Racial Differences in Optic Disc Topography: Title and subTitle BreakBaseline Results From the Confocal Scanning Laser OphthalmoscopyAncillary Study to the Ocular Hypertension Treatment Study FREE

The Ocular Hypertension Treatment Study (OHTS) is a multicenter, randomizedclinical trial designed to evaluate the safety and efficacy of topical ocularhypotensive medication in delaying or preventing the onset of glaucomatousvisual field loss and/or optic nerve deterioration in participants with ocularhypertension at moderate risk for developing primary open-angle glaucoma.¹ A secondary aim of the OHTS is to identify risk factors,including age, race, and cup-disc ratio, that predict which participants withocular hypertension are most likely to develop primary open-angle glaucoma.¹ Given that glaucoma is the leading cause of blindnessin African Americans and the prevalence of glaucoma has been found to be 4to 5 times higher in African Americans than in white subjects,^{2 - 7} theOHTS was designed so that 25% of the participants would be African American.The OHTS provides a unique opportunity to answer important questions aboutthe development of glaucoma in African Americans with ocular hypertension.

In recent years, several new instruments have been developed that showpromise for improving our ability to detect early glaucomatous damage andto follow progression of the disease. One of these instruments, the confocalscanning laser ophthalmoscope (CSLO), has been shown to provide objective,reproducible measurements of optic disc topography.^{8 - 13} TheOHTS CSLO Ancillary Study was designed to evaluate the effectiveness of theCSLO in detecting the presence and progression of glaucomatous optic discchanges and to determine whether optic disc topographic measurements are anaccurate predictor of visual field loss.¹ Oneof the specific aims of the OHTS CSLO Ancillary Study is to compare the racialdifferences in optic nerve topography among patients with ocular hypertension.This article describes the design of the OHTS CSLO Ancillary Study and comparesthe baseline optic disc topography of African American participants with otherOHTS CSLO Ancillary Study participants.

The OHTS design, eligibility criteria, and participant characteristicshave been described in detail elsewhere.¹ Inbrief, to be eligible for participation in the study participants were requiredto have 2 normal and reliable visual field tests (Humphrey 30-2; Carl ZeissMeditec, Dublin, Calif), qualifying intraocular pressure (IOP) readings between24 and 32 mm Hg in one eye and between 21 and 32 mm Hg in the fellow eye,and normal color stereoscopic optic disc photographs for both eyes. Intraocularpressure was measured at least twice according to a standard protocol. Reliablevisual fields were defined as less than 33% false-positive results, less than33% false-negative results, and less than 33% fixation losses. Normal visualfields were classified based on clinical review at the Visual Field ReadingCenter and by STATPAC II (Carl Zeiss Meditec) criteria for global indicesfor corrected pattern standard deviation (CPSD) within the 95% age-specificpopulation norm and a glaucoma hemifield test result within the 97% age-specificpopulation norm.¹⁴ Prior to randomization,2 independent, masked, certified readers at the OHTS Optic Disc Reading Centerjudged the color stereoscopic optic disc photographs as normal. Individualswere excluded from the study if the photographs documented a localized notchor thinning of the neuroretinal rim, a diffuse or localized area of pallor,an optic disc hemorrhage, or an asymmetry in the cup-disc ratios greater than0.2. between the 2 eyes.¹

Twenty-two study centers participated in the OHTS. Participants wererandomized to either the close observation or the medication groups. Randomizationwas stratified by clinic and race.

Seven of the 22 OHTS centers (Henry Ford Medical Center, Troy, Mich;University of California, Davis, Sacramento; Devers Eye Institute, Portland,Ore; University of California, San Diego, La Jolla; Scheie Eye Institute,University of Pennsylvania, Philadelphia; Jules Stein Eye Institute, Universityof California, Los Angeles; and New York Eye and Ear Infirmary, New York)participated in the OHTS CSLO Ancillary Study. Informed consent for the OHTSCSLO Ancillary Study was obtained from all ancillary study participants afterinstitutional review board approval at their respective institutions. Fourhundred fifty-one persons with ocular hypertension provided informed consentfor participation in the OHTS CSLO Ancillary Study.

