ABSTRACT

Purpose: To determine patient demographics and theocular biometric parameters in patients presenting forcataract surgery within the public hospital system, in adefined New Zealand population.

Method: Prospective study of 502 eyes of 488 consecutivepatients undergoing cataract surgery. A clinical assessment,including refraction, keratometry (K),A-scan ultrasound andOrbscan II computerized topography was performed oneach eye.

Results: The mean age of the group was 74.9 ± 9.8 years(mean ± SD) with a female predominance (62%). Ethnicorigin included 72% European, 8% Maori, 10% PacificIslander, 4% Asian, 3% Indian and 3% other ethnic origins.The mean Log MAR visual acuity of eyes prior to cataractsurgery was 0.88 ± 0.57 (approximately 6/48–1). Cornealtopographic (keratometric) maps were classified into fivegroups: 34% round, 10% oval, 31% symmetrical bow tie,12% asymmetrical bow tie and 13% irregular. The meansteepest K measurement was 44.1 ± 1.7 D, the median keratometric astigmatism 0.89 D (range 0.0–6.5 D) and thesteepest corneal meridian was horizontal in 50% and verti-cal in 43%. Seven per cent of corneas were spherical.Refraction revealed a mean sphere of 0.0 ± 3.1 D and amean cylinder of –1.2 (range 0.0–7.5 D). Refractive astig-matism was with-the-rule in 15%, against-the-rule in 50%and oblique in 15%, with 20% spherical. Axial length was amean of 23.14 ± 1.03 mm.

Conclusion: Patients presenting for cataract surgery in thisstudy were predominantly elderly, female, of EuropeanCaucasian ethnicity and exhibited relatively poor corrected

visual acuity in the affected eye. Interestingly, 41% of eyesdemonstrated bow-tie topographic patterns, largelyexhibiting with-the-rule astigmatism. However, assessmentby keratometry or refraction highlighted against-the-rulemore frequently; this may have implications for combinedcataract and astigmatic surgery.The mean axial length wasslightly shorter than expected for a group of predominantlyEuropean ethnic origin, although the mean refractive errorwas emmetropic.

Key words: cataract, demographics, Maori, New Zealand,ocular biometry, Orbscan corneal topography.

INTRODUCTION

Within New Zealand/Aotearoa, cataract surgery is the mostfrequently performed elective surgery.1 Approximately halfof these cataract surgeries are performed in the public healthsystem. The demographics of patients presenting for cataractsurgery in the public system have not previously beenanalysed in a prospective manner. The anatomical structuresof the eye are known to vary with demography. However, noprior studies have considered whether the varied ethnicity ofthe New Zealand population affects trends in ocular biomet-ric parameters with respect to computerized corneal topog-raphy, keratometry, refraction and axial length.

New generation computerized slit-scanning topographi-cal systems have the advantage of obtaining informationfrom the both the front and rear surface of the cornea. Incomparison to purely Placido-based topography systems,this new technology has greater accuracy,2,3 allowing avariety of interactive analyses using proprietary software. Inthe present study, combined slit-scanning and Placido-basedtopography was utilized to assess corneal topography.

The purpose of this study was to assess the demographicsand ocular biometric trends, including computerized corneal

Clinical and Experimental Ophthalmology (2001) 29, 381–386

Original Article

The Auckland Cataract Study: demographic, corneal topographic and ocular biometric parametersAndrew F Riley MBChB, Christina N Grupcheva MD, Tahira Y Malik FRCOphth, Jennifer P Craig PhD andCharles NJ McGhee PhD FRANZCODiscipline of Ophthalmology, University of Auckland, Auckland, New Zealand

� Correspondence: Prof Charles NJ McGhee, Discipline of Ophthalmology, Faculty of Medical and Health Science, University of Auckland, Private Bag 92019,

New Zealand. Email: c.mcghee@auckland.ac.nz

Presented as a scientific poster at the Royal Australian College of Ophthalmologists Annual Scientific Congress, Sydney, November 2000.

topography, in a defined population presenting for cataractsurgery within the New Zealand public health system. Theresults may provide a useful foundation for comparativestudies and planning future surgical interventions.

