Relationship of Ocular Biometry and Retinal Vascular Caliber in Preschoolers

Abstract

Purpose.: To systematically examine the association between ocular biometry and retinal vascular caliber in Singapore Chinese preschoolers aged 48 to 72 months.

Methods.: A total of 469 Singapore Chinese children aged 48 to 72 months were recruited through the Strabismus, Amblyopia and Refractive Error Study in Singaporean Chinese Preschoolers (STARS) from 2006 to 2008. According to standardized protocols, cycloplegic autorefraction, ocular biometry measurements, and retinal photography were performed. Retinal vascular caliber was measured quantitatively and was summarized as the central retinal arteriolar equivalent (CRAE) and central retinal venular equivalent (CRVE), respectively. Ocular magnification was corrected by using the Bengtsson formula.

Results.: The mean retinal arteriolar and venular calibers were 156.08 μm and 219.55 μm in boys, and 161.96 μm and 224.25 μm in girls, respectively. In multiple linear regression adjusted for age, sex, father's education, parental myopia history, mean arterial blood pressure, body mass index, and spherical equivalent, each 1.0 mm increase in axial length was associated with a 3.52 μm decrease in retinal arteriolar caliber (P = 0.023) and a 5.55 μm decrease in retinal venular caliber (P = 0.008). Each 1.00 mm increase in corneal curvature was associated with a 13.79 μm decrease (P = 0.004) in retinal venular caliber.

Conclusions.: In very young children aged 48 to 72 months, narrower retinal arteriolar and venular caliber was associated with elongated axial length. Narrower retinal venular caliber was associated with larger corneal curvature. This suggests that the major structural correlate of myopia might have an effect on retinal microvasculature from early childhood.

The socioeconomic cost associated with myopia is significant because it affects an individual for a lifetime. ¹ This burden is particularly large among Asian young Chinese (16–25 years old) with high myopic prevalence, which was reported as 82.2% in Singapore, ¹ 73.1% in Hong Kong, ² and 81.0% in Taiwan. ³ It has been widely studied that pathologic myopia is associated with potentially blinding complications such as retinal detachment, choroidal neovascularization, macular holes, ^{4 –6} and incident glaucoma and cataract. ^7,8 Using Doppler ultrasonography, reduction of retinal blood flow was indirectly reported in highly myopic eyes (≥ −6.0 diopters [D]), ⁹ even more degenerative myopic eyes (≥ −8.0 D), ¹⁰ myopic eyes with open-angle glaucoma, ¹¹ and myopic eyes with choroidal neovascularization. ¹²  

Through retinal photography, we can noninvasively examine and investigate the retinal vascular changes quantitatively using advanced imaging technology which reflects structural and pathologic features of human microcirculation. Several studies have shown an association with similar direction of effect between both narrowing of retinal arteriolar and venular caliber and refractive error (either spherical equivalent or axial length) in children and adults. ^{13 –17} However, among three reports in which ocular magnification was taken into account, ^14,16,17 only Lim et al. ¹⁴ reported the persistent association. 

Studying microvascular changes by examining retinal vascular caliber on myopic eyes may lead to better understanding of the pathophysiology of myopia. The aim of this study was to determine whether there is a consistent relationship between retinal vascular caliber and refraction among 469 Singapore Chinese preschoolers aged 48 to 72 months, which is a younger pediatric population than studied in the Singapore Cohort Study of the Risk Factors for Myopia (SCORM). ¹⁷  

Methods

Study Population

The Strabismus, Amblyopia and Refractive Error Study in Singaporean Chinese Preschoolers (STARS) is a population-based, cross-sectional study conducted between May 2006 and October 2008. In brief, 3009 Chinese children aged 6 to 72 months, living in government apartments (>85% of the Singaporean population) in the Western and Southwestern regions of Singapore, were invited to participate. The recruitment has been described in previously published studies. ^18,19 Of the 1184 children aged 48 to 72 months who were recruited, retinal photography was successfully obtained in 513 participants with 91.4% gradable retinal images (n = 469). 

The study was approved by Institutional Review Boards of the Singapore Eye Research Institute and the Singapore National Health care Group and conducted according to the tenets of the Declaration of Helsinki. Informed written consent from parents and verbal assent from the child were obtained before any testing. 