The Heidelberg Retina Tomograph (HRT) (Heidelberg Engineering, Dossenheim,Germany) uses confocal scanning diode laser technology to provide topographicalmeasurements of the optic disc and peripapillary retina. Details describingthis instrument^{8 - 12} andthe quality assessment, image processing, and data analysis of the CSLO imagesat the CSLO Reading Center have been presented elsewhere.¹⁵

In brief, only operators certified by the CSLO Reading Center at theUniversity of California, San Diego, acquired images for the OHTS CSLO AncillaryStudy. The HRT examinations were obtained once each year after the pupilswere dilated for the annual dilated fundus examination and photographing ofthe optic disc. Three 10° image series centered on the optic disc wereobtained for both eyes and a mean image was computed using HRT software version2.01. In addition, three 15° image series were obtained for the righteye. Magnification error was corrected using the participant's keratometrymeasurements. Corrective lenses were removed for HRT examination, unless theparticipant has an astigmatism exceeding 1.0 diopter (D). At an astigmatismof more than 1.0 D, corrective lenses were used during image acquisition.

All images were processed and reviewed for quality at the CSLO ReadingCenter.¹⁵ For this analysis, the first acceptablequality 10° or 15° mean topographic image for each eye was included.¹⁵ The optic disc margin was outlined on the mean topographyimage by a trained technician (A.R.S.) while viewing stereoscopic optic discphotographs of the optic disc taken within 12 months of the baseline images.Each outline of the optic disc was reviewed for accurate placement by a secondtrained technician (K.A.D.) with disagreements resolved by consensus.

Topographical parameters included with HRT software and investigatedin this study were as follows: mean cup depth, maximum cup depth, height incontour, height variation contour, mean height contour, cup shape, disc area,cup area, cup–disc area ratio, cup volume (below surface), rim area,rim volume (above reference plane), rim-disc ratio, retinal nerve fiber layerthickness, retinal nerve fiber layer cross section, and reference plane height.Cup shape is a measure of the overall 3-dimensional shape of the optic disccupping calculated as the third moment of the frequency distribution of depthvalues relative to the curved surface of measurements located within the outlineddisc margin. We also examined values from the discriminant analysis formulaof Mikelberg et al¹⁶ (the HRT classificationin current HRT software version 2.01). The HRT classification discriminantfunction value is used to classify an eye as "outside normal limits" or "withinnormal limits."

Information on baseline clinical characteristics including previoustopical ocular hypotensive medication; family history of glaucoma; and participanthistory of hypertension, heart disease, and diabetes mellitus was analyzedto evaluate possible differences by racial group. Similarly, baseline ocularcharacteristics including refractive status, corneal curvature, IOP, visualfield mean deviation, visual field pattern standard deviation (PSD), visualfield corrected pattern standard deviation (CPSD), and Optic Disc ReadingCenter stereophotographic assessment of horizontal and vertical cups–discdiameter ratio were compared in the African American and other racial groups.

Preliminary analyses showed an apparently nonlinear relationship betweenoptic disc area and other variables. A cubic B-spline¹⁷ withonly boundary knots seemed to account for the nonlinearities in optic discarea effects and was used to represent them in analyses in which optic discarea is considered a covariate. To control for confounders, differences inHRT topographic optic disc parameters by race were assessed using a mixed-effectslinear model¹⁸ in which "participant" is treatedas a random effect (and "eye" is nested within participant) and in which thereare fixed effects for race and optic disc area. This model was fitted usingthe restricted maximum likelihood criterion, and conditional F tests wereused to assess the significance of the fixed effects. Note that the modelused in both the unadjusted and adjusted analyses explicitly accounts forthe correlation of measurements taken on the 2 eyes of each participant. Calculationsused the "lme" function¹⁸ of the "nlme" package¹⁸ (version 3.1-36; see http://cran.us.r-project.org/src/contrib/PACKAGES.html#nlme) and the "geepack" package (version 0.2-4) under "R" (version 1.6.2).¹⁹

By self-attribution, 337 (75%) of the 451 participants with informedconsent were white, not of Hispanic origin; 75 (17%) were black, not of Hispanicorigin; 25 (5.5%) were Hispanic; and the balance were Native American, NativeAlaskan, Asian, Pacific Islander, or other. Table 1 lists the demographic and ocular characteristics of theoverall sample and by race, grouped by African American and all other racescombined. The mean (SD) age of the overall sample at enrollment was 54.4 (9.3)years. African Americans were more likely to be female and have hypertensionthan the other participants.

