Factors Associated withChildhood StrabismusFindings from a Population-Based Study

Dana Robaei, MBBS, MPH,1 Kathryn A. Rose, DAppSc(Orth), PhD,2 Annette Kifley, MBBS, MAppStat,1

Michael Cosstick, BAppSc,1,2 Jenny M. Ip, MBBS,1 Paul Mitchell, MD, PhD1

Purpose: To describe strabismus prevalence and associated factors in a representative sample of 6-year-old Australian children.

Design: Population-based cross-sectional study.Participants: One thousand seven hundred thirty-nine predominantly 6-year-old children resident in Sydney

examined in 2003 and 2004.Methods: Cover testing was performed at near and distance fixation, and with spectacles if worn. Logarithm

of the minimum angle of resolution visual acuity was measured in both eyes before and after pinhole correction,after correcting any cylindrical refraction �0.50 diopters and with spectacles, if worn. Cycloplegic autorefraction(cyclopentolate) and detailed dilated fundus examination were performed. Each child’s medical and perinatalhistories were sought in a detailed parental questionnaire.

Main Outcome Measures: Strabismus was defined as any heterotropia at near or distance fixation, or both,on cover testing. Microstrabismus was defined as a deviation of fewer than 10 prism diopters.

Results: Strabismus was diagnosed in 48 children (2.8% of the population), 5 of whom had previously under-gone surgical correction; 26 children (54%) had esotropia, 14 (29%) had exotropia, 7 (15%) had microstrabismus, and1 child had VIth cranial nerve palsy. Prematurity was associated with a 5-fold increase in the risk of esotropia (oddsratio, 5.0; 95% confidence interval, 1.8–14.1). Visual impairment (with presenting correction) was significantly morecommon in children with (22.9%) than without (1.3%) strabismus (P�0.0001). The presence of strabismus wassignificantly associated with hyperopia, astigmatism, anisometropia, and amblyopia (P�0.0001).

Conclusions: This report documents the prevalence of strabismus and its relation to other ocular signsand visual impairment in a representative sample of Australian school children. Presence of strabismus wassignificantly associated with prematurity. Ophthalmology 2006;113:1146 –1153 © 2006 by the American

Academy of Ophthalmology.

Strabismus is a relatively common ocular disorder in child-hood populations. It is the most common cause of amblyo-pia,1–4 and as such contributes importantly to childhoodvisual impairment. Further, there is now growing evidenceon the psychosocial impact of strabismus, with 2 studies

Originally received: August 19, 2005.Accepted: February 13, 2006. Manuscript no. 2005-781.1 Department of Ophthalmology, Centre for Vision Research, WestmeadHospital, and Westmead Millennium Institute, Sydney, Australia.2 School of Applied Vision Sciences, Faculty of Health Sciences, Univer-sity of Sydney, Sydney, Australia.

The authors have no proprietary interests related to the article.

The Sydney Myopia Study is supported by the National Health & MedicalResearch Council, Canberra, Australia (grant no.: 253732); WestmeadMillennium Institute, University of Sydney, Sydney, Australia; and VisionCo-operative Research Centre, University of New South Wales, Sydney,Australia.

Correspondence to Paul Mitchell, MD, PhD, Department of Ophthalmol-ogy, Center for Vision Research, University of Sydney, Westmead Hos-pital, Hawkesbury Road, Westmead, NSW, Australia, 2145. E-mail:

paul_mitchell@wmi.usyd.edu.au.

1146 © 2006 by the American Academy of OphthalmologyPublished by Elsevier Inc.

reporting that strabismus may create a significant negativesocial prejudice5–7 and can significantly reduce an appli-cant’s ability to obtain (or hold) employment.8 In addition,affected adolescents and adults report that strabismus neg-atively affects their self-image and interpersonal relation-ships with others.9 Negative attitudes toward strabismusemerge at a young age, as early as 6 years, as shown in 1study.10 For these reasons, continued epidemiological re-search into the prevalence and risk factors for strabismus(particularly factors that could identify children at risk) andthe impact of strabismus on the visual function of youngchildren is of public health importance.

The literature on childhood strabismus and its manage-ment is substantial; Table 1 summarizes the prevalence ofstrabismus in a number of countries across America, Eu-rope, and Asia. To our knowledge, there have been 2previous studies reporting the prevalence of strabismus inAustralian children,11,12 although both did so as part of ageneral report on childhood ocular abnormalities, and nei-ther attempted an analysis of associated factors. There is

scant literature from large, population-based childhood

ISSN 0161-6420/06/$–see front matterdoi:10.1016/j.ophtha.2006.02.019

Japan

Robaei et al � Strabismus in Sydney Children

samples both within and outside Australia examining fac-tors associated with the presence or development of strabis-mus.13,14 Such research has potential public health implica-tions. Our study aimed to address this deficiency byreporting associations of strabismus with a range of demo-graphic, socioeconomic, and gestational factors in a large,

Table 1. Selected Past Pre

Author (Publication Year) Age Group

School-based studiesFrandsen, (1960)*

a. Childcare centers and preschools 0–6b. Schools for mentally able 7–20c. Schools for mentally handicapped 8–15

Kornder et al (1974)† 6Macfarlane et al (1987)‡ 6–11Preslan and Novak (1996)§ 4–7Lithander et al (1998)� 6 and 1Junghans et al (2002)¶ 3–12Nepal et al (2003)# 5–16

