evaluation of corneal hysteresis and corneal resistance factor after corneal cross-linking for...

CORNEA

Evaluation of corneal hysteresis and corneal resistance factorafter corneal cross-linking for keratoconus

Maria Gkika & Georgios Labiris &

Athanassios Giarmoukakis & Anna Koutsogianni &Vassilios Kozobolis

Received: 31 May 2011 /Revised: 29 November 2011 /Accepted: 2 December 2011 /Published online: 22 December 2011# Springer-Verlag 2011

AbstractBackground To evaluate corneal hysteresis (CH) and cornealresistance factor (CRF) in keratoconic (KC) eyes before andafter corneal collagen cross-linking (CXL). Furthermore, todetermine potential correlations with a series of corneal anddemographic factors.Methods The study consisted of 50 KC eyes that underwentCXL. CH and CRF were measured by the ocular responseanalyzer (ORA). Correlations were attempted with uncor-rected visual acuity (UVA), best spectacle-corrected visualacuity (BSCVA), central corneal thickness (CCT), meankeratometry (Km), astigmatism (Astig.), residual astigma-tism, age, and gender. Fifty non-KC eyes served as controls.Results CH andCRF (mean±SD) for non-KC eyes were 10.1±1.9 mmHg and 9.7±2.4 mmHg respectively, while for KC eyespreoperatively they were 8.2±1.4 mmHg (p00.007) and 7.4±2.3 mmHg (p00.01) respectively. Non-significant differenceswere detected between preoperative and postoperative CH and

CRF measurements in KC eyes (p00.518 and p00.479 respec-tively). Significant correlations were found between ORAparameters and BSCVA, CCT, Km, Astig. and residualastigmatism.Conclusions ORA parameters demonstrate significant dif-ferences between KC and non-KC eyes. Both CH and CRFpresent significant correlations with visual acuity and cornealparameters. CXL exerts a non-significant impact on ORAmeasurements.

Keywords Keratoconus . Corneal collagen cross-linking .

CH . CRF. ORA

Introduction

It is a truism that ocular surface health depends heavily on thepreservation of the biomechanical properties of the cornea. Cer-tain surgical procedures, and a series of corneal diseases, areassociatedwith substantial changes in the corneal tissue structure,and prospectively with altered biomechanical properties [1–5].

Until recently the evaluation of corneal biomechanical prop-erties was encountered, mainly, in research settings. The devel-opment of the Ocular Response Analyzer (ORA; ReichertOphthalmic Instruments, Buffalo, NY, USA), introduced asimple and reliable way for the assessment of a series ofbiomechanical factors of the cornea in clinical settings as well[6]. These factors are: a) corneal hysteresis (CH), and b) cornealresistance factor (CRF).

Keratoconus is a non-inflammatory disease which affectsmainly the central cornea and leads to progressive thinning[7]. In the majority of cases, there is progressive deteriora-tion of the visual acuity due to irregular astigmatism andcorneal scarring. Keratoconic eyes also have altered cornealbiomechanics [8]. According to international bibliography,

The study was performed with informed consent of all participants,following all the guidelines for experimental investigations required bythe Ethics Committee of Democritus University of Thrace and inaccordance with the ethical standards laid down in the 1964Declaration of Helsinki. The authors have full control of all primarydata, and they agree to allow Graefe’s Archive for Clinical andExperimental Ophthalmology to review their data upon request. Theauthors have no proprietary or commercial interest in any materialsdiscussed in this article. No financial support was received for thisstudy. The paper has not been presented at any meeting. Themanuscript has not been previously rejected or evaluated in any formby any other journal.

M. Gkika (*) :G. Labiris :A. Giarmoukakis :A. Koutsogianni :V. KozobolisEye Institute of Thrace and Department of Ophthalmology,Democritus University of Thrace,Dragana, Alexandroupolis, Greece PC: 68100e-mail: [email protected]

Graefes Arch Clin Exp Ophthalmol (2012) 250:565–573DOI 10.1007/s00417-011-1897-0

they seem to be more elastic and less rigid than normal eyes.Corneal collagen cross-linking (CXL) induced by riboflavinand ultraviolet-A irradiation has recently been introduced as atherapeutic option for the treatment of keratoconus [9, 10].Riboflavin serves as a photosensitizer that is activated byUVA light, and oxygen radicals production leads to the devel-opment of strong chemical bonds between collagen fibrils,thereby stiffening the cornea. The increased corneal stiffnessinduced by CXL can prevent progression of keratoconus[10–12]. In order to assess the risk profile of this new treat-ment, it is important to assess how this surgical procedureresults in changes in the corneal tissue structure and affects thecentral corneal thickness (CCT) and curvature of the cornea,and how it may alter corneal biomechanics.