METHODS

The Auckland Cataract Study is a large, prospective obser-vational study of 502 consecutive cataract operations. Thisproject is coordinated by the University of Auckland Dis-cipline of Ophthalmology in collaboration with AucklandPublic Hospital Department of Ophthalmology (AucklandHealthcare Services Ltd). For the purpose of this componentof the study, patients presenting for routine cataract surgerycompleted a questionnaire, were clinically assessed and hadtheir biometric parameters measured. The local AucklandMedical Ethics committee approved this research projectand methodology.

Patients invited to take part in the study had been placedon the reserve waiting list for cataract surgery at AucklandHospital between January 1997 and March 2000 and livedin the greater Auckland metropolitan region. Access tocataract surgery in the New Zealand public service is basedon National Clinical Priority Assessment Criteria (CPAC) asa part of New Zealand Governmental National ReferralGuidelines. In this system, points are allocated for the levelof reduced best spectacle corrected acuity (BSCVA) in theeye with cataract, in relation to the BSCVA of the fellow eyeand in the context of concurrent ocular diseases which maybe more appropriately treated following (e.g. diabeticretinopathy), or that may adversely affect the outcome of(e.g. age-related macular degeneration), cataract surgery. Anumber of social components are also considered such assystemic illness, being a caregiver and car driving. Prior tothe commencement of this study these patients were justbelow the points-threshold for treatment. Due to increasedavailability of funding for cataract surgery, patients on thisreserve waiting list still wishing to be considered for surgerywere asked to attend the University of Auckland Disciplineof Ophthalmology for preoperative assessment. Of 533patients reviewed, 502 eyes of 488 individuals were enrolledin the study and underwent cataract surgery betweenJanuary and September, 2000. Forty-five of the reviewedpatients did not participate as a result of serious illness,migration into another region or death.

Patients were asked to complete a questionnaire whereethnicity data and current medications were obtained. Visualacuity was measured using a single new, calibrated, 4-m illuminated LogMAR chart (University of Otago). Asingle examiner (AFR) obtained full ophthalmic and medicalhistory and undertook an ophthalmic (anterior segment slit-lamp examination, tonometry and dilated fundoscopy) andmedical examination. In addition to the routine ophthalmicexamination, keratometry (manual Topcon OM-4; Topcon,Tokyo, Japan), auto-kerato-refractometry (Topcon KR.8100),corneal elevation topography (Orbscan II; Bausch and Lomb,Rochester, NY, USA) and hand-held A-scan ultrasound

for axial length determination (Tomey AL-2000; Tomey,Erlangen, Germany), were performed on both eyes of allpatients. A minimum of 10 axial length recordings weremade for each eye and the mean calculated.

Orbscan II corneal topography was performed after cali-bration with the supplied test object. All patients wereseated with their chin placed on the chin rest and a sup-porting strap placed around the head. The Orbscan IIcorneal topography system incorporates both a Placido-based topographical assessment and a slit-scanning facilityfor anterior and posterior corneal curvature analysis. For thepurpose of map classification in this study, axial keratometric(power) maps with 1-D incremental colour steps were used.

Two ophthalmologists (AFR, CNG), experienced incorneal topographic analysis, classified the axial keratometrymaps independently. If there was a difference in classificationbetween the examiners the classification was discussed and aconsensus reached. The classification scheme used was basedon that originally described by Bogan et al. (Fig. 1)4 and modified by Rabinowitz et al. (Fig. 2).5 Horizontal (within ±45 degrees of 180 degree meridian) and vertical bow tie patterns (within ± 45 degrees of the vertical meridian) werealso differentiated in the present study.

For keratometry, refraction and biometry data, means andmedians were determined. For the topographic and kerato-metric data, astigmatism was divided into three groups, vertical (within ± 45 degrees of 90 degree meridian), horizontal (within ± 45 degrees of 180 degree meridian) andwithout astigmatism (spherical).