Population-based studiesGraham et al (1974)** 5–6Friedman et al (1980)†† 1–2.5Maul et al (2000)‡‡ 5–15Zhao et al (2000)§§ 5–15Pokharel et al (2000)�� 5–15Ohlsson et al (2001)¶¶ 12–13Kvarnstrom et al (2001)## 10Murthy et al (2002)*** 5–15Dandona et al (2002)††† 7–15Naidoo et al (2003)§§§ 5–15He et al (2004)��� 5–15Goh et al (2005)¶¶¶ 7–15Matsuo and Matsuo (2005)### 6–12

*Frandsen AD. Occurrence of a squint: a clinical-statistical study on the prpopulation. Acta Ophthalmol (Copenh) 1960;62(suppl):9–157.†Kornder LD, Nursey JN, Pratt-Johnson JA, Beattie A. Detection of man1974;77:211–4.‡Macfarlane DJ, Fitzgerald WJ, Stark DJ. The prevalence of ocular diso1987;15:161–74.§Preslan MW, Novak A. Baltimore Vision Screening Project. Ophthalmo�Lithander J. Prevalence of amblyopia with anisometropia or strabismus1998;76:658–62.¶Junghans B, Kiely PM, Crewther DP, Crewther SG. Referral rates for achildren. Ophthalmic Physiol Opt 2002;22:10–25.#Nepal BP, Koirala S, Adhikary S, Sharma AK. Ocular morbidity in scho**Graham PA. Epidemiology of strabismus. Br J Ophthalmol 1974;58:22††Friedman Z, Neumann E, Hyams SW, Peleg B. Ophthalmic screeningOphthalmol Strabismus 1980;17:261–7.‡‡Maul E, Barroso S, Munoz SR, et al. Refractive Error Study in Childre§§Zhao J, Pan X, Sui R, et al. Refractive Error Study in Children: results��Pokharel GP, Negrel AD, Munoz SR, Ellwein LB. Refractive Error2000;129:436–44.¶¶Ohlsson J, Villarreal G, Sjostrom A, et al. Visual acuity, residual amblyin Sweden. Acta Ophthalmol Scand 2001;79:589–95.##Kvarnstrom G, Jakobsson P, Lennerstrand G. Visual screening of Swedish c***Murthy GV, Gupta SK, Ellwein LB, et al. Refractive error in ch2002;43:623–31.†††Dandona R, Dandona L, Srinivas M, et al. Refractive error in children‡‡‡Prevalence at near and distance fixation reported separately.§§§Naidoo KS, Raghunandan A, Mashige KP, et al. Refractive error and vSci 2003;44:3764–70.���He M, Zeng J, Liu Y, et al. Refractive error and visual impairment in u¶¶¶Goh PP, Abqariyah Y, Pokharel GP, Ellwein LB. Refractive error aOphthalmology 2005;112:678–85.###Matsuo T, Matsuo C. The prevalence of strabismus and amblyopia in

population-based sample of young children.

Patients and Methods

Population

The Sydney Myopia Study is a population-based survey of refrac-tion and other eye conditions in a sample of predominantly 6-year-

nce Studies of Strabismus

Sample Size (n) Prevalence (%)

3570 4.510537 4.61418 9.12619 4.5

877 2.5680 3.1

6541 0.92697 0.31100 1.6

4784 5.338000 1.35303 9.95884 2.85067 2.11046 2.73126 3.16447 0.54074 1.8–1.9‡‡‡

4890 1.1–1.3‡‡‡

4364 1.9–3.0‡‡‡

4634 0.786531 1.3

ce of squint and associated signs in different groups and ages of the Danish

trabismus in young children. 2. A retrospective study. Am J Ophthalmol

in 1000 Queensland primary schoolchildren. Aust N Z J Ophthalmol

1996;103:105–9.ng schoolchildren in the Sultanate of Oman. Acta Ophthalmol Scand

ional vision screening among a large cosmopolitan sample of Australian

ildren in Kathmandu. Br J Ophthalmol 2003;87:531–4.

8,000 children, age 1 to 2 1/2 years, in child welfare clinics. J Pediatr

ults from La Florida, Chile. Am J Ophthalmol 2000;129:445–54.Shunyi District, China. Am J Ophthalmol 2000; 129:427–35.y in Children: results from Mechi Zone, Nepal. Am J Ophthalmol

nd ocular pathology in a screened population of 12–13-year-old children

n: an ophthalmological evaluation. Acta Ophthalmol Scand 2001;79:240–4.in an urban population in New Delhi. Invest Ophthalmol Vis Sci

rural population in India. Invest Ophthalmol Vis Sci 2002;43:615–22.

impairment in African children in South Africa. Invest Ophthalmol Vis

children in southern China. Invest Ophthalmol Vis Sci 2004;45:793–9.isual impairment in school-age children in Gombak District, Malaysia.

ese elementary school children. Ophthalmic Epidemiol 2005;12:31–6.

vale

(yrs)

2

evalen

ifest s

rders

logyamo

funct

olch4–31.

of 3

n: resfromStud

opia a

hildreildren

in a

isual

rbannd v

old children resident in the Sydney metropolitan area of Australia.

1147

Ophthalmology Volume 113, Number 7, July 2006

This study in 6-year-old children is the first of 3 projects examin-ing the eye health of children in younger and older age groups.Collectively, these projects are called the Sydney Childhood EyeStudy. The methods used to identify and select the target sample, aswell as its description and study procedures, were reported recently.15

In brief, all schools within the Sydney metropolitan area werestratified by socioeconomic status using Australian Bureau ofStatistics 2001 national census data. These data were used torandomly select 34 primary schools across Sydney, including 5primary schools in the top socioeconomic decile, with the remain-ing 29 schools selected from the bottom 9 socioeconomic deciles.A proportional mix of public and private/religious schools wasincluded. The following report is based on data from predomi-nantly 6-year-old school children examined between August,2003, and October, 2004.