Within this context, this study attempts to compare CHand CRF in a series of keratoconic eyes before and afterCXL. Furthermore, among the objectives of the study was toreveal correlations between ORA parameters, visual acuity,and different corneal parameters.

Material and methods

Setting

This is a prospective, non-randomized trial. The study adheredto the tenets of the Declaration of Helsinki, and written in-formed consent was given by all participants. The institutionalreview board of the Democritus University of Thrace (DUTH)approved the protocol, and the study was conducted at the EyeInstitute of Thrace (ΕΙΤ) in Alexandroupolis, Greece duringthe period between April 2008 and April 2010. EΙΤ is a

University research institute focusing primarily on the con-ditions of the anterior segment of the eye.

Participants

Participants were recruited from the Outpatients Corneaservice of the EIT in a consecutive if eligible basis. Fiftyeyes of 30 patients with keratoconus were recruited for thesake of the study, and formed the keratoconus group (KG).Eligible subjects for the keratoconus group had to presentprogressive keratoconus in consecutive corneal topographies,changes in refractive power, and deterioration of the visualacuity within a period of 2 years. Exclusion criteria includedglaucoma, glaucoma suspicion, and intraocular pressure (IOP)lowering drugs administration. Further to glaucoma, exclu-sion criteria included CCT less than 400 μm, K-readings morethan 60D, a history of herpetic keratitis, corneal scarring,severe eye dryness, pregnancy or nursing, current cornealinfection, or underlying autoimmune disease.

Fifty eyes of 50 non-keratoconic, age-matched individualswho visited our outpatient service formed the control group(CG). Further to keratoconus, and a spherical equivalent errorabove 3D, the same exclusion criteria applied to the controlgroup members as well. All participants wearing contactlenses were instructed to discontinue contact lens wear at leasta month before measurements.

Data collection

CRF and CH parameters were obtained while the patientwas sitting on a chair in front of the ORA device. Uponsuccessful fixation of the patient’s eye on a red blinking target,

Table 1 Demographic data and examination results of study cohort

Total cohort (n0100 eyes; normal, n050; keratoconus, n050) Control group Keratoconus group

Mean SD Range Mean SD Range P

Subjects (no.) 50 30

Male gender (no.) 18 18

Female gender (no.) 32 12

Age (years) 33.3 7.8 19-50 31.1 9.6 16-59 0.207

UVA 0.3 0.1 0.05–0.9 0.2 0.3 0.05–0.9 <0.001*

BSCVA 1.0 0.1 0.7–1.2 0.7 0.3 0.2–1.0 <0.001*

CCT (μm) 539.9 34.4 479–633 449.5 38.2 400–532 <0.001*

Average keratometry (diopters) 43.1 1.6 38.7–46.1 49.2 4.2 38.1–58.2 <0.001*

Corneal astigmatism (diopters) 1.2 0.8 0.0–4.0 3.5 1.7 0.8–7.0 <0.001*

Residual astigmatism (diopters) 0.5 0.8 (−2.25)–2.50 −1.5 1.7 (−5.35)–2.20 <0.001*

CH (mmHg) 10.1 1.9 5.7–15.2 8.2 1.4 6.3–10.5 0.007*

CRF (mmHg) 9.7 2.4 4.4–16.1 7.4 2.3 4.2–11.0 0.010*

*Significant at the 0.05 level

UVA0uncorrected visual acuity, BSCVA0best spectacle-corrected visual acuity, CCT0central corneal thickness; CRF0corneal resistance factor;CH0corneal hysteresis; SD0standard deviation

566 Graefes Arch Clin Exp Ophthalmol (2012) 250:565–573

the operator activated the device. An air puff was released by anon-contact probe, which scanned the central area of the eyeand sent a signal to the ORA. In brief, the air puff causes thecornea to move inward, past applanation, and into slight con-cavity. After milliseconds, the air pumps shut off, the pressuredecreases, and the cornea begins to return to its normal state.The system monitors the entire process and measures twopressure values, which are determined from the inward andoutward applanation processes. The aforementioned measur-ing procedure enables the determination of: a) CH, which isrelated to the viscoelastic structure of the corneal tissue, and is

calculated as the difference between the two pressure values atthe two applanation processes, and b) CRF, which is indicativeof the overall resistance of the cornea and is calculated as alinear function of the two pressures associated with the twoapplanation. In order to ensure accurate results, ORAwas donefour times for each eye. Signals that differed significantly inappearance from the other signals from the same eye weredeleted. All data using ORA were obtained by the sameoperator (M.G.) in a consistent way during office hours.