382 Riley et al.

Figure 1. Classification after Bogan et al.4 Differentiation of fivepatterns based on 1-D colour steps with the steepest colour chosenthat occupied greater than 10% of the central two-thirds of thecornea. Round: The ratio of the shortest to the longest diameter atthe colour zone chosen for the pattern reading was two-thirds ormore. Oval: The ratio of the shortest to the longest diameter at thecolour zone chosen for the pattern reading was less than two-thirds. Symmetrical bow tie: (a) There was a central constriction in theoutline of the colour zone identified for the pattern reading; (b) theratio X0/X2 or X0/X1 is one-third or less; (c) the ratios X1/X2 andY1/Y2 are two-thirds or more. Y represents the diameter of the longaxis of the bow tie and X represents the diameter of each half of thebow tie perpendicular to the long axis. Asymmetrical bow tie: Criteriaa and b for symmetrical bow tie are met; one or both of the ratiosdescribed in criteria c for symmetrical bow tie was less than two-thirds. Irregular: No clear pattern could be identified according tothe above criteria.

Statistical analysis was performed using the SPSS softwarepackage (Statistical Program for Social Scientists).

RESULTS

The prospective study group included 502 eyes of 488 con-secutive individuals with a mean age of 74.9 years (range34.7–94.3 years) and a marked female preponderance (62%).From the questionnaire response, the ethnic backgroundconsisted of 72% European Caucasians, 8% Maori, 10%Pacific Islander, 4% Asian (East of India), 3% Indian and 3%non-specified. There were 249 right eyes and 253 left eyes.

A number of eyes (n = 33) were excluded from topo-graphical analysis due to incomplete map patterns or inabil-ity to perform Orbscan because of poor patient health orcooperation. In addition, the study group also included eyes(n = 18) that had undergone previous surgery, or had ocular

pathology that altered corneal topography, and these mapswere also excluded from further evaluation. Therefore, com-puterized corneal topography maps of 451 (89.8%) eyes,obtained prior to surgery, were classified according tomethods described in Fig. 1. The maps were categorizedinto the five groups as follows: 34.1% round, 10.2% oval,30.8% symmetrical bow tie, 11.5% asymmetrical bow tieand 13.3% irregular. These data are shown in Fig. 3 and sub-divided according to Rabinowitz et al. in Table 1. We furthersubdivided the bow tie (astigmatic) patterns and found 99 of154 (64%) symmetrical bow ties and 26 of 52 (50%) asym-metrical bow ties exhibited a ‘with-the-rule’ distributionbetween 45° and 135°.

Manual keratometry readings were obtained on all 502eyes and the results are shown in Table 2. The mean steep-est K reading in dioptres was 44.1 ± 1.7 D (mean ± SD, range38.6–50.5 D) and the mean astigmatism was 0.9 ± 0.8 D(range 0.0 to –6.5 D, median 0.8 D). The distribution of theaxis in the steepest meridian was 50% horizontal, 43% ver-tical and only 7% having no measurable astigmatism.

Refractive error was measured in 435 of 502 eyes beforeand after surgery. The actual refractive error, based onspherical equivalent, was classified into six groups (Table 3).Overall mean spherical refractive error was 0.0 ± 3.1 D(range +7.5 D to –18.0 D and median +0.25 D) and meancylinder was –1.22 ± 1.03 D (range 0.0 to –7.5 D andmedian –1.0 D; Table 4). Refractive astigmatism was with-the-rule in 15%, against-the-rule in 50%, and oblique in15% of eyes with the remaining 20% being spherical. Dueto density of cataract, poor vision and inability to obtainsubjective refraction, no results were available on theremaining 67 eyes.