ProceduresWritten consent from at least 1 parent in addition to the assent ofeach child was obtained before examination. Approval for thestudy was obtained from the Human Research Ethics Committee,University of Sydney; Department of Education and Training,New South Wales; church education bodies; and private schools.

The presence of strabismus was ascertained in a standardizedmanner by experienced orthoptists who performed a Hirschbergtest, followed by the cover–uncover test to detect manifest stra-bismus. If no strabismus was elicited, the alternating cover test todetect heterophoria was performed. Measurement of the size ofany deviation used the prism bar cover test. The prism strength indiopters (D) that neutralized any movement of the eyes wasrecorded as the size of the deviation. This cover test sequence wascarried out on all children at near (33 cm) and distance (6 m)fixation, with and without spectacles, if worn. Distance visualacuity was tested in each eye separately using a logarithm ofminimum angle of resolution chart. The chart was retroilluminatedwith automatic calibration to 85 cd/m2 (VectorVision CSV-1000,VectorVision, Inc., Dayton, OH) and read at 8 feet (244 cm).Visual acuity was assessed with and without spectacle correction,if worn, and with a 1.2-mm pinhole aperture for reduced vision(�20/25) or if there was more than a 1-line (5-letter) differencebetween the 2 eyes. In addition, noncycloplegic autorefraction wasperformed on all children with any reduced vision, and any cyl-inder larger than 0.50 D was corrected by subjective refraction. Amatching HOTV card was available for children unable to namethe letters. For each eye, visual acuity was recorded as the numberof letters read correctly from 0 (�20/200) to 70 (20/10). If noletters could be read at 8 feet (244 cm), the chart was moved to 3feet (91 cm), giving 4 additional levels of visual acuity: 20/250,20/320, 20/400, and 20/500. If no letters could be identified on thechart, visual acuity was assessed as counting fingers at 2 feet (61cm), hand movements, perception of light, or no perception oflight.

Cycloplegia was obtained after 2 cycles of cyclopentolate 1%(1 drop) and tropicamide 1% (1 drop) instilled 5 minutes apart,after corneal anesthesia with amethocaine hydrochloride 1%. In asmall proportion of children whose pupils were slow to dilate,phenylephrine hydrochloride 2.5% was used to maximize mydri-asis. Cycloplegia was considered full when the pupil was fixed andthe diameter was 6 mm or more.

A Canon autorefractor (model RK-F1, Canon Inc., Tokyo,Japan) was used to perform cycloplegic autorefraction and kera-tometry. Children also had a comprehensive eye examinationassessment of ocular movements, stereopsis screening using theLang II test, color vision assessment, ocular biometry, slit-lampexamination, optical coherence tomography, and mydriatic digital

retinal photography.

1148

Parents were asked to complete a comprehensive 193-itemquestionnaire. Sociodemographic information covering ethnicity,country of birth, education, occupations, and parental age wasincluded. Information about the child’s birth and medical historyas well as maternal obstetric history was sought. Maternal smokinghistory also was sought in the questionnaire.

Definitions

Strabismus was defined as any heterotropia at near or distance fixa-tion, or both. Microstrabismus was defined as a deviation of fewerthan 10 prism diopters in the presence of demonstrable binocularvision on the Lang II test. In the absence of demonstrable binocularvision, deviations of this magnitude were simply classified as strabis-mus. Table 2 shows further classification of esotropia and exotropiainto various subtypes, consistent with recommendations by the Com-mittee for Classification of Eye Movement Abnormalities and Stra-bismus, sponsored by the National Eye Institute.16

Myopia was defined as spherical equivalent (SE) refraction of�0.50 D or less. Hyperopia was defined as SE refraction of �2.0D or more, and was deemed significant at �3.0 D or more.Astigmatism was defined as cylinder of 1.0 D or more and aniso-

Table 2. Definitions of Selected Strabismus Subtypes

Subtype Definition

Comitant esotropiaRefractive accommodative Esotropia is decreased by �10� when

wearing full hyperopic correctionin a child with a hyperopicrefractive error.

Non-refractiveaccommodative

Comitant esotropia at near only,eliminated by hyperopic correctionat near.

Mixed accommodative Esotropia at distance and near, withfully corrected near deviation beinglarger than distance deviation by�10�.

Nonaccommodative (basic) Esotropia is decreased by �10� whenwearing full hyperopic correctionin a child with a hyperopicrefractive error.

Sensory Comitant esotropia in an eye withvision loss resulting from eye and/or brain disease.

Infantile esotropia syndrome Large angle, early onset (�6 monthsage), constant esotropia in aneurologically healthy child.

Comitant exotropiaBasic Near deviation within 10� of

distance deviation.Divergence excess Distance deviation �10� greater

than near deviation.Convergence weakness Near deviation �10� greater than

distance deviation.Sensory Comitant exotropia in an eye with

vision loss resulting from eye and/or brain disease.

Infantile exotropia syndrome Large angle, early onset (�6 monthsage), constant exotropia in aneurologically healthy child.

Incomitant strabismus Incomitant esotropia or exotropiawith evidence of cranial nervedysfunction, neuromuscular,muscular, or orbital disease.

� � prism diopters.