Surgical procedure

The same surgical procedure was applied to all keratoconuspatients, which included: instillation of proparacaine hydro-chloride 0.5% drops for topical anaesthesia, application of asponge saturated with 10% alcohol to the central cornea for30 seconds, and subsequent de-epithelialization by means ofa hockey knife. Following de-epithelialization, a mixture of0.1% riboflavin in 20% Dextran solution was instilled to thecornea for 30 minutes (two drops every 2 minutes) prior to theirradiation, until the stroma was completely penetrated andaqueous was stained yellow. The ultraviolet-A radiation sourcethat was used is UV-XTM (IROCAG, Zurich, Switzerland). Indetail, an 8.0 mm diameter of central cornea was irradiated for30 minutes by ultraviolet-A light with a wavelength of 370 nmand an irradiance of 3 mW/cm2. Instillation of riboflavin drops(one drop every 2 minutes) was continued during irradiation aswell, in order to sustain the necessary concentration of theriboflavin.Moreover, balanced salt solution (BSS) was appliedevery 6 minutes to moisten the cornea.

Postoperative management

After treatment all patients were prescribed topical ofloxacindrops qid, fluorometholone qid and diclofenac nitrate qid,

Table 2 Examination results of keratoconus group after CXL treatment

(Keratoconus eyes, n050) 3rd month postop 6th month postop 12th month postop

Mean SD P Mean SD P Mean SD P

UVA 0.3 0.3 0.007* 0.4 0.3 <0.001* 0.4 0.3 <0.001*

BCVA 0.7 0.2 0.018* 0.7 0.2 0.014* 0.7 0.2 0.010*

CCT (μm) 445.3 38.1 0.876 448.5 38.2 0.987 449.0 37.7 0.511

Average keratometry (diopters) 47.9 4.3 <0.001* 48.6 3.6 0.049* 48.7 3.7 0.037*

Corneal astigmatism (diopters) 3.1 1.6 0.041* 3.1 1.4 0.011* 3.2 1.5 0.047*

Residual astigmatism (diopters) −1.3 1.8 0.755 −1.4 1.4 0.810 −1.5 1.6 0.829

CH (mmHg) 8.8 1.6 0.345 8.7 1.1 0.518 8.7 1.4 0.077

CRF (mmHg) 7.9 2.0 0.765 7.6 1.6 0.479 7.9 1.9 0.517


UVA0uncorrected visual acuity, BSCVA0best spectacle-corrected visual acuity, CCT0central corneal thickness; CRF0corneal resistance factor; CH0corneal hysteresis; SD0standard deviation

Fig. 1 Box-and-whisker plots (median and interquartile range) of CHand CRF in normal and keratoconic eyes before treatment. (CH0cornealhysteresis, CRF0corneal resistance factor)

Graefes Arch Clin Exp Ophthalmol (2012) 250:565–573 567

accompanied by frequent instillation of artificial tears. Softtherapeutic lens was applied until complete re-epithelializationof the cornea was detected. Follow-up visits were performedon the 1st day, 7th day, 1st, 3rd, 6th and 12th month after theoperation.

Statistical analysis

The numerical data accrued from all procedures were statis-tically evaluated using SPSS (software version 17.0, Inc.). Ap value <0.05 was considered significant. Significant differ-ences between measurements were assessed with analysis ofvariance with post-hoc tests and Student’s t-tests. Variablecorrelations were evaluated with Pearson analysis (r), andmultiple regression models with CH and CRF as dependentvariables and gender, age, uncorrected visual acuity (UVA),best spectacle-corrected visual acuity (BSCVA), averagekeratometry (Km), corneal astigmatism (Astig.), residualastigmatism, and CCT as independent variables were per-formed for the CG, the KG preoperatively and the KGpostoperatively.

Results

Demographic data and preoperative data for the KG and theCG participants are presented in Table 1. Independent sam-ples t-tests were conducted in order to determine significantdifferences in all examined measures. Box-whisker plots ofthe CH and CRF for both groups are presented in Fig. 1.

Examination results for the keratoconus group after CXLtreatment are presented in Table 2. A one-way repeatedmeasures ANOVA and paired-samples Student’s t-test wereconducted to compare preoperative and postoperative meas-urements. Figure 2 shows mean CH and CRF of keratoconiceyes up to 12 months after treatment.

Correlations between measured parameters are presentedin Table 3. Preliminary analyses were performed to ensure

no violation of the assumptions of normality, linearity andhomoscedasticity. Figures 3, 4, 5, 6, 7, 8 show the relation-ships between corneal parameters (CCT, Km and Astig.) andORA parameters (CH and CRF) in normal and keratoconiceyes before treatment.