Baseline ocular characteristics that were used to establish eligibilityare noted because their range is truncated in the sample. Eye-specific measurementsreported in Table 1 are the averagesof both eyes. The baseline IOP for analytic purposes is the mean baselineIOP measurement, which was taken at the baseline-randomization examinationafter eligibility had been established in the qualifying assessment periodfor the OHTS. The baseline IOP measurement, which represents 2 or 3 IOP readingsper eye taken during the baseline-randomization examination,¹ was(mean [SD]) 25.2 (2.4) mm Hg (average of right and left eyes) for the entiresample, 25.0 (2.3) mm Hg for African Americans, and 25.3 (2.4) mm Hg for allof the other participants. Sixty-four percent of participants in the OHTSCSLO Ancillary Study had baseline IOP measurements greater than 24 mm Hg inboth eyes. The mean (SD) horizontal and vertical cup–disc diameter ratio(average of values for the right and left eyes) as determined by Optic DiscReading Center assessment of color stereophotographs was significantly greater(both P<.001) for African Americans (0.42 [0.18]and 0.46 [0.19], respectively) than for the other participants (0.35 [0.19],and 0.38 [0.19]), respectively. The Humphrey 30-2 visual field thresholdsof the 2 qualifying visual fields were averaged. While the CPSD of the qualifyingvisual fields did not differ by race, the PSD for African Americans (mean[SD], 2.06 [0.79] dB) was significantly different from that for the otherparticipants (1.95 [0.37] dB, P = .007), as was themean (SD) deviation (−0.02 [1.76] dB and 0.38 [1.39] dB, P = .002), respectively.

Overall, 40% of the OHTS CSLO Ancillary Study participants reporteda family history of glaucoma and 30% reported previous use of topical ocularhypotensive medication prior to study enrollment. Neither of these factorsdiffered significantly by race. Significantly more African Americans thanother participants reported a history of diabetes mellitus (P = .04) and hypertension (P<.001). Inaddition, African Americans differed from the other participants in this samplefor sex (P = .02), marital status (P<.001), and educational level (P<.001)(Table 1). These results are similarto those reported for the 1636 participants in the OHTS.¹

Of the 451 consenting participants, 11 participants discontinued participationor became inactive before HRT imaging was completed. In addition, 1 participantdid not have good-quality 10° images available for this analysis. Therefore,a total of 439 participants are included in this analysis of whom 74 (17%)are African American. As the OHTS clinical trial began before the fundingfor the OHTS CSLO Ancillary Study was approved, and 1 study center was added2 years later to increase African American participation, not all participantscompleted their imaging at the baseline OHTS visit. In addition, before fundingwas initiated, only 15° images were obtained on both eyes, so that insome cases, 10° images were not obtained at the first OHTS CSLO AncillaryStudy visit. Of the 439 participants with good-quality 10° or 15°images included in this analysis, 102 (23%) had images obtained at the baselinevisit, 157 (36%) at the 6- or 12-month visit, 127 (29%) at the 18- or 24-monthvisit, 44 (10%) at the 30- or 36-month visit, and 9 (2%) at later visits.The proportion of African Americans who had images obtained on or before the12-month visit after randomization was similar to that of the other participants,54% and 59%, respectively.