The mean axial length (n = 502), based on a mean of 10measurements of the eyes listed for cataract surgery, was23.14 ± 1.03 mm (range 20.20–28.32 mm). The axial lengthof the fellow eye was 23.13 ± 0.99 mm (range 20.52–28.78mm; Table 5). There was no statistically significant differ-ence between the two groups of eyes (independent Student’st-test P = 0.7). The axial length of eyes of male patients wassignificantly longer than the eyes of female patients (23.45± 0.94 mm and 22.94 ± 1.03 mm, respectively, P < 0.001).Matching of the male : female ratio at 1:1, resulted in a meanaxial length of 23.18 ± 1.02 mm. The axial length for Maori

The Auckland Cataract Study 383

Figure 2. Classification afterRabinowitz et al.5 Steepest colourthat occupied at least 10% of thecentral two thirds was used todetermine the classification. Abow tie is considered skewedwhen the smaller of the anglesbetween the lobes is less than150 degrees.

Figure 3. Distribution of keratometric topographic maps(n = 451) after the classification according to Bogan et al.4 ( )Round, n = 154; ( ) oval, n = 46; (�) symmetrical bow tie,n = 139; ( ) asymmetrical bow tie, n = 52; (�) irregular, n = 60.

(n = 41) and Pacific islanders (n = 50) was 23.17 ± 1.03 mmand 23.25 ± 0.75 mm, respectively; these were not signifi-cantly different from those of the European population(23.16 ± 1.03 mm). Although other ethnic groupings (Asianand Indian) did not demonstrate any statistical difference inaxial length compared to the European population, thesegroups were too small to enable useful statistical comparison.

DISCUSSION

In this study we have defined several demographic andocular biometric characteristics for a group of New Zealandpatients presenting for cataract surgery in the public hospi-tal service. To our knowledge there are no prior publicationsin the literature on this population. This population of rela-tively elderly patients with a female predominance included18% of Polynesian descent. Access to cataract surgery isconstrained by a Government-directed points system inNew Zealand and the visual acuity in the eye to be operatedon was worse than 6/24 in 53% of our patients. However,the threshold for cataract surgery in the Australasian regionhas reduced to 6/6–6/9 in symptomatic patients.6,7

Our observations demonstrate that the mean axial length(23.14 ± 1.03 mm) is slightly shorter in the present studythan the findings in previous studies for other populations.8,9

Indeed, in a large study by Hoffer of 6950 phakic eyes, amean axial length of 23.65 ± 1.35 mm and average kerato-metric readings of 43.81 ± 1.6 D were recorded.10 Thegender of the patients was not documented in Hoffer’sstudy, whereas in the present study, female eyes were

more frequent and statistically shorter than male eyes(23.45 ± 0.94 mm), thus influencing the overall mean measurement.

Hand-held ultrasound biometry measurements, as per-formed in the present study, have been compared to slit-lamp supported biometry and the two techniques have beenshown to give comparable results and are considered accept-able alternatives.11 According to the literature, a minimumof three readings should be taken and the mean calculated to enhance results.12 To maximize accuracy in the presentstudy, a minimum of 10 axial length measurements weremade and the mean recorded. Nonetheless, theoretically thesingle longest reading might be more accurate in identifyingmaximum axial length, as corneal flattening has been postu-lated as a source of error when using the manual ultrasoundprobe. Interestingly, as no significant differences in theultrasound measurement of axial length, either in relation tocataract density or morphology, have been established,13 therelatively advanced cataracts encountered in some patientsin this study should not have affected individual or overallmean measurements.

An average keratometric reading of 43.7 ± 1.6 D wasrecorded in 502 eyes of this study, which is comparable withthe average of 43.97 ± 1.54 D identified by Bogan et al.4Interestingly, in a previously reported population-basedstudy a high level of uncorrected refractive error was identified, and this rate increased with age and presence ofcataract.14 The mean refractive error in the present studygroup was very close to emmetropia (0.0 ± 3.1 D). How-ever, a wide range of spherical equivalent (+7.5 D to –18.5 D)was identified with a relatively large standard deviation,

384 Riley et al.

Table 1. Distribution of keratometric topographic maps (n = 451)after the classification according to Rabinowitz et al.5