There was not any statistically significant relationshipbetween UVA and ORA parameters. However, there was asignificant positive relationship between BSCVA and ORAparameters. The slope of the regression equation betweenCH and BSCVA for normal eyes was 6.04, whereas that forkeratoconic eyes was 2.13. Similarly, the slope of the re-gression equation between CRF and BSCVA for normaleyes was 7.88, whereas that for keratoconic eyes was 2.91.Additionally, there was a significant positive relationshipbetween CCT and ORA parameters. The slope of the regres-sion equation between CH and CCT for normal eyes was0.027, whereas that for keratoconic eyes was 0.025. Simi-larly, the slope of the regression equation between CRF andCCT for normal eyes was 0.038, whereas that for kerato-conic eyes was 0.037. Furthermore, there was a significant

0

2

4

6

8

10

12

14

16

300 350 400 450 500 550 600 650CCT

CH

NormalKeratoconusNormalKeratoconus

Fig. 3 Scatterplot of relationship between CH (mmHg) and CCT (μm)in normal and keratoconic eyes before treatment. (CH0corneal hyster-esis, CCT0central corneal thickness)

Table 3 Pearson correlations (r)

CH CRF

R P R P

Gender 0.125 0.503 0.117 0.532

Age 0.038 0.773 0.002 0.987

UVA 0.028 0.836 0.008 0.953

BSCVA 0.469 <0.001* 0.463 <0.001*

CCT 0.590 <0.001* 0.620 <0.001*

Average keratometry −0.560 <0.001* −0.636 <0.001*

Corneal astigmatism −0.347 0.007* −0.346 0.007*

Residual astigmatism 0.387 0.005* 0.315 0.025*


UVA0uncorrected visual acuity, BSCVA0best spectacle-corrected vi-sual acuity, CCT0central corneal thickness; CH0corneal hysteresis;CRF0corneal resistance factor

0

1

2

3

4

5

6

7

8

9

10

Preoperatively 6 months post

Mea

n a

nd

SD

CH CRF

3 months post 1 month post

Fig. 2 Mean CH and CRF of keratoconic eyes after CXL. (CH0corneal hysteresis, CRF0corneal resistance factor, CXL0corneal col-lagen cross-linking)


negative relationship between Km and the ORA parameters.The slope of the regression equation between CH and Km fornormal eyes was −0.569, whereas that for keratoconic eyes was−0.252. Similarly, the slope of the regression equation betweenCRF and Km for normal eyes was −0.874, whereas that forkeratoconic eyes was −0.470. In addition, a significant negativerelationship was found between Astig. and ORA parameters.The slope of the regression equation between CH andAstig. fornormal eyes was −0.207, whereas that for keratoconic eyes was−0.346. Similarly, the slope of the regression equation betweenCRF and Astig. for normal eyes was −0.492, whereas that forkeratoconic eyes was −0.307. Also, a significant positive rela-tionship was found between residual astigmatism and ORAparameters. The slope of the regression equation betweenCH and residual astigmatism for normal eyes was 0.753,whereas that for keratoconic eyes was −0.059. Similarly, theslope of the regression equation between CRF and residualastigmatism for normal eyes was 0.807, whereas that forkeratoconic eyes was −0.520. The correlations between gen-der and ORA parameters and between age and ORA param-eters were not significant (Table 3).

According to multiple regression models for the threegroups (CG, KG preoperatively, and KG postoperatively),

with CH and CRF as dependent variables, there were sig-nificant correlations between variables used in the models,but all of them were <0.7, so we can accept that the possi-bility of multicollinearity or singularity in the models is low.Also, assumptions of normality, linearity, homoscedasticity,and independence of residuals were checked and not foundviolated. Beta values of all variables are presented inTables 4, 5, 6.

Discussion

Previously, various investigators attempted to measure theocular rigidity of the keratoconic eyes in order to reveal thepathologic processes that affect the corneal tissue [8,13–15]. A significant limitation of these studies was thatmost of them were performed in vitro. Edmund et al. [8]compared the viscoelasticity of keratoconic and normaleyes, and found that the distensibility of eyes was higherin normal eyes, so they concluded that distensibility may bean important factor in the pathogenesis of keratoconus. Thesame research group reported [14] that corneal rigidity tendsto be lower in keratoconus, and concluded that a decrease in

0

2

4

6

8

10

12

14

16

0,0 1,0 2,0 3,0 4,0 5,0 6,0 7,0 8,0Astig.

CH


Fig. 7 Scatterplot of relationship between CH (mmHg) and Astig. (D) innormal and keratoconic eyes before treatment. (CH0corneal hysteresis,Astig.0corneal astigmatism)

0

2

4

6

8

10

12

14

16

18

20

20 30 40 50 60 70Km

CR

F


Fig. 6 Scatterplot of relationship between CRF (mmHg) and Km (D)in normal and keratoconic eyes before treatment. (CRF0corneal resis-tance factor, Km0mean keratometry)

0

2

4

6

8

10

12

14

16

18

20 30 40 50 60 70Km

CH


Fig. 5 Scatterplot of relationship between CH (mmHg) and Km (D) innormal and keratoconic eyes before treatment. (CH0corneal hystere-sis, Km0mean keratometry)

