Article

What colour are newborns’ eyes? Prevalence of iris colour in the Newborn Eye Screening Test (NEST) study

Cassie A. Ludwig, Natalia F. Callaway, Douglas R. Fredrick, Mark S. Blumenkranz and Darius M. Moshfeghi

Byers Eye Institute, Stanford University School of Medicine, Palo Alto, California, USA

ABSTRACT.

Purpose: This study aims to assess the birth prevalence of iris colour among newborns in a prospective, healthy, full-term newborn cohort.

Methods: The Newborn Eye Screening Test (NEST) study is a prospective cohort study conducted at Lucile Packard Children’s Hospital at Stanford

University School of Medicine. A paediatric vitreoretinal specialist (DMM)

reviewed images sent to the Byers Eye Institute telemedicine reading centre and

recorded eye colour for every infant screened. Variables were graphed to assess

for normality, and frequencies per subject were reported for eye colour, sex,

ethnicity and race.

Results: Among 192 subjects screened in the first year of the NEST study with external images of appropriate quality for visualization of the irides, the birth

prevalence of iris colour was 63.0% brown, 20.8% blue, 5.7% green/hazel, 9.9%

indeterminate and 0.5% partial heterochromia. The study population was

derived from a quaternary care children’s hospital. We report the birth

prevalence of iris colour among full-term newborns in a diverse prospective

cohort.

Conclusion: The study demonstrates a broad range of iris colour prevalence at birth with a predominance of brown iris coloration. Future studies with the

NEST cohort will assess the change in iris colour over time and whether the

frequencies of eye colour change as the child ages.

Key words: eye colour – newborns – paediatric ophthalmology – prevalence – universal newborn

screening

Acta Ophthalmol. 2016: 94: 485–488 ª 2016 Acta Ophthalmologica Scandinavica Foundation. Published by John Wiley & Sons Ltd

doi: 10.1111/aos.13006

Introduction

The Newborn Eye Screening Test (NEST) study is a universal newborn ophthalmic screening initiative utilizing retinal imaging photography to assess newborns for ophthalmic disease and report the birth prevalence of disease. During the first year of screening, parents consistently asked the research

team the same questions after their child was screened: (1) Are my baby’s eyes healthy? and (2) What colour are his or her eyes? The answer to the first question is still being investigated, but the second question we could answer in the examination room.

A review of the literature by these authors revealed that no published study reports the prevalence of iris

colour among a large prospective cohort of newborns assessed shortly after birth. In this study, we aimed to answer the following question: what is the prevalence of iris colour shortly after birth among newborns?

Materials and Methods

Ethics

This study was approved by the Insti- tutional Review Board and Ethics Committee at Stanford University School of Medicine. Informed consent was obtained from a parent of an infant participating in NEST screening, and the study was conducted in a HIPAA-compliant fashion. All research adhered to the tenets of the Declaration of Helsinki.

Study design

The NEST study is a prospective cohort study conducted at Lucile Packard Children’s Hospital (LPCH) at Stanford University School of Medicine. Screening was offered to all infants born at the children’s hospital who did not receive retinopa- thy of prematurity (ROP) screening. This study includes participants screened during the first year of enrolment (25 July 2013 to 25 July 2014). Exclusion criteria were patients who were bilaterally anophthalmic, lacked images or medical charts for review, had infectious conjunctivitis, or were deemed too unstable for examination.

485

Acta Ophthalmologica 2016

Data acquisition

Digital images were taken by a NICU- certified nurse and stored on the RetCam III (Clarity Medical Systems, Pleasanton, CA, USA) computer hard drive as well as input into the NEST telemedicine database management system. The research nurse recorded baseline demographic characteristics for each infant. Study data were col- lected and managed using REDCap electronic data capture tools (Harris et al. 2009). All image acquisition and transmittal was handled in a HIPAA- compliant fashion with strict attention to the confidentiality of personal data.

Image review

A paediatric vitreoretinal specialist (DMM) reviewed images sent to the Byers Eye Institute telemedicine read- ing centre. The reviewing ophthalmol- ogist recorded eye colour and ophthalmic examination findings for every infant screened.

Intra-observer reliability

The intra-observer reliability was assessed using a random subset of 40 consecutive infants whose images were regraded 1 year following the initial grading. We found an overall agree- ment of 87.5% with an unweighted kappa of 0.742 (95% CI: 0.550–0.934) indicating substantial agreement.

