O

Ct

FJJ

HM

a

A

R

A

A

K

G

D

G

c

2

a r c h s o c e s p o f t a l m o l . 2 0 1 1;8 6(9):287–291

ARCHIVOS DE LA SOCIEDADESPAÑOLA DE OFTALMOLOGÍA

www .e lsev ier .es /o f ta lmologia

riginal article

omparison of Goldmann applanation and dynamic contouronometry measurements: Effects of corneal morphometry�

. Sáenz-Francés ∗, R. García-Catalán, M. Jerez-Fidalgo, A. Fernández-Vidal,.M. Martínez-de-la-Casa, C. Méndez-Hernández, E. Santos-Bueso, J. Reche-Frutos,. García-Sánchez, J. García-Feijoo

ospital Clínico Universitario San Carlos, Instituto de Investigaciones Oftalmológicas Ramón Castroviejo, Universidad Complutense,adrid, Spain

r t i c l e i n f o

rticle history:

eceived 3 August 2010

ccepted 15 April 2011

vailable online 15 February 2012

eywords:

laucoma

ynamic contour tonometry

oldmann applanation tonometry

a b s t r a c t

Objectives: To compare intraocular pressure (IOP) measurements made by Goldmann appla-

nation tonometry (GAT) and dynamic contour tonometry (DCT).

Methods: IOPs were measured by GAT and DCT in 63 eyes of 63 healthy subjects. A compar-

ison was made by intraclass correlation coefficient. Passing–Bablok plot was constructed

to establish the existence of systematic and/or proportional biases. Multivariate regression

analysis was used to examine whether the measurements of both instruments were affected

by the power of the steepest and flattest corneal axes, their orientation, age or central corneal

thickness (CCT).

Results: The intra-class correlations (ICCs) were 0.57 (95% confidence interval (95% CI):

0.29–0.74). Mean differences were 1.68 (DCT minus GAT) (95% CI: 0.92–2.44). Passing–Bablok

analysis (X = DCT, Y = GAT) revealed a systematic bias (A = −14.35, 95% CI: −24.51 to [−9.14])

and a proportional bias (B = 1.74, 95% CI: 1.43–2.26). Multivariate regression analysis revealed

that the DCT was independent of the corneal characteristics analyzed while GAT was biased

by CCT (B = 0.042, 95% CI: 0.002–0.082).

Conclusions: While GAT was biased by corneal CCT; DCT readings were independent of

corneal morphometry.

© 2010 Sociedad Española de Oftalmología. Published by Elsevier España, S.L. All rights

reserved.

� Please cite this article as: Sáenz-Francés F, et al. Concordancia entre la tonometría de aplanación de Goldmann y la tonometría deontorno dinámico: efectos de la morfometría corneal. Arch Soc Esp Oftalmol. 2011; 86:287–91∗ Corresponding author.

E-mail address: federicosaenzfrancessb@gmail.com (F. Sáenz-Francés).

173-5794/$ – see front matter © 2010 Sociedad Española de Oftalmología. Published by Elsevier España, S.L. All rights reserved.

288 a r c h s o c e s p o f t a l m o l . 2 0 1 1;8 6(9):287–291

Concordancia entre la tonometría de aplanación de Goldmann y latonometría de contorno dinámico: efectos de la morfometría corneal

Palabras clave:

Glaucoma

Tonometría de contorno

dinámico

Tonometría de aplanación

de Goldmann

r e s u m e n

Objetivos: Determinar la concordancia entre la presión intraocular (PIO) medida mediante

tonometría de aplanación de Goldmann (TAG) y tonometría de contorno dinámico (TCD).

Métodos: La PIO se midió mediante TAG y TCD en 63 ojos de 63 voluntarios sanos. La con-

cordancia se determinó mediante coeficiente de correlación intraclase (CCI). Se empleó el

método de Passing-Bablok para establecer la presencia de sesgos proporcionales y/o sis-

temáticos. El análisis de regresión multvariable se empleó para determinar si las diferencias

estaban condicionadas por la potencia de los ejes mayor y menor de la córnea, por su

orientación, por el grosor corneal central (GCC) y por la edad.