0

2

4

6

8

10

12

14

16

18

300 350 400 450 500 550 600 650CCT

CR

F


Fig. 4 Scatterplot of relationship between CRF (mmHg) and CCT(μm) in normal and keratoconic eyes before treatment. (CRF0cornealresistance factor, CCT0central corneal thickness)


the corneal matrix and in the corneal tissue mass might be animportant pathogenic factor in the development of keratoco-nus. Andreassen et al. [16] also found that the corneal tissue inthe keratoconic eye was more elastic than that in normalsubjects. Brooks et al. [17] calculated the ocular rigiditycoefficient of keratoconic eyes from the combination of appla-nation tonometry and impression tonometry, using theFriedenwald [18] nomogram and the line of best fit. Theyfound that the ocular rigidity of the keratoconic eyes wassignificantly lower than the control when corneal thinning of40% or more was present. Foster and Yamamoto [13] alsocalculated the ocular rigidity coefficient in a similar way, butfailed to demonstrate any statistically significant difference incorneal rigidity unless corneal thinning was 60% or more.Therefore, they concluded that the Friedenwald method ofcalculating ocular rigidity was not accurate, and did not reflectthe true viscoelastic properties of keratoconic eyes. Hartsteinand Becker [15] also calculated rigidity by the Friedenwaldnomogram, and found lower corneal rigidity in keratoconiceyes. In general, all of these studies showed a reduced rigidityin keratoconic eyes, but all measurement methods suggestedare impractical in clinical settings.

With regard to the results of our study, there is evidencefor significant differences in all examined measures between

KG and CG. CH and CRF in normal eyes were higher thanthat in keratoconic eyes. Our results are in agreement withprevious studies performed with ORA [1, 6, 18–20].

With regard to possible relationships between visual acuityand ORA parameters, non-significant correlations of CH andCRF with UVAwere detected (Table 3). However, regressionanalysis revealed a strong positive relationship between ORAparameters and BSCVA (i.e., the higher CH and CRF, thehigher BSCVA and vice versa) (Table 3). This strong correla-tion suggests that changes in CH and CRF may be indicativeof changes in visual acuity after CXL.

An analysis of a possible relationship between the CCTand ORA parameters of normal and keratoconic eyes wasperformed (Figs. 3 and 4), and a significant correlation wasfound between CCT and CH and between CCT and CRF(Table 3). These correlations were similar for both groups.When a simple regression line was applied, a positive rela-tionship was revealed (i.e., the higher the CCT, the higherthe CH and CRF and vice versa). The results of severalstudies [18, 21] agree with our results, while other studiesdemonstrated that only CRF and not CH showed a strongcorrelation with CCT [22, 23]. However, in our study, thecorrelation coefficients (Table 3) imply that CH, CRF, andCCT are related, but are not measurements of the samebiomechanical parameter. In the absence of another reliablemeasure of viscoelasticity, it is difficult to assess to whatextent the CH and CRF values are thickness (CCT)-depen-dent rather than viscoelasticity-dependent.

Furthermore, possible relationships between Km andORA parameters were explored (Figs. 5 and 6, Table 3).Regression analysis revealed a negative relationship (i.e.,the higher the Km, the lower the CH and CRF and viceversa). A similar relationship was observed when we com-pared Astig. and ORA parameters (Figs. 7 and 8, Table 3).Also, a statistically significant positive correlation exists be-tween ORA parameters and residual astigmatism (Table 3).According to our knowledge, this is the first study that corre-lates ORA parameters with mean keratometry, corneal astig-matism, and residual astigmatism in normal and keratoconic

Table 4 Control group (CG)beta values of independentvariables included in themultiple regression modelswith CH and CRF as dependentvariables

CRF0corneal resistance factor;CH0corneal hysteresis

CH multiple regression model CRF multiple regression model

Beta P Beta P

Gender 0.071 0.827 0.137 0.643

Age 0.232 0.541 0.128 0.706

UVA 0.135 0.592 0.071 0.755

BSCVA 0.189 0.549 0.134 0.638

CCT 0.370 0.236 0.394 0.167

Average keratometry −0.350 0.196 −0.374 0.064

Corneal astigmatism −0.077 0.826 −0.077 0.808

Residual astigmatism 0.242 0.524 0.193 0.574

0

2

4

6

8

10

12

14

16

18

0,0 1,0 2,0 3,0 4,0 5,0 6,0 7,0 8,0Astig.

CR

F


Fig. 8 Scatterplot of relationship between CRF (mmHg) and Astig.(D) in normal and keratoconic eyes before treatment. (CRF0cornealresistance factor, Astig.0corneal astigmatism)


eyes. Since both keratometry and corneal astigmatism areindicative for keratoconus staging, this strong correlation sug-gests that CH and CRF may be useful both in the diagnosisand the classification of keratoconus. Also, correlation be-tween ORA parameters and residual astigmatism shows thatthe effect of CH and CRF on residual astigmatism and possi-ble changes in them could affect, for example, visual outcomeor patients’ ability to wear contact lenses.