Topographic optic disc parameter measurements from the 439 participantswith usable images are given in Table 2. The mean SD of the mean topography image, a measure of image quality,was good in both African American and other eyes, 17.1 (8.1) µm, and18.5 (7.9) µm, respectively. A mean SD of less than 50 µm wasused as 1 measure of the quality of the acquired images.¹⁵ Statisticallysignificant differences were found between African Americans and other participantsfor most topographic optic disc parameters. African American participantshad significantly larger optic disc areas than the other participants, 2.17(0.41) mm² and 1.87 (0.38) mm², respectively (P<.001) (Figure 1).As indicated by the cup area, cup volume, and mean cup depth measurementsin the univariate, unadjusted analysis, African American eyes tended to havesignificantly larger and deeper cups (P≤.008)than the other participants. The African Americans also had significantlylarger cup-disc ratio (P = .048), neuroretinal rimarea (P<.001), and volume (P = .02) measurements than the other participants. The rim–discarea ratio was smaller in African American eyes than the other participants'eyes. No difference between African Americans and other racial groups wasfound for the mean values of cup shape (P = .66),retinal nerve fiber layer thickness (P = .56), HRTclassification value (P = .09), mean height contour(P = .08), and reference plane height (P = .07). However, after adjusting for optic disc area in the model,the differences in the other topographic optic disc parameter measurementsbetween African Americans and the other racial groups are greatly reducedand no longer reach statistical significance (Table 2).

Our study demonstrated in a large, well-characterized cohort of ocularhypertensive participants that African Americans have significantly largeroptic discs, neuroretinal rims, optic cups, and cup-disc ratios than the othergroups. Furthermore, this current study established that after adjusting foroptic disc area, differences in topographic optic disc parameters such ascup area, cup volume, rim area, rim volume, and rim-disc ratio between AfricanAmericans with ocular hypertension and other racial groups are reduced andno longer statistically significant.

These results in subjects with ocular hypertension confirm previousreports in healthy participants^{20 - 24} andpatients with glaucoma²⁵ that showed AfricanAmericans have significantly larger optic discs, optic cups, and cup-discratios than other racial groups. The difference between African Americansand other racial groups in cup–disc area ratio measured with the HRTcorresponds to differences in horizontal and vertical cup–disc diameterratio found by assessment of stereophotographs at the OHTS Optic Disc ReadingCenter.¹ In addition, these OHTS HRT resultsalso confirm reports that rim-disc ratio tends to be smaller in African Americaneyes,²⁴ despite larger disc areas in AfricanAmerican eyes. In contrast to these previous reports, the current study foundthat the differences in topographic optic disc parameters between ocular hypertensiveAfrican Americans and the other groups are largely explained by the largeroptic disc area in the African Americans.

It could be argued that adjusting for other optic disc parameters, suchas rim area or cup area also will explain the racial differences in opticdisc parameter measurements. However, unlike disc area, cup and rim area arealso features that are used to detect glaucoma, and both change with increasingseverity of the disease. Furthermore, several studies have shown that correctionfor optic disc size is clinically useful for detecting glaucoma.^{26 - 27} Forthese reasons, we adjusted for optic disc area in the analysis model.

Earlier reports recommended that different criteria for normal appearanceof the optic disc be used for African Americans and whites²⁰ sinceAfrican Americans generally have larger optic discs than other racial groups.The current study confirmed these recommendations and also highlighted thegrowing evidence of the importance for differential assessment of the opticnerve head based on optic disc size.^{26 ,28 - 30} Asubstantial number of participants not of African American race also had largeoptic discs. The challenge to the ophthalmologist is how best to judge thesize of the optic disc during a clinical examination. An accurate assessmentof disc area based on the clinical examination can be difficult to obtainsince both magnification of the eye and the lens used can influence the sizeof the optic disc as viewed by the clinician. Several slitlamp methods forestimating optic disc size have been proposed.^{31 - 33} Thesemethods can provide clinically useful classification of optic discs as small,medium, or large, but do not provide a reproducible or accurate estimate ofthe size of the optic disc.³³ Confocal scanninglaser ophthalmoscopy provides objective and quantitative information on discsize and other topographic optic disc parameter measurements that can assistthe clinician in determining whether the disc is glaucomatous or not.