Resultados: El CCI fue 0,57 (intervalo de confianza [IC] al 95%: 0,29-0,74). La diferencia media

entre los dos instrumentos fue 1,68 mmHg (TCD menos TAG) (IC 95%: 0,92-2,44). El análisis

de Passing-Bablok (X = DCT, Y = GAT) reveló la presencia de un sesgo sistemático (A = −14,35,

IC 95%: −24,51-[−9,14]) y otro proporcional (B = 1,74, IC 95%: 1,43-2,26). El análisis de regre-

sión multivariable mostró una total independencia del TCD de las características corneales

mientras que la TAG estaba afectada por el GCC (B = 0,042 CI 95%: 0,002-0,082).

Conclusiones: La TCD se mostró independiente de las características morfométricas de la

córnea mientras que la TAG se afectó por el GCC.

© 2010 Sociedad Española de Oftalmología. Publicado por Elsevier España, S.L. Todos los

Introduction

Intraocular pressure (IOP) is the main risk factor for the devel-opment and progression of glaucoma. In addition, it is theonly factor in which therapeutic intervention has demon-strated to improve the control of the disease.1 Nowadays,Goldmann applanation tonometry (GAT) constitutes the goldstandard for measuring IOP, although subject to various mor-phometric characteristics of the cornea such as corneal centralthickness (CCT), the power of the major and minor axes andthe orientation thereof.2,3

In recent years IOP measuring instruments have beendeveloped with the aim of overcoming the shortcomings of theGAT results. These include the bounce tonometer, the ocularresponse analyzer and the dynamic contour tonometer (DCT)or Pascal tonometer,4 which is the object of this paper.

The DCT has a tip similar to the GAT cone, equipped witha concave surface designed to adapt to the corneal surface,maintain its shape and curvature in order to cause minimumcorneal distortion in contrast with the distortion caused byGAT applanation. A digital sensor inserted in the contact sur-face carries out transcorneal IOP measurements. In this way,theoretically the DCT measurements are independent of thecorneal characteristics which affect the GAT.4

The objective of this study is to assess the match betweenGAT and DCT in normal subjects as well as to determine theinfluence of CCT, corneal curvature radius and age on the mea-surement of both tonometers.

Subjects, material and methods

Volunteers were recruited between workers and rela-tives of patients of the San Carlos Clinical University

derechos reservados.

Hospital of Madrid, Spain. The inclusion criteria comprisedthe ability to sign an informed consent, normal ophthal-mological exploration results including IOP below 22 mmHgand the absence of systemic diseases. In all patients, theeye to be analyzed was decided by automatic randomization(www.randomization.com).

The study excluded patients with a history of ocularsurgery or trauma or with any corneal alteration which couldhinder adequate applanation tonometry.

The study protocol was approved by the Ethical Committeeof the San Carlos University Clinical Hospital of Madrid. All thepatients signed an informed consent included in said protocol.

All the patients submitted to a full ophthalmologicalassessment, comprising best corrected visual acuity (Marte-sian logarithmic scale), anterior segment biomicroscopy,gonioscopy, IOP measurements with Goldmann applana-tion tonometer and dynamic contour tonometer, funduscopy,Octopus TOP-G1X automated perimetry.

The match was determined with the intraclass correlationcoefficient (ICC). The Passing–Bablok method was utilized toestablish the presence of proportional and/or systematic bias.Two multivariate regression analyses were utilized, one foreach tonometer, to determine whether the measurements ofboth were influenced by the power of the major and minoraxes of the cornea, by the orientation, the central cornealthickness (CCT) and by age. In both models, the predictive val-ues were the power of the major cornea axis, the minor corneaaxis, the orientation of the minor axis rated at 30◦ intervals,age and central corneal thickness. In order to choose the bestregression equation, adjusted R2 was utilized as criterion.

All the statistical analyses were made utilizing the SPSS

software version 15.0 for Windows (SPSS Inc., Chicago, IL,USA). The Kolmogorov–Smirnov test was utilized to certifythe normality of quantitative data distribution. The requiredstatistical significance level was set at p < 0.05.

a r c h s o c e s p o f t a l m o l . 2 0 1 1;8 6(9):287–291 289

Table 1 – Statistics describing the analyzed variables.