Retinal Photography and Measurement of Retinal Vascular Caliber

After pupillary dilation using three sets of 1% cyclopentolate and 2.5% phenylephrine drops given five minutes apart, digital retinal photographs were taken using a 45° digital retinal camera (Model CR6-NM45, Canon Inc., Tokyo, Japan). Two images were obtained, one centered on the optic disc and the other centered on the macula, within which the optic disc-centered image was used for grading and analysis. Methods used to measure retinal vascular caliber from retinal photographs followed a standardized protocol as described in previous reports in adults and children. ^{7,15 –17} A computer imaging analysis program (IVAN; University of Wisconsin, Madison, WI) was used to measure the caliber of all retinal arterioles and venules located in zone one half- to one-disc diameter from the optic disc margin in the retinal photograph (zone B). Using the revised Knudtson-Parr-Hubbard formula to compute retinal vascular caliber, the largest six arterioles and venules were used in calculating average vascular caliber, and estimates of the average diameters of arterioles and venules were summarized as the central retinal arteriolar equivalent (CRAE) and central retinal venular equivalent (CRVE). 

A single grader, masked to blood pressure measurements and participant characteristics, performed all retinal vascular caliber measurements for this cohort. Intragrader reliability was assessed in 50 randomly selected retinal photographs, and the intra-class correlation coefficient was 0.95 for CRAE and 0.96 for CRVE. 

To correct for ocular magnification on retinal vascular caliber measurements which was caused by telecentric camera and ocular refractive media, we used a correction factor (1–0.0017 × spherical equivalent [SE]) described by Bengtsson and Krakau ²⁰ for our fundus camera (Canon). 

Autorefraction and Ocular Biometric Measurement

Cycloplegic autorefraction was performed using an autorefraction camera (Canon Autorefraction RK-F1; Canon Inc., Tokyo, Japan), where the average from five consecutive readings of spherical and cylindrical refractive error was obtained. The autorefractor readings were considered acceptable if the difference between the lowest and highest reading was 0.25 D or less. Spherical equivalent (SE) was calculated as the sphere plus half of the cylinder (SE = sphere + [cylinder/2]). 

Axial length, corneal curvature and anterior chamber depth of the right eye were obtained using a noncontact partial coherence interferometer (IOLMaster, Carl Zeiss, Oberkochen, Germany). A total of five consecutive readings with a signal-to-noise ratio of > 2.0 and with a difference between the lowest and highest reading being 0.05 mm or less were used to determine axial length. Then a total of 5 consecutive readings of corneal curvature and anterior chamber depth were obtained subsequently. The mean of five readings of each ocular biometric parameter was used in the analysis. 

Other Measurements

Blood pressure was measured at clinics following a standard protocol, ¹⁸ by using an automatic sphygmomanometer (Omron HEM 705 LP, Omron Health Care Inc.; Bannockburn, IL) with an appropriate pediatric cuff size after 5 minutes of rest. Two separate measurements were taken and their average was calculated. If the difference between the first two readings was greater than 10 mm Hg in systolic blood pressure (SBP) and/or 5 mm Hg in diastolic blood pressure (DBP), a third reading was be taken. The average of the two closest readings was used for analysis. Mean arterial blood pressure (MABP) was equal to DBP plus one third of the difference between the systolic and diastolic pressures (MABP = DBP + ⅓[SBP − DBP]). 

Height and weight were both measured in the standing position according to a standard protocol by using a height and weight measuring scale (Seca model 220, Hamburg, Germany). ¹⁹ Height was recorded to the nearest 1.0 mm while weight was recorded to the nearest 0.1 kg. Body mass index (BMI) was calculated as weight (kg) divided by the height squared (m²). 

Interview

Parental sociodemographic information was obtained through either English or Chinese questionnaires. The father's education was classified into four categories as follows: (1) below or equal to secondary school; (2) O/N level (O level, 4 years secondary school; N level, 5 years of secondary school); (3) A levels/diploma; and (4) University education or higher. Parental myopia history was collected through clinic interview. It was classified into two groups as follows: none of the parents had myopic condition, or at least one parent had myopic condition. Subjects' birth parameters information such as birth weight, birth length, head circumference, and gestation week was obtained from the subjects' health booklets which were brought along by their parents/guardians to the clinic. 