Topographic map No. eyes (%)

Round 133 (29)Oval 46 (10)Symmetrical bow tie 128 (28)Symmetrical bow tie 11 (2)

with skewed radial axisAsymmetrical bow tie 13 (3)

with superior steepeningAsymmetrical bow tie 36 (8)

with inferior steepeningAsymmetrical bow tie 3 (1)with skewed radial axis

Superior steepening 8 (2)Inferior steepening 13 (3)Irregular 60 (13)

Table 3. Refraction in spherical equivalent (dioptres)* for 435eyes, before and after cataract surgery

Spherical equivalent Preoperative Postoperative(dioptres) refraction refraction

≥ +6.0 1 0+1.0 to +5.75 147 (33.8%) 22 (5.0%)± 0.75 136 (31.3%) 298 (68.5%)–1.0 to –5.75 149 (34.2%) 115 (26.4%)≤ –6.0 2 0

For the purposes of comparison the groups have been defined asemmetropia (± 0.75 D), low to moderate myopia (–1.00 to –5.75 D), low to moderate hypermetropia (+1.00 to +5.75 D), andhigh myopia or high hypermetropia (≤ –6.00 D or ≥ +6.00 D,respectively). *All spherical equivalents rounded to the nearest 0.25 D.

Table 2. Manual keratometry readings in dioptres (n = 502 eyes)

Mean SD Median Minimum Maximum

Steepest meridian 44.13 1.69 44.12 38.60 50.50Flattest meridian 43.25 1.64 43.25 36.50 47.75Astigmatism 0.89 1.71 0.75 0.00 6.50

indicating a large number of eyes with a refractive error,some of which may be due to cataract. The distribution ofthe spherical equivalent of the refractive error before surgerywas almost equal for the three main groups: emmetropia(31%), mild to moderate myopia (34%) and mild to moder-ate hypermetropia (34%). However, after cataract surgeryemmetropia was most prevalent (69%), myopia wasdecreased (26%), and only 5% of patients had mild to moderate hypermetropia. These data correlate well with thetendency for phacoemulsification surgery to be used tocorrect existing refractive error and our intention to providea final refractive end point between emmetropia and –0.50 Dexcept in patients in whom it would produce unacceptableanisometropia.

The number of patients with against-the-rule astigmatismon keratometry and refraction was greater than the numberof patients with with-the-rule astigmatism. This is consistentwith a previously identified age-related shift in axis.15–18

The relative distribution of subgroups following classifica-tion of corneal topography, after Bogan et al.4 and Rabinowitzet al.,5 was different to that previously described but thesestudies were conducted with generally younger populations.Comparisons are presented in Table 6. The five-category

classification proposed by Bogan et al. was based on a popu-lation of 212 patients with a mean age of 37.1 ± 13.6 years.4Rabinowitz et al., in a population of 195 normal patients, pro-posed a classification system that resulted in an increase overthe Bogan system from five to 10 categories.5 In respect tocorneal ectasias and other corneal pathologies this lattersystem is more useful. However, there was little value to begained by an expanded classification system in the relatively‘normal’ population assessed in the present study and we utilized a modified classification system with seven predom-inant patterns. This included separating bow ties into horizontal and vertical groups, which offers additional infor-mation on overall anterior corneal astigmatism.

Within the subgroup of eyes exhibiting bow tie astigma-tism (of at least 1 D), we identified that 61% of bow ties hada vertical axis. In contrast, against-the-rule astigmatism wasthe predominant pattern when considering astigmatismmeasured by keratometry and refraction in this study group.However, unlike refraction and keratometry based analyses,bow tie patterns are often non-orthogonal as descriptors ofastigmatism. Therefore, in considering each limb of the bowtie, the classification might encompass oblique and with-the-rule or oblique and against-the-rule categories. In order

The Auckland Cataract Study 385

Table 4. Pre-operative refraction in dioptres in minus cylinder form (n = 435 eyes)