Non-significant correlations between ORA parametersand age were detected (Table 3). Similar results were pre-sented by Shah et al. [18] In contrast, Ortiz et al. [1]demonstrated that CH was decreased in older eyes, and thedifference between the youngest age group (9 to 14 years)and oldest age group (60 to 80 years) was significant (p00.01, t-test), Kamiya et al. [24] in their study found a weak,but significant, negative correlation between age and CH(r0−0.17, p00.02) and CRF (r0−0.18 p00.01), while Kidaet al. [25] concluded that aging may lower CH and CRF.

Moreover, in accordance to former researchers no signif-icant correlations between ORA parameters and genderwere detected (Table 3) [9].

Overall R square of multiple regression model with CH asdependent variable was 0.439, which means that the modelexplains 43.9 per cent of the variance in CH (p<0.05) andoverall R square of multiple regression model with CRF as

dependent variable was 0.536, which means that the modelexplains 53.6 per cent of the variance in CRF (p<0.05). Rsquare of multiple regression model for CG with CH asdependent variable was 0.519, which means that the modelexplains 51.9 per cent of the variance in CH (p<0.05). Of allthe variables included, CCT makes the strongest unique con-tribution to the prediction of the dependent variable CH(beta00.370), but not at a statistical significant level (p>0.05). In fact, none of the variables makes a significant uniquecontribution to this prediction (p>0.05 for all independentvariables). Similarly, R square of multiple regression modelfor CG with CRF as dependent variable was 0.610, whichmeans that the model explains 61.0 per cent of the variance inCRF (p<0.05). In this model too, CCT makes the strongestunique contribution to explaining the dependent variable CRF(beta00.394). However, in this case too, none of the variablesmakes a significant unique contribution (p>0.05 for all inde-pendent variables). In KG preoperatively, R square of multipleregression model with CH as dependent variable was 0.258,while R square of multiple regression model with CRF asdependent variable was 0.260 (p<0.05). In accordance withresults of multiple regression analysis for CG, in these modelstoo, of all variables included, CCTmakes the strongest uniquecontribution to explaining the dependent variables CH andCRF with beta values 0.265 and 0.294 respectively, but not at

Table 6 Keratoconus group(KG) postoperatively—betavalues of independent variablesincluded in the multiple regres-sion models with CH and CRFas dependent variables




Beta P Beta P

Gender 0.271 0.314 0.091 0.515

Age −0.112 0.635 −0.084 0.567

UVA 0.176 0.681 0.517 0.956

BSCVA 0.527 0.348 0.118 0.343

CCT −0.696 0.686 −0.655 0.409


Corneal astigmatism 0.712 0.046* 0.757 0.036*

Residual astigmatism −0.568 0.312 −0.377 0.096

Table 5 Keratoconus group(KG) preoperatively—betavalues of independent variablesincluded in the multipleregression models with CH andCRF as dependent variables



Beta P Beta P

Gender 0.034 0.818 0.061 0.674

Age −0.189 0.194 −0.167 0.248

UVA −0.047 0.795 −0.026 0.887

BSCVA 0.124 0.475 0.052 0.762

CCT 0.265 0.111 0.294 0.077


Corneal astigmatism −0.145 0.353 −0.016 0.918

Residual astigmatism −0.094 0.510 −0.135 0.344


a statistically significant level, since p values for CCT are>0.05 in both models. In these cases too, none of the variablesis making significant unique contribution to the prediction ofCH and CRF (p>0.05). After CXL, R square of multipleregression model for KG with CH as dependent variable was0.284, while R square of multiple regression model with CRFas dependent variable was 0.241 (p<0.05). In these models, ofall variables included, corneal astigmatism makes the stron-gest unique contribution to explaining the dependent variablesCH and CRFwith beta values 0.712 and 0.757 respectively. Inthese cases, corneal astigmatism is making a statistically sig-nificant unique contribution to the prediction of the dependentvariables CH and CRF, since p00.046 and p00.036 respec-tively (<0.05 in both cases). All of the other variables includedare not making significant unique contributions (p>0.05).

With regard to the impact of CXL on corneal parameters,significant differences were encountered for UVA, BSCVA,Km and Astig., but not for CCT and residual astigmatism.