Variable description

Mean Average Std. dev. Percentiles

25 50 75

Age 63.52 68.00 16.93 52.00 68.00 78.00Major axis potential 44.65 44.60 1.73 43.70 44.60 45.70Minor axis potential 43.17 43.70 3.80 42.70 43.70 44.70DCT IOP 17.90 18.00 2.70 15.80 18.00 19.70GAT IOP 16.22 16.00 4.14 14.00 16.00 18.00CCT 555.08 560.00 31.09 531.00 560.00 575.00

ann applanation tonometry; DCT: dynamic contour tonometry.

R

SIlwoTTcF

vmw(((r91G

atamdct0

9

12

15m

mH

g

18

21

DCT GAT

Fig. 1 – Box diagrams showing the distribution ofintraocular pressure measured with dynamic contour

Discussion

Even though at the present time GAT is the gold standardfor assessing IOP, it is known that certain anatomic or

620

CCT: central corneal thickness; IOP: intraocular pressure; GAT: Goldm

esults

ixty-three healthy volunteers were included in this study.t was randomly decided (www.randomization.com) to ana-yze the left eye of each patient. Of the 63 volunteers, 31ere males and 32 females, all Caucasian. The characteristicsf the analyzed quantitative variables are summarized inable 1; the orientation of the minor cornea axis is shown inable 2; the IOP distribution measured with both tonometersan be seen in Fig. 1 and the CCT distribution is shown inig. 2.

The ICC (absolute match) was of 0.57 (confidence inter-al [CI] at 95%: 0.29–0.74). The difference between the meaneasurements of both instruments (t-test for paired data)as of 1.68 mmHg (DCT minus GAT) (CI 95%: 0.92–2.44)

Fig. 1). The Passing–Bablok regression straight line analysisFig. 3) was the result of comparing the DCT measurementsabscisa axis) with the GAT measurements (coordinates axis)evealed the presence of a systematic bias (A = −14.35. CI5%: −24.51 to [−9.14]) and a proportional (B = 1.74. CI 95%:.43–2.26), therefore being the lesion of this straight line:AT = −14.35 + 1.74 × DCT.

Multivariate regression analysis allowed to determine thebsence of interaction and of confusion phenomena betweenhe predictive variables. In addition, on the basis of thedjusted R2 determination coefficient, it allowed us to deter-ine that the best regression equation to predict the IOP

id not include any predictive variable (independent of the

orneal characteristics) for DCT and only included CCT inhe case of GAT (adjusted R2 = 0.11; p = 0.04; B = 0.042 CI 95%:.002–0.082).

Table 2 – Orientation of the minor corneal axis.

Orientation of the minor corneal axis

Frequency Percentage

Arc sector 0–30◦ 14 22.2Arc sector 30–60◦ 6 9.5Arc sector 60–90◦ 11 17.5Arc sector 90–120◦ 8 12.7Arc sector 120–150◦ 7 11.1Arc sector 150–180◦ 17 27.0

Total 63 100.0

tonometry and Goldmann applanation tonometry.

500

CCT

520

540

mcm 560

580

600

Fig. 2 – Box diagrams showing central corneal thicknessdistribution.

290 a r c h s o c e s p o f t a l m o l

Regression of Passing-Bablok

GA

T

30

25

20

15

10

5

12.5 15.0 20.017.5DCT

25.022.5

Fig. 3 – Passing–Bablok regression between dynamic

r

contour tonometry and Goldmann applanation tonometry.

morphometric characteristics such as CCT or corneal curva-ture can influence its measurements. A number of studieshave evidenced the limitations of GAT, with CCT being themain factor influencing its measurements.5–7 The importanceof pachymetry in glaucoma was emphasized in the OcularHipertension Glaucoma Study (OHTS), which demonstrated thatCCT was a predictive factor for the conversion of ocular hyper-tension (OHT) into glaucoma.8,9 This fact, together with theCCT differences found between ocular hypertensive corneaeand those of patients diagnosed with normotensive glaucoma,as well as the inefficiency of nomograms developed for cor-recting the effect of pachymetry in TAG measurements10–12

have determined the need of seeking new tonometry systemsmore independent of CCT and other corneal morphometricparameters.