There are conflicting reports on whether optic disc size influencessusceptibility to glaucomatous damage at a given level of IOP.²⁸ Over30 years ago it was hypothesized that greater pressure is exerted on a largedisc than it is on a smaller one at a given level of IOP.³⁴ Morerecent reports have provided evidence supporting this hypothesis.^{20 ,22} Other investigators, however, havenot found a relationship between optic disc size and susceptibility to glaucomatousdamage.³⁵ Furthermore, studies have found thateyes with small optic discs possess fewer optic nerve fibers than eyes withlarger optic discs.^{36 - 37}

Eyes with larger discs and more fibers have a larger anatomical reserveof neurons; they can lose more fibers before visual function is compromised.With its extensive follow-up of a well-defined ocular hypertensive populationof diverse racial backgrounds and measurement of optic disc area, the OHTSCSLO Ancillary Study provides a unique opportunity to evaluate the complexrelationship between IOP, optic disc area, race, and the development of glaucomatousoptic neuropathy.

This study demonstrated in a large cohort of subjects with ocular hypertensionthat African Americans have significantly larger optic discs, optic cups,neuroretinal rims, and cup-disc ratios than the other racial groups. Furthermore,this study established that differences in topographic optic disc parametersbetween African Americans with ocular hypertension and the other racial groupsare largely explained by the larger optic disc area in the African Americans.These results highlight the need to consider race and optic disc size whenevaluating the appearance of the optic disc in glaucoma.

Corresponding author and reprints: Linda M. Zangwill, PhD, DiagnosticImaging Laboratory, Hamilton Glaucoma Center and Department of Ophthalmology,University of California, San Diego, 9500 Gilman Dr, La Jolla, CA 92093 (e-mail: zangwill@eyecenter.ucsd.edu).

Submitted for publication June 10, 2003; final revision received August15, 2003; accepted September 10, 2003.

Dr Zangwill has received research support (equipment) from HeidelbergEngineering, Laser Diagnostic Technologies Inc, and Carl Zeiss Meditec. DrWeinreb has been a consultant or has received honoraria from Alcon Inc andNovartis AG; he also has been a consultant, received honoraria, grant support,or a patent (received or pending) from Allergan Speciality Therapeutics Inc,Heidelberg Engineering, Humphrey Instruments, Merck Research Laboratories,and Pharmacia & Upjohn Co; he owns stock in Merck & Co Inc. Dr Colemanhas received grant support from Alcon Laboratories Inc and has been a consultantor received honoraria from Allergan Specialty Therapeutics Inc and Pharmacia& Upjohn Co. Dr Piltz-Seymour has been a consultant or has received honorariafrom Merck & Co Inc and Pharmacia & Upjohn Co. Dr Liebmann has beena consultant or has received honoraria from Allergan Specialty TherapeuticsInc, Heidelberg Engineering, Norvartis AG, and Pharmacia & Upjohn Co.Dr Cioffi has been a consultant or has received honoraria from Merck &Co Inc and Alcon Laboratories Inc; he has been a consultant, received honoraria,grant support, or a patent from Allergan Ciba Vision, Heidelberg Engineering,Humphrey Instruments, Novartis AG, and Pharmacia & Upjohn Co. Dr Brandthas been a consultant or has received honoraria from Allergan Inc, Merck &Co Inc, and Pharmacia & Upjohn Co. Dr Kass has been a consultant or hasreceived honoraria from Merck & Co Inc and Pharmacia & Upjohn Co.

This study was supported by grants EY11158, EY09341, and EY09307 fromthe National Eye Institute and the National Center of Minority Health andHealth Disparities, National Institutes of Health, Bethesda, Md; Merck ResearchLaboratories, White House Station, NJ; and by an unrestricted grant from Researchto Prevent Blindness Inc, New York, NY.