Statistical analysis

Variables were graphed to assess for normality, and frequencies per subject were reported for eye colour, sex, ethnicity and race. Groups were com- pared using crude bivariate statistical tests appropriate for the variable of interest with statistical significance set as a two-tailed alpha of <0.05.

Results

We screened 202 subjects during the first year of enrolment with 192 sub- jects (95.0%) with external images of appropriate quality for visualization of the irides (Fig. 1). The overall preva- lence of eye colour is shown in Figure 2 with brown being the most common iris colour (63.0%, 121/192). We found one case of partial heterochromia and no cases of complete heterochromia.

The reviewing vitreoretinal specialist chose standard iris colour images (Fig. 1) to compare images for colour assessment. Any images that did not appear close to these colour pho- tographs were deemed ‘indeterminate’ colour (9.9%, 19/192).

Table 1 compares the frequency of eye colour by sex, race and ethnicity.

There was no significant difference in colour by sex. By ethnicity, signifi- cantly more newborns identified by parents as Not Hispanic or Latino had blue irides compared to those identified as Hispanic or Latino (p <0.0001, Chi-squared test). A com- parison of iris colour by self-identified race was limited because race was

(A) (B)

(D) (E)

(C)

Fig. 1. Iris colourphotographs of a (A) Brown (B)Blue (C) Green/Hazel (D) Indeterminate* and(E) Partially heterochromic eye. Eye colour was determined by one reviewing paediatric vitreoretinal

specialist (DMM). The photographs were taken by RetCam III by a NICU-nurse trained in retinal

image photography as a part of the Newborn Eye Screening Test (NEST). The images displayed in

this figure represent the classic image for the eye colour description. *In some cases, a subject’s eye colour did not easily fit these categories and was deemed indeterminate (panel D).

Fig. 2. Frequency of iris colour among newborns detected after birth via retinal image

photography screening. The majority of newborns had brown eyes at birth, with blue being the

next most common eye colour. One paediatric ophthalmologist reviewed the images and assessed

eye colour for each eye independently. 9.9% of infants had colours that did not clearly match the

preassigned eye colour templates and were deemed ‘Indeterminate’. One case of partial

heterochromia was identified, and no cases of complete heterochromia were observed. Data are

given as the percentage of newborns with count noted circumferentially.

486

Acta Ophthalmologica 2016

missing a significant portion of the data (37.6% missing). All other vari- ables had less than 5% missing values. Crude bivariate methods were used to compare eye colour by race. Newborns whose parents identified them as White/Caucasian more often had blue irides compared to those identified as Asian (p < 0.0001, Fisher’s exact test) or Native Hawaiian or Pacific Islander (p = 0.0032, Fisher’s exact test). In contrast, significantly more newborns identified as Asian (p < 0.0001, Fish- er’s exact test) and Native Hawaiian or Pacific Islander (p = 0.0104, Fisher’s exact test) had brown irides than new- borns identified by parents as White/ Caucasian.

Discussion

We report the birth prevalence of newborn eye colour shortly after birth as determined by retinal image photog- raphy in a universal newborn screening initiative. Newborn iris colour at birth is brown in 63.0% (121/192) of infants, blue in 20.8% (40/192) of infants, green/hazel in 5.7% (11/192) of infants, indeterminate in 9.9% (19/192) of infants and partially heterochromic in 0.5% (1/192) of infants. The study cohort represents a diverse patient population presenting to a quaternary care children’s hospital and enrolled in

the Newborn Eye Screening Test (NEST) study at Stanford University aimed to detect ophthalmic pathology at birth. Not surprisingly, the study also confirmed ethnic and racial pre- dispositions to select eye colours with significantly more brown-eyed Asian and Native Hawaiian/Pacific Islander infants and significantly more blue- eyed Caucasian infants. No statistically significant correlations between sex and eye colour were seen, and no significant relationships between green/hazel eyes and race were revealed in this study; however, the sample size remains too small for these categories at this time.

A newborn’s iris colour holds sig- nificance for both families and clini- cians. Families are curious and eager to learn more about the newest member of their family. Clinicians, on the other hand, monitor iris colour as the child ages because iridial discoloration may be a sign of pathology. Genetic, bio- chemical and infectious lesions such as Down’s syndrome, neurofibromatosis, albinism, pigment dispersion and even herpetic infections can present with iris discoloration (Mackey et al. 2011).