Robaei et al � Strabismus in Sydney Children

metropia as SE refraction difference between the 2 eyes of at least1.0 D. Absence of significant ametropia was defined as SE refrac-tion of more than �0.50 D to less than �2.0 D. Amblyopiainitially was defined as corrected visual acuity less than 0.3 loga-rithm of minimum angle of resolution units (i.e., fewer than 40letters; equivalent to Snellen acuity less than 20/40) in the affectedeye, not attributable to any underlying structural abnormality of theeye or visual pathway, together with a difference of at least 2logarithm of minimum angle of resolution lines between the 2eyes. When amblyopia was suspected on grounds of minimallyreduced corrected visual acuity in the presence of an amblyogenicrisk factor, corroborative evidence from the parental questionnairewas sought. Parental report of lazy eye or amblyopia was takeninto account for each potential amblyopic child, as were reports ofany previous patching treatment.

The child’s ethnic origin was derived from that of both parents.Categories of ethnicity included all major ethnic groups in Aus-tralia. Parental education was defined as the highest level ofeducation completed by either parent. This ranged from neverhaving attended school to having completed a higher degree suchas a master’s or doctoral degree. Socioeconomic status was basedon home ownership by the child’s parents as well as their employ-ment status. Parents also were asked to extract birth data from theirchildren’s health record booklet.

Data Handling and Statistical AnalysisData were entered into a Microsoft Access database (Microsoft,Redmond, WA). All statistical analyses were performed usingStatistical Analysis System software, version 8.2 (SAS Institute,Cary, NC). Mixed models and generalized estimating equationswere used to adjust for clustering within schools. Where clustereffects were not significant, chi-square tests and t tests were used.All sample means and mean differences are reported with theirstandard errors. Odds ratios (ORs) and 95% confidence intervals(CIs) are presented.

Results

PatientsOf 2238 eligible children, 1765 (78.9%) children were givenparental permission to participate and questionnaire data wereprovided by parents. Of this group, 25 were not examined becausethey were absent from school during the week of the survey. Covertest findings were available for 1739 of the remaining 1740 chil-dren. The mean age of participants was 6.7 years (range, 5.5–8.4years); 49.4% of children were female and 50.6% were male. Most(70.4%) were 6 years of age, whereas one quarter (25.5%) were 7years of age.

Prevalence of StrabismusStrabismus was diagnosed in 48 children (2.8% of the population),5 of whom previously had undergone surgical correction; 26children (54%) had esotropia and 14 (29%) had exotropia, all ofthe comitant type, and 7 (15%) had microstrabismus, 1 of whomhad microesotropia. There were no cases of isolated vertical de-viation, whereas 11 cases were noted in combination with hori-zontal strabismus, most commonly esotropia. Only 1 child hadincomitant strabismus; the presence of VIth cranial nerve palsy inthis child was likely the result of previous ventriculoperitonealshunt failures and the need for multiple reinsertions. Further sub-classification of the types of strabismus is shown in Table 3.

Fixation was preferred by the left eye in 17 children (35.4%) and

the right eye in 18 children (37.5%), whereas the remaining 13children had alternating strabismus. Of the 48 strabismus patients,14 (29.2%) were previously undiagnosed, with most of these (11patients; 78.6%) having intermittent strabismus.

Demographic, Socioeconomic, and GestationalAssociations with Strabismus

Findings on sociodemographic and perinatal characteristics ofchildren with and without strabismus are shown in Table 4. Thepresence of strabismus was not significantly associated with gen-der (P � 0.74) or measures of socioeconomic status such asparental education (P � 0.67), employment status (P � 0.29), orhome ownership (P � 0.45). It was, however, significantly lowerin children with nonwhite than white ethnicity; further subgroupanalysis revealed this to be true for esotropia only. Children ofnonwhite ethnicity were 3 times less likely to have esotropia thanwhite children (OR, 0.34; 95% CI, 0.15–0.78) because of thelower prevalence of significant hyperopia (�3.00 D or more) inthis group (2.4% compared with 4.6% in white children; P �0.01).

Presence of strabismus was associated with a number of ges-tational factors, as shown in Table 5. Significant associations werefound with prematurity (gestational age, 36 weeks or less), lowbirth weight (�2500 g), and parent-reported admission to a neo-natal intensive care unit, although the strength of the relationshipwas not always comparable for esotropia and exotropia. Prematu-rity, for example, conferred a 5-fold increase in the prevalence ofesotropia (OR, 5.0; 95% CI, 1.8–14.1); the effect of prematurity onexotropia was not assessable. Similarly, paternal age more than 2standard deviations from the mean was associated with a 5-foldincrease in the risk of esotropia (OR, 4.9; 95% CI, 1.6–15.0), butnot exotropia (OR, 1.0; 95% CI, 0.1–6.8). Crude trends showedthat this relationship was probably related to older, rather thanyounger, paternal age.