DCT has been broadly analyzed in various studies, whichdemonstrated DCT is a highly reproducible tonometer.4,13,14

The relationship of DCT with GAT has also been exten-sively studied. Accordingly, in contrast with our results, Realinet al.15 compared GAT and DCT measurements in normal andglaucomatous populations, finding in the former a weak cor-relation between GAT and CCT but independently of DCT.In glaucomatous patients, these authors did not find a rela-tionship between the GAT measurements and CCT, althoughthey did find a weak reverse association between DCT andpachymetry. Lanza et al.16 found that the differences betweenthe measurements of both instruments were affected by CCTbut not by the corneal radii. In a sample of normal subjects,Jordao et al.17 found a strong correlation between CCT andGAT and, in contrast with our findings, a weak correlationbetween DCT and CCT. Surprisingly, Halkiadakis et al.18 didnot found any correlation between CCT with GAT and DCT in apopulation of glaucomatous and ocular hypertensive patients.Hamilton et al.19 found a weak correlation between the differ-ences between GAT and DCT with corneal thickness changes

derived from secondary edema due to the use of contactlenses (the CCT variations in relation to the use of contactlenses exhibited a mean value of 48.3 �m SD 14.4 �m). Millaet al.20 found that the optimum match (ICC of 0.54) between

. 2 0 1 1;8 6(9):287–291

both instruments occurred with CCT between 540 and 545 �mbut diminished drastically beyond said range of values. Eventhough Cerruti et al.21 matched our results, they did not find acorrelation between DCT and CCT; however, they did observethat the measurements of this instruments were influencedby corner curvature. In contradiction with these findings aswell as with our observations, Grieshaber et al.22 observed asimilar influence of CCT on GAT and DCT.

From the above it can be deduced that there is no consensusin scientific literature about the dependency of DCT vis-à-visCCT and keratometry; moreover, there is no absolute agree-ment about the relationship of GAT vis-à-vis said parameters.However, there is a general consensus on the fact that DCTtends to overestimate IOP vis-à-vis TAG,4,13–22 of fact verifiedin this study.

Our results in relation to the degree of agreement betweenboth instruments match those found in the literature. In ourstudy, it was considered of interest to construct a multi-variate regression model which includes, in addition to ageand central corneal thickness, keratometry-related variables.With this model it was possible to establish the absence ofinteraction and confusion effect between the analyzed predic-tive variables. In addition, we found complete independenceof DCT vis-à-vis CCT, keratometry powers, the orientationthereof and age while we observed dependency of the GATmeasurements on the CCT values. This suggests that DCTperforms more independently of the corneal characteristicsthan GAT and accordingly the former may be adequate as anew gold standard. If this is assumed, it would be necessaryto change the IOP paradigm and to redefine the thresholdbetween normality and OHT because the overestimation of thenew tonometry system vis-à-vis the current reference methodwould become obvious.

Funding

Carlos III Health Institute, “Cooperative Research ThematicNetwork. Project RD07/0062: Ocular pathology of aging, visualquality and quality of life”.

Conflict of interest

The authors have no conflict of interests to declare.

e f e r e n c e s

1. Leske MC. The epidemiology of open-angle glaucoma:a review. Am J Epidemiol. 1983;118:166–91.

2. Whitacre MM, Stein R. Sources of error with the use ofGoldmann-type tonometers. Surv Ophthalmol. 1993;38:1–30.

3. Whitacre MM, Stein RA, Hassanein K. The effect of cornealthickness on applanation tonometry. Am J Ophthalmol.1993;115:592–6.

4. Martínez-de-la-Casa JM, Garcia-Feijoo J, Vico E,Fernandez-Vidal A, Benitez del Castillo JM, Wasfi M, et al.

Effect of corneal thickness on dynamic contour, rebound,and Goldmann tonometry. Ophthalmology. 2006;113:2156–62.

5. Argus WA. Ocular hypertension and central cornealthickness. Ophthalmology. 1995;102:1810–2.

o l . 2

1

1

1

1

1

1

1

1

1

1

2

2

2

a r c h s o c e s p o f t a l m

6. Chatterjee A, Shah S, Bessant DA, Naroo SA, Doyle SJ.Reduction in intraocular pressure after excimer laserphotorefractive keratectomy: correlation with pre-treatmentmyopioa. Ophthalmology. 1997;104:355–9.