At least 16 different genes influence iris colour (White & Rabago-Smith 2011). Nevertheless, two adjacent genes on chromosome 15 control the major- ity of iris colour variation: the oculo- cutaneous albinism type II gene

(OCA2) and the hect domain and RCC1-like domain-containing protein 2 gene (HERC2) (White & Rabago- Smith 2011). Specifically, polymor- phisms in one intron of the OCA2 gene that is regulated by a promoter region in the intron of HERC2 correlate with blue/non-blue iris colour variation (Duffy et al. 2007). Likewise, blue- brown iris colour variation corre- sponds to a single SNP in the HERC2 intron (Sturm et al. 2008).

At the biochemical level, iris colour appears to vary by the quantity of melanin within the iris, with brown irides simply possessing more melanin than blue irides (Wielgus & Sarna 2005; White & Rabago-Smith 2011). Lighter features provide evolutionary advan- tage to those at higher latitudes by aiding in vitamin D production at low levels of ultraviolet (UV) radiation (Roberts 1977; Sanfilippo et al. 2015). Conversely, darker features provide evolutionary advantage to those at lower latitudes by protecting against high levels of UV radiation (Roberts 1977; Sanfilippo et al. 2015).

Changes in melanin production over time can yield changes in iris colour from infancy and may explain the popular myth that all infants are born with blue eyes. The Louisville Twin Study demonstrated that 10% to 20% of children experienced a change in iris

Table 1. Prevalence of iris colour of newborns in the Newborn Eye Screening Test (NEST) study by sex, ethnicity and race.

Characteristic No.

% (n)

Brown Blue Indeterminate Green/Hazel Partial Heterochromia

Overall 192 63.0 (121/192) 20.8 (40/192) 9.9 (19/192) 5.7 (11/192) 0.5 (1/192)

Sex

Female 89 65.2 (58/89) 19.1 (17/89) 11.2 (10/89) 4.5 (4/89) 0.0 (0/89)

Male 103 61.1 (63/103) 22.3 (23/103) 8.7 (9/103) 6.8 (7/103) 1.0 (1/103)

Race/ethnicity

Hispanic or Latino 82 74.4 (61/82) 7.3 (6/82)* 11.0 (9/82) 7.3 (6/82) 0.0 (0/82)

Not Hispanic or Latino 109 54.1 (59/109) 31.2 (34/109)* 9.2 (10/109) 4.6 (5/109) 0.9 (1/109)

Race***

Did not Answer 73 76.1 (56/73) 5.5 (4/73) 9.6 (7/73) 8.2 (6/73) 0.0 (0/73)

White/Caucasian 64 34.4 (22/64) 54.(35/64)** 7.8 (5/64) 1.6 (1/64) 1.6 (1/64)

Asian 45 80.0 (36/45) 2.2 (1/45)** 13.3 (6/45) 4.4 (2/45) 0.0 (0/45)

Native Hawaiian or Pacific Islander 9 77.8 (7/9) 0.0 (0/9) 0.0 (0/9) 22.2 (2/9) 0.0 (0/9)

Black or African American 1 0.0 (0/1) 0.0 (0/1) 100.0 (1/1) 0.0 (0/1) 0.0 (0/1)

*Significantly more newborns identified by parents as Not Hispanic or Latino had blue irides compared to those identified as Hispanic or Latino

(p = <0.0001, Chi-squared test). **Infants whose parents identified as White/Caucasian were more likely to have blue irides when compared to infants whose parents identified them

as Asian (p < 0.0001, Fisher’s exact test) or Native Hawaiian or Pacific Islander (p = 0.0032, Fisher’s exact test). ***Self-identified race was the only variable missing a significant portion (37.6%). All other variables had less than 5% missing values.

Percentages may not add up to 100 due to rounding. The NEST study is a universal newborn screening initiative at Lucile Packard Children’s

Hospital where universal RetCam III screening is offered to all infants who otherwise did not meet retinopathy of prematurity screening criteria. The

table above reports the 1-year birth prevalence of iris colour. Overall, brown was the most common eye colour in the cohort.

487

Acta Ophthalmologica 2016

colour between 3 months and 6 years of age (Bito et al. 1997). Furthermore, a subpopulation of 10–15% of Cau- casian subjects in the study had changes up until adulthood (Bito et al. 1997). However, the study lacks preva- lence of iris colour at birth as infants were first screened at 3 months. Fur- thermore, the study has restricted gen- eralizability because it was limited to twins. Given the lack of information regarding iris colour at birth, our team developed this study to prospectively evaluate the birth prevalence of eye colour among universally screened infants drawn from a diverse patient population in hopes of alleviating this gap in the literature.