Given the significant correlation between gestational factors, all

Table 3. Proportion of Strabismus by Type

Category n (%)

Esotropia 26 (54.2)Constant

Basic nonaccommodative 10Refractive (full) accommodative 3Refractive (partial) accommodative 3Mixed accommodative 3

IntermittentConvergence excess 4Divergence insufficiency 2Basic 1

Exotropia 14 (29.2)Constant 1Intermittent

Basic 6Convergence insufficiency 4Divergence excess 3

Microtropia 7 (14.6)Microexotropia 6Microesotropia 1

Vertical deviations 11 (22.9)In isolation 0In combination with esotropia 8In combination with exotropia 3

Incomitant strabismusVIth cranial nerve palsy 1 (2.1)

factors were examined in multivariable-adjusted models, and sig-

1149

Ophthalmology Volume 113, Number 7, July 2006

nificant or confounding factors were retained. After adjustment forage and nonwhite ethnicity, extremes of paternal age and lack ofbreastfeeding were significantly associated with presence of stra-bismus, whereas admission to a neonatal intensive care unit wasmarginally nonsignificant after adjusting for birth weight. The

Table 5. Strabismus Prevalence in Ch

Associated FactorAssociated F

Present, n (

Male gender 23 (2.6)Nonwhite ethnicity

Esotropia 5 (0.8)Exotropia 8 (1.3)

Gestational factorsBirth weight �2500 g 7 (7.6)

Esotropia 4 (4.5)Exotropia 2 (2.3)

Gestational period (�37 wks) 8 (6.4)Esotropia 7 (5.7)Exotropia 0 (0.0)

Admission to NICU 9 (9.7)Esotropia 4 (4.6)Exotropia 4 (4.6)

Maternal smoking during pregnancy 8 (4.2)Maternal illness during pregnancy 7 (3.1)Extremes of maternal age at birth

(�2 SD from mean)3 (4.8)

Esotropia 2 (3.2)Exotropia 1 (1.6)

Extremes of paternal age at birth(�2 SD from mean)

5 (6.3)

Esotropia 4 (5.1)Exotropia 1 (1.3)

History of breastfeeding 34 (2.5)

Table 4. Comparison of Selected Sociodemographic and Pe

Characteristic

Characterwith

(n

Sociodemographic characteristicsMean age (yrs) (CI) 6.8 (Gender (male) [n (%)] 23 (Ethnicity [n (%)]

White 35 (Southeast Asian 7 (Other 6 (

Parental employment (both employed) [n (%)] 18 (Home ownership [n (%)] 1106 (Parental education (tertiary) [n (%)] 749 (

Perinatal characteristicsMean maternal age at birth (yrs) (CI) 30.5 (Mean paternal age at birth (yrs) (CI) 35.3 (Mean birth weight (g) (CI) 3298 (First-born child [n (%)] 20 (Premature birth (�36 weeks) [n (%)] 8 (Low birth weight (�2500 g) [n (%)] 7 (Caesarean section [n (%)] 13 (Admission to NICU [n (%)] 9 (Maternal smoking in pregnancy [n (%)] 8 (Passive maternal smoking in pregnancy [n (%)] 5 (Maternal illness in pregnancy [n (%)] 7 (

CI � confidence interval; NICU � neonatal intensive care unit.

NICU � neonatal intensive care unit; N/A � not assessabl

1150

ethnicity association, however, became nonsignificant after adjust-ing for significant hyperopia. An interaction between low birthweight and breastfeeding status was observed; 38.5% of low birthweight children who were not breastfed had strabismus, comparedwith only 2.6% of those who were breastfed (P � 0.02). After

with and without Associated Factors

Associated FactorAbsent, n (%)

Odds Ratio(95% Confidence Interval)

25 (2.9) 0.89 (0.48–1.63)

22 (2.0) 0.34 (0.15–0.78)12 (1.1) 1.10 (0.42–2.87)

31 (2.2) 3.49 (1.32–9.22)18 (1.3) 3.35 (0.91–12.34)13 (1.0) 2.46 (0.30–20.02)34 (2.4) 2.82 (1.25–6.34)16 (1.1) 5.04 (1.80–14.11)18 (1.3) N/A34 (2.4) 4.25 (1.62–11.15)18 (1.3) 3.60 (1.05–12.36)16 (1.1) 4.07 (1.08–15.32)37 (2.6) 1.51 (0.67–3.39)41 (2.7) 1.13 (0.40–3.17)40 (2.7) 1.57 (0.55–4.48)

21 (1.4) 1.98 (0.67–5.80)18 (1.2) 1.22 (0.13–11.45)33 (2.3) 2.74 (0.93–8.07)

15 (1.1) 4.91 (1.61–15.03)17 (1.3) 1.06 (0.16–7.05)11 (5.1) 0.49 (0.20–1.20)

al Characteristics in Children with and without Strabismus

n Childrenismus8)

Characteristic in Childrenwithout Strabismus

(n � 1692) P Value

.0) 6.7 (6.6–6.8) 0.19857 (50.6) 0.74

1072 (63.4) 0.0005292 (17.3) 0.5327 (19.3) 0.08838 (53.4) 0.2926 (71.9) 0.4520 (45.7) 0.67

32.5) 30.6 (29.9–31.2) 0.9037.2) 33.6 (33.0–34.2) 0.08–3566) 3377 (3333–3420) 0.56

697 (49.0) 0.81117 (7.8) 0.007985 (5.9) 0.0065

292 (19.1) 0.0484 (5.7) 0.0013

181 (12.4) 0.65160 (10.6) 0.70219 (13.0) 0.81

ildren

actor%)

rinat

istic iStrab

� 4

6.6–747.9)

72.9)14.6)12.5)45.2)66.0)47.9)

28.4–33.3–303147.1)19.4)18.3)30.3)20.9)14.9)8.6)14.5)

e; SD � standard deviation.

Robaei et al � Strabismus in Sydney Children

adjustment for age, ethnicity, and parental age at birth, this inter-action was no longer significant (P � 0.1).