On the other hand, post-CXL values of CH and CRFwere non-significantly higher, in all postoperative visits.According to international bibliography, our study is amongthe few ones that attempt to investigate CH and CRF trendsafter CXL. Similar results were presented by Goldich et al.[26] and by Sedaghat et al. [27] in a series of ten and 56 eyesrespectively, with a follow-up of 6 months. Vinciguerra etal. [28] extended the post-operative period to 12 months;however, they recruited only 24 eyes in their report. Inaccordance to our results, they also demonstrated non-significant post-CXL changes both for CH and CRF. In fact,the outcome of our study seems to be in contrast to theresults of previous in vitro studies, which demonstrated thatCXL actually increases the biomechanical properties of thecornea. Confocal [29], histologic [30] and thermomechan-ical [31] studies of corneas after CXL confirmed the earlierresults. Prior to the interpretation of these potentially contra-dicting results, it should be taken into account that Glass etal. [32] reported a correlation between the ORA measure-ments and the deformation area from the air pulse. Theydemonstrated that the deformation response differs betweennormal corneas and keratoconic corneas. In fact, keratoconiccorneas are not homogeneous, which results in asymmetricor uneven deformation and thus variable readings. More-over, they found that CH is a complex indicator of theviscous and elastic properties, and that the interaction ofthese factors influences CH. These findings suggest thateven though changes in elastic and viscous behaviour mayoccur, they might not necessarily produce a change in CH.Moreover, the results of our study indicate a significantcorrelation between mean keratometry and CH and CRF.

Nevertheless, further studies are needed to assess whetherthis increase of CH and CRFmeasurements after CXL at a non-significant level is because CH and CRF are not significantlyaffected after CXL, or whether the CXL technique used could

be improved. Comparison between CXL-treated keratoconuseyes with non-treated keratoconus or alternatively-treated ker-atoconus eyes would also have been of interest.

Summarizing, this study attempts to contribute to thebody of knowledge regarding the importance of CH andCRF in keratoconus, especially after CXL treatment. Inaccordance with former surveys, we confirmed that signifi-cant differences do exist between normal and keratoconiceyes for both parameters. Further analysis of our resultsrevealed significant correlations between ORA parametersand BSCVA, CCT, Astig., residual astigmatism, and Km. Toour knowledge, this is the first study that demonstrates thesecorrelations. Multiple regression analysis models revealedthat for CG and KG preoperatively CCT makes the strongestunique contribution to the prediction of CH andCRF, but not ata statistical significant level. So, we can conclude that none ofthe independent variables can predict the dependent variablesCH and CRF. For KG after CXL, CCT seems to play asecondary role, and corneal astigmatism makes the strongestunique contribution at a statistically significant level. However,the R square values of bothmultiple regression analysis modelsfor KG after CXL are rather law. On the other hand, weconfirmed the non-significant impact of CXL on ORA meas-urements for a period of 12 months. Since earlier reports reliedon a limited number of recruited eyes, our results provide arobust indication of non-significant increase for both CH andCRF after CXL. Additional studies will explore the full poten-tial of these factors in keratoconus and CXL.

References

1. Ortiz D, Pinero D, Shabayek MH, Arnalich-Montiel F, Alio JL(2007) Corneal biomechanical properties in normal, post-laser insitu keratomileusis, and keratoconic eyes. J Cataract Refract Surg33:1371–1375

2. Chihara E, Takahashi H, Okazaki K, Park M, Tanito M (2005) Thepreoperative intraocular pressure level predicts the amount ofunderestimated intraocular pressure after LASIK for myopia. Br JOphthalmol 89:160–164

3. Davison PF, Glabavy EJ (1986) Connective tissue remodelling incorneal and scleral wounds. Invest Ophthalmol Vis Sci 27:1478–1484

4. Kaufmann C, Bachmann LM, Thiel MA (2003) Intraocular pres-sure measurement using dynamic contour tonometry after laser insitu keratomileusis. Invest Ophthalmol Vis Sci 44:3790–3794

5. Munger R, Dohadwala AA, Hodge WG, Jackson WB, MintsioulisG, Damji KF (2001) Changes in measured intraocular pressureafter hyperopic photorefractive keratectomy. J Cataract RefractSurg 27:1254–1262

6. Luce DA (2005) Determining in-vivo biomechanical properties ofthe cornea with an ocular response analyzer. J Cataract RefractSurg 31:156–162

7. Rabinowitz YS (1998) Keratoconus (major review). Surv Ophthalmol42:297–319

8. Edmund C (1987) Assessment of an elastic model in the patho-genesis of keratoconus. Acta Ophthalmol 65:545–550

9. Spoerl E, Huhle M, Seiler T (1998) Induction of cross-links incorneal tissue. Exp Eye Res 66:97–103


10. Wollensak G, Spoerl E, Seiler T (2003) Riboflavin/ultraviolet-A-induced crosslinking for the treatment of keratoconus. Am J Oph-thalmol 135:620–627

11. Wollensak G, Spoerl E, Seiler T (2003) Stress–strain measure-ments of human and porcine corneas after riboflavin-ultraviolet-A-induced cross-linking. J Cataract Refract Surg 29:1780–1785