The strength of this study is that it is a large, population-based, prospective study enrolling infants of all races, sexes and socio-economic classes into a universal newborn screening initiative with images documenting the examina- tion findings. One reviewer used stan- dard colour photographs to compare new iris photos and determine eye colour. Thus, the study results are likely generally applicable to the pop- ulation. The low number of Black or African American and American Indian or Alaska Native families lim- ited the generalizability of this study to these groups. Furthermore, eye colour is an inherently subjective assessment and thus prone to the bias of the single image reviewer. The use of standard- ized colour images for borderline cases aimed to standardize the classification process; however, the potential for reviewer bias remains. We do not anticipate there was any differential misclassification of the eye colour, and thus, any bias would likely not affect

the overall results of the study. Of note, it is possible that better scoring of eye colour may have been achieved through use of a brighter flash system.

To these authors’ knowledge, this study is the first, large-scale universal newborn study that reports the preva- lence of eye colour among newborns within the first few days of birth. Future studies of the NEST cohort will assess the change in iris colour over time and whether the frequencies of eye colour change as the child ages.

References

Bito LZ, Matheny A, Cruickshanks KJ, Non-

dahl DM & Carino OB (1997): Eye color

changes past early childhood. The Louisville

Twin Study. Arch Ophthalmol 115: 659–663. Duffy DL, Montgomery GW, Chen W, Zhao

ZZ, James MR, Hayward NK, Martin NG

& Sturm RA (2007): A three-single-nucleo-

tide polymorphism haplotype in intron 1

of OCA2 explains most human eye-color

variation. Am J Hum Genet 80:

241–252. Harris PA, Taylor R, Thielke R, Payne J,

Gonzalez N & Conde JG (2009): Research

electronic data capture (REDCap) - A

metadata-driven methodology and workflow

process for providing translational research

informatics support. J Biomed Inform 42:

377–381. Mackey DA, Wilkinson CH, Kearns LS &

Hewitt AW (2011): Classification of iris

colour: review and refinement of a classifi-

cation schema. Clin Exp Ophthalmol 39:

462–471. Roberts DF (1977): Human pigmentation: its

geographical and racial distribution and

biological significance. J Soc Cosmet Chem

28: 329–342. Sanfilippo PG, Wilkinson CH, Ruddle JB,

Zhu G, Martin NG, Hewitt AW & Mackey

DA (2015): Don’t it make your brown eyes

blue? A comparison of iris colour across

latitude in Australian twins Clin Exp

Optom. 98: 172–176. Sturm RA, Duffy DL, Zhao ZZ, Leite FP,

Stark MS, Hayward NK, Martin NG &

Montgomery GW (2008): A single SNP in

an evolutionary conserved region within

intron 86 of the HERC2 gene determines

human blue-brown eye color. Am J Hum

Genet 82: 424–431. White D & Rabago-Smith M (2011): Geno-

type-phenotype associations and human eye

color. J Hum Genet 56: 5–7. Wielgus AR & Sarna T (2005): Melanin in

human irides of different color and age of

donors. Pigment Cell Res 18: 454–464.

Received on August 17th, 2015.

Accepted on January 7th, 2016.

Correspondence:

Darius M. Moshfeghi, MD

Clinical Professor, Ophthalmology

Stanford Byers Eye Institute

2452 Watson Ct.

Palo Alto, California 94303

Tel: +1 (650) 736 2050

Fax: +1 (650) 721 2884

Email: [email protected]

This research was made possible by a grant

from the Giannini Foundation. Cassie Ann

Ludwig and Natalia Fijalkowski Callaway

were funded by the NIH TL1 trainee grant

TR 001084 offered by the Stanford centre for

Clinical and Translational Research and Edu-

cation. We thank Andrew Martin at the

Stanford centre for Clinical Informatics who

provided database design and management.

The Research Electronic Data Capture (RED-

Cap) database tool hosted by Stanford

University is maintained by the Stanford

centre for Clinical Informatics grant sup-

port (Stanford CTSA award number UL1

RR025744 from NIH/NCRR).

488

Acta Ophthalmologica 2016