Ocular Characteristics of Children withStrabismus

Table 6 summarizes findings on selected ocular characteristics andoutcomes in children with and without strabismus. Children withstrabismus were significantly more hyperopic than children with-out, with a mean SE refraction of �2.46 D (95% CI, �1.85 to�3.07 D) versus �1.23 D (95% CI, �1.16 to �1.29 D). Theprevalence of all significant refractive errors in the right eye wasgreater in children with than without strabismus; 3 children withstrabismus (6.4%) were myopic, 13 (27.5%) were significantlyhyperopic, and 8 (17.2%) were anisometropic; astigmatism waspresent in 12 (25.4%). Corresponding rates in children withoutstrabismus were as follows: 22 (1.3%) were myopic (P � 0.02), 37(2.3%) were significantly hyperopic (P�0.0001), 20 (1.2%) wereanisometropic (P�0.0001), and 72 (4.2%) had astigmatism(P�0.0001). Strabismus alone (or in combination with anisome-tropia) was the most common potential cause of amblyopia in oursample, being associated with 56.3% of all cases. The amblyopiaprevalence was predictably higher in children with strabismus,present in 18 (37.6%) strabismic and 14 (0.9%) nonstrabismicchildren (P�0.0001). Given the higher prevalence of significantrefractive errors and amblyopia in strabismic children, the meanuncorrected visual acuities for both the better and worse eyes alsowere significantly lower in this group than in nonstrabismic chil-dren. The same also was true for mean corrected visual acuity.

Table 6. A Comparison of Selected Ocular Cha

Characteristic

Characterwith

(

Mean uncorrected visual acuity (logMARletters)

Worse eyeBetter eye

Mean corrected visual acuity (logMARletters)

Worse eyeBetter eye

Corrected visual impairment (�20/40 inworse eye) [n (%)]

1

Mean spherical equivalent (D)

Mean cylinder (D)Astigmatic axis [n (%)]

With the rule 1Against the ruleOblique 1

Myopia [n (%)]Hyperopia (��3.00 D) [n (%)] 1Astigmatism [n (%)] 1Anisometropia [n (%)]Amblyopia [n (%)] 1

1

Full binocular stereoscopic vision [n (%)] 1

D � diopters; logMAR � logarithm of the minimal an*In children with esotropia.†In children with exotropia.

Visual impairment (visual acuity worse than 20/40) in the worse

eye after optical correction was found in 11 strabismic children(22.9%), compared with only 21 nonstrabismic children (1.3%;P�0.0001).

Discussion

Estimates of the prevalence of strabismus in the currentliterature of population-based studies have ranged from0.5% in an urban sample of Indian children17 to 9.9% in aChilean sample18; our rate of 2.8% is lower than that pre-viously reported for other white populations.14,19,20 Thismay be explained partly by the ethnic heterogeneity of oursample, with a significant proportion being of East Asianbackground. A lower prevalence of strabismus among chil-dren of East Asian ethnicity has been reported in a numberof studies, some of which are cited in Table 1.21–23

The presence of strabismus in our sample was signifi-cantly lower in children of nonwhite ethnicity. Further sub-classification of cases revealed this to be true only foresotropia; a similar finding has been reported for blackAmerican children compared with white children.24 Ahigher prevalence of hyperopia among white children thanamong nonwhite children recently was established in theRefractive Error Study in Children, a population-basedstudy sponsored by the World Health Organization.18,25,26

Similarly, the prevalence of significant hyperopia (�3.00 D

ristics in Children with and without Strabismus

in Childrenismus8)

Characteristic in Childrenwithout Strabismus

(n � 1692) P Value

48.8 �0.000150.7 �0.0001

49.4 �0.000151.0 �0.0001

9) 21 (1.3) �0.0001

�1.23 �0.0001* —

2† —�0.29 �0.0001

3) 319 (31.9) 0.431) 268 (28.8) 0.571) 368 (38.7) 0.16) 22 (1.3) 0.0245) 37 (2.3) �0.00014) 72 (4.2) �0.00012) 20 (1.2) �0.00016) 14 (0.9) �0.00012)* — �0.00010)† — �0.00010) 1557 (93.6) �0.0001

f resolution.

racte

isticStrab

n � 4

39.345.3

44.248.3

1 (22.

�2.46�3.27�1.4�0.61

1 (25.8 (25.8 (49.3 (6.43 (27.2 (25.8 (17.8 (37.3 (48.4 (20.8 (60.

gle o

or more) in our sample was significantly higher in white

1151

Ophthalmology Volume 113, Number 7, July 2006

than nonwhite children (4.6% vs. 2.4%; P � 0.01). Thehigher prevalence of esotropia among white children in oursample therefore is likely to be related to a close linkbetween accommodative esotropia and hyperopia. As ex-pected, the association between white ethnicity and esotro-pia disappeared after adjusting for this confounding vari-able. The prevalence of exotropia was not significantlydifferent among white and nonwhite children in our sample.Our study, however, did not have adequate power to examinea difference specifically between white and East Asian chil-dren. Larger population-based studies will be needed to eluci-date true differences in the prevalence of the various types ofstrabismus among children of diverse ethnic backgrounds,independent of specific refractive errors.

Prevalence variability across studies in part relates tohow cases are ascertained and by whom. Reliance on aretrospective review of clinical records, albeit made byophthalmologists, is limited by the nonuniform documenta-tion of findings and the tests performed. Although exami-nation by orthoptists has been questioned,27 it remains areliable method of strabismus examination in the context ofa large population-based study. Australian orthoptists arehighly trained and also typically perform more cover testexaminations in clinical settings than most comprehensiveophthalmologists. Although a confirmation of all of ourfindings by a pediatric ophthalmologist would have beenideal, it was not possible in this study because of its size andpopulation-based nature, with examinations at many differ-ent schools.