Mean SD Median Minimum Maximum

Sphere –0.001 3.10 0.25 –18.00 7.50Cylinder –1.22 1.03 –1.00 –7.50 0.00Spherical equivalent –0.49 2.96 0.00 –18.63 6.88

Table 5. Axial length in millimetres as measured by hand-held A-scan ultrasound (n = 502)

Mean SD Median Minimum Maximum

Operated eye 23.14 1.03 23.08 20.20 28.31Other eye 23.13 0.99 23.02 20.52 28.78

Table 6. Comparison of keratometric and topographic map classification between the present study and those of Bogan et al.4 andRabinowitz et al.5

Bogan et al. Rabinowitz et al. Present study(n = 212) (n = 195) (n = 451)

Proportion with classification (%)Round 22.6 41.3 34Oval 20.8 20.8 10Symmetrical bow tie 17.5 21.8 31Asymmetrical bow tie 32.1 10.2 12Irregular 7.1 5.9 13

Keratometric and refractive findings (mean ± SD; D)Keratometry 43.97 ± 1.54 – 43.69 ± 1.64Keratometric astigmatism 0.80 ± 0.70 – 0.88 ± 0.78Spherical equivalent –1.00 ± 2.02 – –0.49 ± 2.96Refractive cylinder 0.48 ± 0.77 – –1.22 ± 1.03

The combined bow tie astigmatism patterns were detected in 49.6%, 32% and 43% of studies, respectively, although mean refractivecylinder (–1.2 D) was greatest in the present study.

to clarify, and simplify this issue, bow tie categories were classified as with-the-rule when both limbs were within 45°of the vertical meridian and against-the-rule when bothlimbs were within 45° above or below the horizontal merid-ian. Therefore, although comparison of topographic/keratometric and refractive astigmatism highlights trends,variations in the size of the subgroups studied (Tables 1,2,4)and differences in classification of axis prevent true quanti-tative comparison of such data in this study. Although pos-terior corneal power and lenticular astigmatism have notbeen analysed in this group, these might explain some of theincongruity identified when comparing trends in topo-graphy, refraction and keratometry based astigmatism.Other patterns with elements of astigmatism in the topo-graphy pattern, but into which the astigmatism was not categorized in this study, are oval and unclassified topo-graphies and corneas with less than 1 D of cylinder.

The New Zealand population presenting for cataractsurgery in the public health system is characterized by spe-cific demographic and ocular biometric parameters. Themajority of eyes exhibited astigmatism as identified by auto-refraction (80%), manual keratometry (93%) and Orbscanslit-scanning topography (43% ≥ 1 D). Although the mostcommon topographic pattern was round, the previouslydescribed predominance of against-the-rule astigmatism inolder populations is confirmed by keratometry and refrac-tion in this study and may influence the types of incisionperformed in cataract surgery19,20 or the indications foradditional astigmatic procedures. In terms of planning ofastigmatic surgery, differences in the magnitude and axis ofastigmatism were noted between topography, keratometryand refraction, although only the latter measures total ocularastigmatism. Axial length is the most important parameter inthe choice of intraocular lens implant,21 and in the presentstudy, as with previous studies, the mean axial length wassignificantly longer in males than in females.22 Interestingly,the overall mean axial length was less than that reported for non-New Zealand populations, although no statisticaldifferences were noted between Maori, Pacific Island andEuropean-Caucasian eyes.

ACKNOWLEDGEMENTS

This Auckland Cataract Study was funded in large part by anunrestricted grant from Auckland Healthcare Services Ltd.Dr CN Grupcheva was supported by a fellowship from theMaurice and Phyllis Paykel Trust. We acknowledge withdeep gratitude the invaluable support and assistance ofadministrative, nursing and medical staff within AucklandPublic Hospital and the University of Auckland Disciplineof Ophthalmology.

REFERENCES

1. Elder M, Tarr K, Leaming D. The New Zealand cataract andrefractive surgery survey 1997/1998. Clin. Exp. Ophthalmol.2000; 28: 89–95.