Participating Clinics, Committees, and ResourceCenters in the OHTS CSLO Ancillary Study

Clinical Centers, Principal Investigators, ClinicCoordinators, and Staff

Devers Eye Institute,Portland, Ore: George A. Cioffi, MD, principal investigator (PI); StevenMansberger, MD; Kathryn Sherman, JoAnne M. Fraser, COT. Henry Ford Medical Center, Troy, Mich: Nauman R. Imami, MD (PI); DeborahDarnley-Fisch, MD; Aldo Fantin, MD; G. Robert Lesser, MD; Melanie Gutkowski,COMT, CO; Jim Bryant, COT; Amy Carroll; Ingrid C. Fugmann, COMT; Wendy Gilroy,COMT; Monica Lacoursiere; Sue Loomis, COT. Charles R. DrewUniversity, Jules Stein Eye Institute, University of California, Los Angeles: Anne L. Coleman, MD, PhD (PI); Richard S. Baker, MD; Michelle C. Banks,MD; Y. P. Dang, MD; Satvinder K. Gujral, MD; Simon K. Law, MD; Dana P. Tannenbaum,MD; Thao T. Thach, MD; Mary T. Gonzales; Bobbie Ballenberg, COMT; SalvadorMurillo; Jackie Sanguinet; Manju Sharma. New York Eye &Ear Infirmary, New York: Jeffrey M. Liebmann, MD (PI); Robert Ritch,MD; Celso A. Tello, MD; Jean L. Walker, BS, COA; Deborah L. Simon, COA. Scheie Eye Institute, University of Pennsylvania, Philadelphia: Jody R. Piltz-Seymour, MD (PI); Teresa L. Brevetti, MD; Andrew Mohammed,MD; Jane L. Anderson, MS, CCRC; Cheryl McGill, COA; Janice T. Petner, COA. University of California, Davis, Sacramento: James D. Brandt,MD (PI); Sean D. Adrean, MD; Shaun Brierly, MD; Jeffrey J. Casper, MD; JohnT. Dragicevich, MD; Thomas R. Johansen, MD; Michele C. Lim, MD; Michael B.Mizoguchi, MD; Alan M. Roth, MD; Ivan R. Schwab, MD; Ingrid J. Clark, COA;Denise M. Owensby, BS, COA; Marilyn A. Sponzo, COA. Universityof California, San Diego, La Jolla: Robert N. Weinreb, MD (PI); J.Rigby Slight, MD; Tess A. Abunto, MD.

Resource Centers, Principal Investigators, ClinicCoordinators, and Staff

Confocal Scanning LaserOphthalmoscopy Reading Center, University of California, San Diego: RobertN. Weinreb, MD (PI); Linda Zangwill, PhD; Keri Dirkes, MPH. Ocular Hypertension Treatment Study Coordinating Center, Washington UniversitySchool of Medicine, St Louis, Mo: Mae O. Gordon, PhD (PI); Steven Kymes,PhD; J. Philip Miller, AB; Ellen Long, CCRA; Patricia Morris; Ann K. Wilder,RN, BSN, CCRP. Ocular Hypertension Treatment Study Chairman'sOffice, Washington University School of Medicine: Michael A. Kass,MD (PI); Deborah Dunn. Project Office, National Eye Institute,Rockville, Md: Donald F. Everett, MA.

Executive and Steering Committee

MichaelA. Kass, MD (chair); Ingrid Adamson, MD; Douglas R. Anderson, MD, George A.Cioffi, MD; Donald F. Everett, MA; Douglas Gaasterland, MD; Mae E. Gordon,PhD; Dale K. Heuer, MD; Eve J. Higginbotham, MD; Chris A. Johnson, PhD; JohnL. Keltner, MD; Richard K. Parrish II, MD; M. Roy Wilson, MD; Pamela Frady,COMT, CCRC; Patricia A, Morris; Ann K. Wilder, RN, BSN.

Barry Davis,MD, PhD (chair); Ingrid Adamson, MD; Roy Beck, MD, PhD; John Connett, PhD;Claude Cowan, MD; Donald F. Everett, MA; Mae O. Gordon, PhD; Michael A. Kass,MD; Ronald Munson, PhD; Mark Sherwood, MD; Gregory L. Skuta, MD.