The presence of strabismus was significantly associatedwith prematurity and low birth weight, confirming findingsfrom a number of previous studies.13,24,28,29 This effect wasespecially true for premature children who were not breast-fed; this subgroup was probably more premature and there-fore unable to be breastfed. Nonetheless, the interaction ofprematurity and breastfeeding on strabismus developmentmay be of clinical importance, and therefore warrants fur-ther investigation. Subclassification of cases into esotropiaand exotropia revealed similar magnitudes of associationwith prematurity, low birth weight, and admission to aneonatal intensive care unit, all of which conferred a 3- to5-fold increase in the odds of both esotropia and exotropia.There is, to the best of our knowledge, only 1 previous studyinvestigating the differential influence of birth weight onesotropia and exotropia in the 1 sample.24 Low birth weightwas uniformly associated with an increase in risk for bothesotropia and exotropia in that study.

The pathogenesis of the different types of strabismus isnot well known, partly because most previous studies hadinsufficient power to examine key associations with esotro-pia and exotropia separately. A high prevalence of strabis-mus in children with known central nervous system abnor-malities has been widely established in the literature,30–34

and possible noninclusion of many developmentally de-layed children in this school population could account inpart for the relatively low prevalence of strabismus. Currentevidence suggests that damage during the critical period ofbrain development is likely to be an important determinantof both types of strabismus. This critical period begins with

rapid multiplication of cells at midgestation and ends with

1152

the myelination of neuronal cells, a process that is notcomplete until 3 to 4 years after birth.35 Although thisinterval spans the onset of most types of strabismus,36–38

the method and timing of the insult may be importantdeterminants in the type of strabismus that develops.

Finally, within our sample of children, all refractiveerrors, that is, myopia, significant hyperopia, astigmatism,and anisometropia, were significantly more frequent in chil-dren with than without strabismus. Predictably, these chil-dren also were much more likely to have amblyopia orresidual visual impairment after optical correction than chil-dren without strabismus. Mean corrected visual acuity in thebetter eye was significantly lower in children with thanwithout strabismus, largely a result of undercorrected re-fractive error (predominantly hyperopia) in the nonstrabis-mic eye. Although these findings are not new, they point tochildhood strabismus as being a condition that is unlikely tobe present in isolation. Inclusion of cover testing by trainedexaminers in childhood vision screening programs thereforenot only would lead to an increased detection of strabismus,but also is likely to also point to a specific group of childrenwho are at high risk of having significant (often uncor-rected) refractive errors and reduced visual acuity.

The main strength of this study is its population-baseddesign. Our high response rate allows findings to be reliablygeneralized to Australian school children. The standardizedexamination protocol reduces the likelihood of spuriousmeasurement variability. Further, this study includes previ-ously undiagnosed cases of strabismus, otherwise unac-counted for in studies based on retrospective review ofclinical records.

Our study also has some limitations. First, cover testfindings were not confirmed by a pediatric ophthalmologist,although children not previously under care were referred toan ophthalmologist. For the reasons mentioned above, wedo not believe that this has significantly compromised ourfindings. Second, our study probably had inadequate powerto detect some known associations with strabismus. Forexample, we were unable to find a significant relationshipbetween strabismus and maternal smoking during preg-nancy, shown in a number of previous studies.13,24,39 Inaddition, we had insufficient information on family historyof strabismus, which also has been shown to increase therisk of childhood strabismus.40,41

In conclusion, this study documents the prevalence ofstrabismus and associated factors in a population-basedsample of predominantly 6-year-old children. In attemptingto address the influence of perinatal factors on the develop-ment of esotropia and exotropia separately, it became clearthat further research in this area is required.

References

1. Oliver M, Nawratzki I. Screening of pre-school children forocular anomalies. II. Amblyopia. Prevalence and therapeuticresults at different ages. Br J Ophthalmol 1971;55:467–71.

2. Yassur Y, Yassur S, Zaifrani S, et al. Amblyopia amongAfrican pupils in Rwanda. Br J Ophthalmol 1972;56:368–70.

3. Shaw DE, Fielder AR, Minshull C, Rosenthal AR. Amblyopia—

factors influencing age of presentation. Lancet 1988;2:207–9.

Robaei et al � Strabismus in Sydney Children

4. Preslan MW, Novak A. Baltimore Vision Screening Project.Ophthalmology 1996;103:105–9.

5. Uretmen O, Egrilmez S, Kose S, et al. Negative social biasagainst children with strabismus. Acta Ophthalmol Scand2003;81:138–42.

6. Olitsky SE, Sudesh S, Graziano A, et al. The negative psychos-ocial impact of strabismus in adults. J AAPOS 1999;3:209–11.

7. Satterfield D, Keltner JL, Morrison TL. Psychosocial aspectsof strabismus study. Arch Ophthalmol 1993;111:1100–5.

8. Coats DK, Paysse EA, Towler AJ, Dipboye RL. Impact oflarge angle horizontal strabismus on ability to obtain employ-ment. Ophthalmology 2000;107:402–5.

9. Menon V, Saha J, Tandon R, et al. Study of the psychosocialaspects of strabismus. J Pediatr Ophthalmol Strabismus 2002;39:203–8.

10. Paysse EA, Steele EA, McCreery KM, et al. Age of theemergence of negative attitudes toward strabismus. J AAPOS2001;5:361–6.