2. Yaylali V, Kaufman SC, Thompson HW. Corneal thicknessmeasurements with the Orbscan Topography System and ultra-sonic pachymetry. J. Cataract Refract. Surg. 1997; 23: 1345–50.

3. Srivannaboon S, Reinstein DZ, Sutton HF, Holland SP.Accuracy of Orbscan total optical power maps in detectingrefractive change after myopic laser in situ keratomileusis. J. Cataract Refract. Surg. 1999; 25: 1596–9.

4. Bogan SJ, Waring GO, Ibrahim O, Drews C, Curtis L. Classifi-cation of normal corneal topography based on computer-assisted videokeratography. Arch. Ophthalmol. 1990; 108: 945–9.

5. Rabinowitz YS, Yang H, Brickman Y et al. Videokeratographydatabase of normal human corneas. Br. J. Ophthalmol. 1996; 80:610–16.

6. Keeffe JE, Taylor HR. Cataract surgery in Australia1985–1994. Aust. N.Z. J. Ophthalmol. 1996; 24: 313–17.

7. McCarty CA, Keeffe JE, Taylor HR. The need for cataractsurgery. Projections based on lens opacity, visual acuity, andpersonal concern. Br. J. Ophthalmol. 1999; 83: 62–5.

8. Hoffer KJ. Axial dimension of the human cataractous lens.Arch. Ophthalmol. 1993; 111: 914–18.

9. Sayegh FN. The correlation of corneal refractive power, axiallength, and the refractive power of the emmetropizingintraocular lens in cataractous eyes. Ger. J. Ophthalmol. 1996; 5:328–31.

10. Hoffer KJ. Biometry of 7,500 cataractous eyes. Am. J.Ophthalmol. 1980; 90: 360–68.

11. Whelehan IM, Heyworth P, Tabandeh H, McGuigan S, FossAJ. A comparison of slit-lamp supported versus hand-held biometry. Eye 1996; 10: 514–16.

12. Butcher JM, O’Brien C. The reproducibility of biometry andkeratometry measurements. Eye 1991; 5: 708–11.

13. Liou HL, McCarty CA, Jin CL, Taylor HR. Prevalence andpredictors of undercorrected refractive errors in the Victorianpopulation. Am. J. Ophthalmol. 1999; 127: 590–96.

14. Connell B, Brian G, Bond MJ. A case-control study of bio-metry in healthy and cataractous Eritrean eyes. OphthalmicEpidemiol. 1997; 4: 151–5.

15. Fledelius HC, Stubgaard M. Changes in refraction and cornealcurvature during growth and adult life. A cross-sectional study.Acta Ophthalmol. 1986; 64: 487–91.

16. Hayashi K, Masumoto M, Fujino S, Hayashi F. Changes incorneal astigmatism with aging. Acta Soc. Ophthalmol. Jap. 1993;97: 1193–6.

17. Ninn-Pedersen K. Relationships between preoperative astig-matism and corneal optical power, axial length, intraocularpressure, gender, and patient age. J. Refract. Surg. 1996; 12:472–82.

18. Hayashi K, Hayashi H, Hayashi F. Topographic analysis of thechanges in corneal shape due to aging. Cornea 1995; 14: 527–32.

19. Roman S, Givort G, Ullern M. Choice of the site of incisionfor cataract surgery without suture according to preoperativeastigmatism. J. Fr. Ophthalmol. 1997; 20: 673–9.

20. Roman S, Ullern M. Astigmatism caused by superior and tem-poral corneal incisions in cataract surgery. J. Fr. Ophthalmol.1997; 20: 277–83.

21. Olsen T. Sources of error in intraocular lens power calculation.J. Cataract Refract. Surg. 1992; 18: 125–9.

22. Osuobeni EP. Ocular components values and their inter-correlations in Saudi Arabians. Ophthalmol. Physiol. Optics 1999;19: 489–97.

386 Riley et al.