12. Caporossi A, Baiocchi S, Mazzotta C, Traversi C, Caporossi T (2006)Parasurgical therapy for keratoconus by riboflavin-ultraviolet type Arays induced cross-linking of corneal collagen: preliminary refractiveresults in an Italian study. J Cataract Refract Surg 32:837–845

13. Foster CS, Yamamoto GK (1978) Ocular rigidity in keratoconus.Am J Ophthalmol 86:802–806

14. Edmund C (1988) Corneal rigidity and ocular rigidity in normaland keratoconic eyes. Acta Ophthalmol 66:134–140

15. Hartstein J, Becker B (1970) Research into the pathogenesis ofkeratoconus: a new syndrome: low ocular rigidity, contact lensesand keratoconus. Arch Ophthalmol 84:728–729

16. Andreassen TT, Simonsen AH, Oxlund H (1980) Biomechanical prop-erties of keratoconus and normal corneas. Exp Eye Res 31:435–441

17. Brooks AMV, Robertosn IF, Mahoney AM (1984) Ocular rigidity andintraocular pressure in keratoconus. Aust J Ophthalmol 12:317–324

18. Shah S, Laiquzzaman M, Bhojwani R, Mantry S, Cunliffe I (2007)Assessment of the biomechanical properties of the cornea with theocular response analyzer in normal and keratoconic eyes. InvestOphthalmol Vis Sci 48(7):3026–3031

19. Shah S, LaiquzzamanM (2009) Comparison of corneal biomechanicsin pre and post-refractive surgery and keratoconic eyes by OcularResponse Analyser. Cont Lens Anterior Eye 32(3):129–132

20. Mangouritsas G, Morphis G, Mourtzoukos S, Feretis E (2009)Association between corneal hysteresis and central corneal thick-ness in glaucomatous and non-glaucomatous eyes. Acta Ophthal-mol 87(8):901–905

21. Johnson RD, Nguyen MT, Lee N, Hamilton DR (2011) Cornealbiomechanical properties in normal, forme fruste keratoconus, andmanifest keratoconus after statistical correction for potentiallyconfounding factors. Cornea 30(5):516–523

22. Touboul D, Roberts C, Kérautret J, Garra C, Maurice-Tison S,Saubusse E, Colin J (2008) Correlations between corneal hysteresis,

intraocular pressure, and corneal central pachymetry. J CataractRefract Surg 34(4):616–622

23. Wasielica-Poslednik J, Berisha F, Aliyeva S, Pfeiffer N, HoffmannEM (2010) Reproducibility of ocular response analyzer measure-ments and their correlation with central corneal thickness. GraefesArch Clin Exp Ophthalmol 248(11):1617–1622

24. Kamiya K, Shimizu K, Ohmoto F (2009) Effect of aging oncorneal biomechanical parameters using the ocular response ana-lyzer. J Refract Surg 25(10):888–893

25. Kida T, Liu JH, Weinreb RN (2008) Effects of aging on cornealbiomechanical properties and their impact on 24-hour measure-ment of intraocular pressure. Am J Ophthalmol 146(4):567–572

26. Goldich Y, Barkana Y, Morad Y, Hartstein M, Avni I, Zadok D(2009) Can we measure corneal biomechanical changes after colla-gen cross-linking in eyes with keratoconus?—a pilot study. Cornea28(5):498–502

27. Sedaghat M, Naderi M, Zarei-Ghanavati M (2010) Biomechanicalparameters of the cornea after collagen crosslinking measured bywaveform analysis. J Cataract Refract Surg 36(10):1728–1731

28. Vinciguerra P, Albè E, Mahmoud AM, Trazza S, Hafezi F, RobertsCJ (2010) Intra- and postoperative variation in ocular response ana-lyzer parameters in keratoconic eyes after corneal cross-linking. JRefract Surg 26(9):669–676

29. Mazzotta C, Balestrazzi A, Traversi C, Baiocchi S, Caporossi T,Tommasi C, Caporossi A (2007) Treatment of progressive keratoco-nus by riboflavin-UVA-induced cross-linking of corneal collagen;ultrastructural analysis by Heidelberg Retinal Tomograph II in vivoconfocal microscopy in humans. Cornea 26:390–397

30. Wollensak G, Wilsch M, Spoerl E, Seiler T (2004) Collagen fiberdiameter in the rabbit cornea after collagen crosslinking by ribo-flavin/UVA. Cornea 23:503–507

31. Spoerl E,Wollensak G, Dittert D-D, Seiler T (2004) Thermomechan-ical behaviour of collagen-cross-linked porcine cornea. Ophthalmo-logica 218:136–140

32. Glass DH, Roberts CJ, Litsky AS, Weber PA (2008) A viscoelasticbiomechanical model of the cornea describing the effect of viscos-ity and elasticity on hysteresis. Invest Ophthalmol Vis Sci 49(9):3919–3926


evaluation of corneal hysteresis and corneal resistance factor after corneal cross-linking for...

Documents