11. Junghans B, Kiely PM, Crewther DP, Crewther SG. Referralrates for a functional vision screening among a large cosmo-politan sample of Australian children. Ophthalmic Physiol Opt2002;22:10–25.

12. Macfarlane DJ, Fitzgerald WJ, Stark DJ. The prevalence ofocular disorders in 1000 Queensland primary schoolchildren.Aust N Z J Ophthalmol 1987;15:161–74.

13. Hakim RB, Tielsch JM. Maternal cigarette smoking duringpregnancy. A risk factor for childhood strabismus. Arch Oph-thalmol 1992;110:1459–62.

14. Graham PA. Epidemiology of strabismus. Br J Ophthalmol1974;58:224–31.

15. Ojaimi E, Rose KA, Smith W, et al. Methods for a population-based study of myopia and other eye conditions in schoolchildren: the Sydney Myopia Study. Ophthalmic Epidemiol-ogy 2005;12:59–69.

16. National Eye Institute. A classification of eye movement ab-normalities and strabismus (CEMAS). Report of a NationalEye Institute-sponsored workshop. 2001:1-56. Available at:http://www.nei.nih.gov/news/statements/cemas.pdf. AccessedJune 23, 2005.

17. Murthy GV, Gupta SK, Ellwein LB, et al. Refractive error inchildren in an urban population in New Delhi. Invest Oph-thalmol Vis Sci 2002;43:623–31.

18. Maul E, Barroso S, Munoz SR, et al. Refractive Error Study inChildren: results from La Florida, Chile. Am J Ophthalmol2000;129:445–54.

19. Frandsen AD. Occurrence of a squint: a clinical-statisticalstudy on the prevalence of squint and associated signs indifferent groups and ages of the Danish population. ActaOphthalmol (Copenh) 1960;62(suppl):9–157.

20. Kornder LD, Nursey JN, Pratt-Johnson JA, Beattie A. Detec-tion of manifest strabismus in young children. 2. A retrospec-tive study. Am J Ophthalmol 1974;77:211–4.

21. He M, Zeng J, Liu Y, et al. Refractive error and visualimpairment in urban children in southern China. Invest Oph-

thalmol Vis Sci 2004;45:793–9.

22. Goh PP, Abqariyah Y, Pokharel GP, Ellwein LB. Refractiveerror and visual impairment in school-age children in GombakDistrict, Malaysia. Ophthalmology 2005;112:678–85.

23. Matsuo T, Matsuo C. The prevalence of strabismus and am-blyopia in Japanese elementary school children. OphthalmicEpidemiol 2005;12:31–6.

24. Chew E, Remaley NA, Tamboli A, et al. Risk factors for esotro-pia and exotropia. Arch Ophthalmol 1994;112:1349–55.

25. Zhao J, Pan X, Sui R, et al. Refractive Error Study in Children:results from Shunyi District, China. Am J Ophthalmol 2000;129:427–35.

26. Pokharel GP, Negrel AD, Munoz SR, Ellwein LB. RefractiveError Study in Children: results from Mechi Zone, Nepal.Am J Ophthalmol 2000;129:436–44.

27. Govindan M, Mohney BG, Diehl NN, Burke JP. Incidence andtypes of childhood exotropia: a population-based study. Oph-thalmology 2005;112:104–8.

28. Blazso S, Giesel V. Correlation between strabismus and cen-tral nervous system injuries. J Pediatr Ophthalmol Strabismus1971;8:18–22.

29. Bremer DL, Palmer EA, Fellows RR, et al. Strabismus inpremature infants in the first year of life. Arch Ophthalmol1998;116:329–33.

30. Kristjansdottir R, Sjostrom A, Uvebrant P. Ophthalmologicalabnormalities in children with cerebral white matter disorders.Eur J Paediatr Neurol 2002;6:25–33.

31. Jacobson L, Ygge J, Flodmark O, Ek U. Visual and perceptualcharacteristics, ocular motility and strabismus in children withperiventricular leukomalacia. Strabismus 2002;10:179–83.

32. Pennefather PM, Tin W. Ocular abnormalities associated withcerebral palsy after preterm birth. Eye 2000;14:78–81.

33. Black P. Visual disorders associated with cerebral palsy. Br JOphthalmol 1982;66:46–52.

34. Lossef S. Ocular findings in cerebral palsy. Am J Ophthalmol1962;54:1114–6.

35. Dobbing J, Sands J. Quantitative growth and development ofhuman brain. Arch Dis Child 1973;48:757–67.

36. Mohney BG. Common forms of childhood esotropia. Oph-thalmology 2001;108:805–9.

37. Mohney BG, Huffaker RK. Common forms of childhoodexotropia. Ophthalmology 2003;110:2093–6.

38. Kvarnstrom G, Jakobsson P, Lennerstrand G. Visual screeningof Swedish children: an ophthalmological evaluation. ActaOphthalmol Scand 2001;79:240–4.

39. Christianson RE. The relationship between maternal smokingand the incidence of congenital anomalies. Am J Epidemiol1980;112:684–95.

40. Podgor MJ, Remaley NA, Chew E. Associations betweensiblings for esotropia and exotropia. Arch Ophthalmol 1996;114:739–44.

41. Abrahamsson M, Magnusson G, Sjostrand J. Inheritance ofstrabismus and the gain of using heredity to determine popu-lations at risk of developing strabismus. Acta Ophthalmol

Scand 1999;77:653–7.

1153