with ultrasound pachymetry,and reproducibility was obtained by using a latchingposition on a rotating platform. However, small varia-tions in measurement location and an influence of thiscontact method on the actual measurement valuescould not be ruled out. Our control measurements inGroup 1 showed a fairly stable corneal thicknesswhen no deepithelialization was applied. The eyeswere postmortem eyes excised from pigs shortly afterdeath. However, epithelial permeability of even veryfresh porcine eyes may differ from that of intact hu-man epithelium. Further experiments might be neces-sary using in vivo eyes (for example rabbit eyes) orpostmortem human eyes or in the clinical context.This study did not address the biomechanical andcrosslinking efficacy of each composition or the thick-ness of the precorneal eyedrop film. This is an impor-tant factor for some compositions and should befurther investigated in the future.

In conclusion, we showed the swelling and deswel-ling effect of various 0.1% riboflavin compositions thatare or should be in use in the clinical context. We alsosimulated the standard protocols and placed emphasison critical points. Further studies are needed of modi-fied riboflavin eyedrop compositions (eg, with a re-duced dextran concentration) to satisfy surgeon andpatient needs.

REFERENCES1. Wollensak G, Spoerl E, Seiler T. Riboflavin/ultraviolet-A-

induced collagen crosslinking for the treatment of keratoconus.

Am J Ophthalmol 2003; 135:620–627

J CATARACT REFRACT SURG -

2. Spoerl E, Huhle M, Seiler T. Induction of cross-links in corneal

tissue. Exp Eye Res 1998; 66:97–103

3. WollensakG, Spoerl E,WilschM, Seiler T. Keratocyte apoptosis

after corneal collagen cross-linking using riboflavin/UVA treat-

ment. Cornea 2004; 23:43–49

4. Wollensak G, Spoerl E, Reber F, Seiler T. Keratocyte cytotoxic-

ity of riboflavin/UVA-treatment in vitro. Eye 2004; 18:718–722.

Available at: http://www.nature.com/eye/journal/v18/n7/pdf/

6700751a.pdf. Accessed September 27, 2011

5. Wollensak G, Sp€orl E, Reber F, Pillunat L, Funk R. Corneal

endothelial cytotoxicity of riboflavin/UVA treatment in vitro. Oph-

thalmic Res 2003; 35:324–328

6. WollensakG, Spoerl E,WilschM, Seiler T. Endothelial cell dam-

age after riboflavin-ultraviolet-A treatment in the rabbit.

J Cataract Refract Surg 2003; 29:1786–1790

7. Wollensak G, Sp€orl E, Seiler T. Behandlung von Keratoconus

durch Kollagenvernetzung [Treatment of keratoconus by colla-

gen cross-linking]. Ophthalmologe 2003; 100:44–49

8. Wollensak G, Spoerl E, Seiler T. Stress-strain measurements of

human and porcine corneas after riboflavin-ultraviolet-

A-induced cross-linking. J Cataract Refract Surg 2003; 29:

1780–1785

9. Caporossi A, Mazzotta C, Baiocchi S, Caporossi T. Long-term

results of riboflavin ultraviolet A corneal collagen cross-linking

for keratoconus in Italy: the Siena Eye Cross Study. Am J Oph-

thalmol 2010; 149:585–593

10. Wittig-Silva C, Whiting M, Lamoureux E, Lindsay RG,

Sullivan LJ, SnibsonGR.A randomized controlled trial of corneal

collagen cross-linking in progressive keratoconus: preliminary

results. J Refract Surg 2008; 24:S720–S725

11. Wollensak G. Crosslinking treatment of progressive keratoco-

nus: new hope. Curr Opin Ophthalmol 2006; 17:356–360

12. Spoerl E, Mrochen M, Sliney D, Trokel S, Seiler T. Safety of

UVA-riboflavin cross-linking of the cornea. Cornea 2007; 26:

385–389

13. Kymionis GD, Kounis GA, Portaliou DM, Grentzelos MA,

Karavitaki AE, Coskunseven E, Jankov MR, Pallikaris IG. Intra-

operative pachymetric measurements during corneal collagen

cross-linkingwith riboflavin and ultraviolet A irradiation.Ophthal-

mology 2009; 116:2336–2339

14. Dohlman CH, Hedbys BO, Mishima S. The swelling pressure

of the corneal stroma. Invest Ophthalmol 1962; 1:158–162.

Available at: http://www.iovs.org/cgi/reprint/1/2/158 Accessed

September 27, 2011

15. Doughty MJ. Re-assessment of the potential impact of physio-

logically relevant pH changes on the hydration properties of

the isolated mammalian corneal stroma. Biochim Biophys Acta

1999; 1472:99–106

16. Doughty MJ. Swelling of the collagen-keratocyte matrix of the

bovine corneal stroma ex vivo in various solutions and its rela-

tionship to tissue thickness. Tissue Cell 2000; 32:478–493

17. Pels E, Schuchard Y. The effects of high molecular weight dex-

tran on the preservation of human corneas. Cornea 1984-1985;

3:219–227

18. Hafezi F, Mrochen M, Iseli HP, Seiler T. Collagen crosslinking

with ultraviolet-A and hypoosmolar riboflavin solution in thin cor-

neas. J Cataract Refract Surg 2009; 35:621–624

19. Sp€orl E, Schreiber J, Hellmund K, Seiler T, Knuschke P. Unter-

suchungen zur Verfestigung der Hornhaut am Kaninchen [Stud-

ies on the stabilization of the cornea in rabbits]. Ophthalmologe

2000; 97:203–206

20. Wollensak G, Aurich H, Wirbelauer C, Pham D-T. Potential use

of riboflavin/UVA cross-linking in bullous keratopathy. Ophthal-

mic Res 2009; 41:114–117

VOL 38, MARCH 2012

532 LABORATORY SCIENCE: MODULATION OF CORNEAL THICKNESS BY RIBOFLAVIN

21. TahzibNG,SoetersN, Van der Lelij A. Pachymetry during cross-

linking [letter]. Ophthalmology 2010; 117:2041–2041.e1; reply

by GD Kymionis, DM Portaliou, 2041�2042

22. Wollensak G, Aurich H, Wirbelauer C, Sel S. Significance of the

riboflavin film in corneal collagen crosslinking. J CataractRefract

Surg 2010; 36:114–120

23. Raiskup F, Spoerl E. Corneal cross-linking with hypo-osmolar ri-

boflavin solution in thin keratoconic corneas. Am J Ophthalmol

2011; 152:28–32

24. Otori T. Electrolyte content of the rabbit corneal stroma. Exp Eye

Res 1967; 6:356–367

OTHER CITED MATERIALA. Hersh PS, Fry KL. Corneal thickness changes during corneal

collagen crosslinking. IOVS 2009; 50: ARVO E-Abstract 5490

J CATARACT REFRACT SURG -

Available at: http://abstracts.iovs.org//cgi/content/abstract/50/5/

5490?sidZ2d8274d9-53ce-4255-92c6-4ddd2c6810b4. Accessed

September 27, 2011

VO

L 38, MARCH 2012

First author:Jan M. Vetter, MD

Department of Ophthalmology,University Medical Center of theJohannes Gutenberg University Mainz,Mainz, Germany