7. Herndon LW, Choudhri SA, Cox T, Damji KF, Shields MB,Allingham RR. Central corneal thickness in normal,glaucomatous, and ocular hypertensive eyes. ArchOphthalmol. 1997;115:1137–41.

8. Kass MA, Heuer DK, Higginbotham EJ, Johnson CA, Keltner JL,Miller JP, et al. The Ocular Hypertension Treatment Study: arandomized trialdetermines that topical ocularhypotensivemedicationdelays or prevents the onset ofprimary open-angle glaucoma. Arch Ophthalmol.2002;120:701–13.

9. Gordon MO, Beiser JA, Brandt JD, Heuer DK, Higginbotham EJ,Johnson CA, et al. The Ocular Hypertension Treatment Study:baseline factors that predict the onset of primary open-angleglaucoma. Arch Ophthalmol. 2002;120:714–20.

0. Doughty MJ, Zaman ML. Human corneal thickness and itsimpact on intraocular pressure measures: a review andmetaanalysis approach. Surv Ophthalmol. 2000;44:367–408.

1. Orssengo GJ, Pye DC. Determination of the true intraocularpressure and modulus of elasticity of the human cornea invivo. Bull Math Biol. 1999;61:551–72.

2. Gunvant P, O’Leary DJ, Baskaran M, Broadway DC, Watkins RJ,Vijaya L. Evaluation of tonometric correction factors. JGlaucoma. 2005;14:337–43.

3. Herdener S, Hafizovic D, Pache M, Lautebach S, Funk J. Is thePASCAL-Tonometer suitable for measuring intraocularpressure in clinical routine? Long- and short-term

reproducibility of dynamic contour tonometry. Eur JOphthalmol. 2008;18:39–43.

4. Herdener S, Pache M, Lautebach S, Funk J. Dynamic contourtonometry (DCT) versus Goldmann applanation tonometry

0 1 1;8 6(9):287–291 291

(GAT) – a comparison of agreement and reproducibility.Graefes Arch Clin Exp Ophthalmol. 2007;245:1027–30.

5. Realini T, Weinreb RN, Hobbs G. Correlation of intraocularpressure measured with Goldmann and dynamic contourtonometry in normal and glaucomatous eyes. J Glaucoma.2009;18:119–23.

6. Lanza M, Borrelli M, De Bernardo M, Filosa ML, Rosa N.Corneal parameters and difference between Goldmannapplanation tonometry and dynamic contour tonometryin normal eyes. J Glaucoma. 2008;17:460–4.

7. Jordão ML, Pedroso de Carvalho Lupinacci A, Ferreira EL,Enomoto IJ, Costa VP. Influence of age, central cornealthickness, and quality score on dynamic contour tonometry.Eye. 2008:12 [epub ahead of print].

8. Halkiadakis I, Patsea E, Chatzimichali K, Skouriotis S,Chalkidou S, Amariotakis G, et al. Comparison of dynamiccontour tonometry with Goldmann applanationtonometry in glaucoma practice. Acta Ophthalmol. 2009;87:323–8.

9. Hamilton KE, Pye DC, Kao L, Pham N, Tran AQ. The effectof corneal edema on dynamic contour and Goldmanntonometry. Optom Vis Sci. 2008;85:451–6.

0. Milla E, Duch S, Buchacra O, Masuet C. Poor agreementbetween Goldmann and Pascal tonometry in eyes withextreme pachymetry. Eye. 2009;23:536–42.

1. Ceruti P, Morbio R, Marraffa M, Marchini G. Comparison ofGoldmann applanation tonometry and dynamic contourtonometry in healthy and glaucomatous eyes. Eye.2009;23:262–9.

2. Grieshaber MC, Schoetzau A, Zawinka C, Flammer J, Orgul S.

Effect of central corneal thickness on dynamic contourtonometry and Goldmann applanation tonometryin primary open-angle glaucoma. Arch Ophthalmol. 2007;125:740–4.