Statistical Analysis

Multiple linear regression was constructed to estimate the relationship between retinal vascular caliber and three ocular parameters (axial length, corneal curvature, and anterior chamber depth) by using 2 multivariate models. Model 1 was controlled for age, sex, and SE. Model 2 was controlled for all variables in model 1, and then was further adjusted for father's education, parental myopia history, MABP, BMI, and birth weight. 

Ocular parameters were categorized into quartiles and analyzed as continuous variables. Test of trend was determined by treating quartiles of ocular parameters in their association with uncorrected and corrected retinal vascular caliber. 

Potential modifiers were examined in stratified analyses. All probabilities quoted are two-sided, and a significant P value was defined as <0.05. All statistical analyses were performed using statistical software (PASW 18.0; SPSS Inc.; Chicago, IL). 

Results

All the variables were approximately normally distributed. The mean age of our study population was 60.88 ± 7.52 months in boys and 60.48 ± 7.19 months in girls. Table 1 shows the demographics of the study population stratified by sex. Among 469 Singapore Chinese children aged 48 to 72 months, except ocular biometric parameters and retinal vascular caliber, there is no significant difference between boys and girls in age, sex, father's education, SE, MABP, BMI, and gestation week. The boys in our study had relatively greater axial length (22.62 ± 0.75 mm vs. 22.08 ± 0.71 mm; P < 0.001), greater cornea curvature (7.76 ± 0.24 mm vs. 7.65 ± 0.21 mm; P < 0.001), and greater anterior chamber depth (3.41 ± 0.25 mm vs. 3.32 ± 0.22 mm; P < 0.001) than girls, while boys had narrower CRAE (156.08 ± 15.27 μm vs. 161.96 ± 15.65 μm; P < 0.001) and narrower CRVE (219.55 ± 19.13 μm vs. 224.25 ± 21.84 μm; P = 0.013) than girls. 

Table 1.

 

Characteristics of Included Children from STARS

Characteristics	Total (n)	Sex		P Value*
		Boys (Mean ± SD)	Girls (Mean ± SD)
Age, mo	469	60.88 (7.52)	60.48 (7.19)	0.575
Sex, %	469	239 (50.96)	230 (49.04)
Father's education, %†				0.931†
≤Secondary school	127	71 (29.96)	67 (29.13)
O/N level	100	42 (17.72)	45 (19.57)
A level/diploma	110	58 (24.47)	52 (22.61)
University education	132	66 (27.85)	66 (28.70)
Axial length, mm	469	22.62 (0.75)	22.08 (0.71)	<0.001
Spherical equivalent (SE), D	469	0.62 (1.05)	0.79 (1.18)	0.097
Corneal curvature, mm	469	7.76 (0.24)	7.65 (0.21)	<0.001
Anterior chamber depth, mm	469	3.41 (0.25)	3.32 (0.22)	<0.001
Mean arterial blood pressure, mm Hg	469	75.00 (7.83)	74.95 (8.09)	0.948
Body mass index	469	15.79 (2.54)	15.40 (1.73)	0.056
Gestation, wk	469	38.47 (1.45)	38.54 (1.79)	0.672
Retinal vascular caliber, μm
CRAE	469	156.08 (15.27)	161.96 (15.65)	<0.001
CRVE	469	219.55 (19.13)	224.25 (21.84)	0.013

 SD, standard deviation.

*  Student's t-test.

†  χ² test.

Table 2 describes the association between retinal vascular caliber and ocular biometric parameters. In multiple linear regression models adjusting for age, sex, right eye spherical equivalent, father's education, parental myopia history, MABP, BMI, and birth weight, each 1.0 mm increase in axial length was associated with a 3.52 μm reduction (P = 0.023) in retinal arteriolar caliber and a 5.55 μm reduction (P = 0.008) in retinal venular caliber. Each 1.0 mm increase in corneal curvature was only associated with 13.79 μm reduction (P = 0.004) in retinal venular caliber while not with retinal arteriolar caliber (P = 0.093). There was no association found between retinal vascular caliber and anterior chamber depth in both linear regression models. 

Table 2.

 

Multiple Linear Regression of Association between Corrected Retinal Vascular Caliber and Refraction and Ocular Biometric Parameters

	N	Retinal Arteriolar Caliber, β ± SE (μm)		Retinal Venular Caliber, β ± SE (μm)
		Model 1*	Model 2†	Model 1*	Model 2†
Axial length, mm	442
Per unit increase (1.0 mm)		−2.62 ± 1.23	−3.52 ± 1.54	−4.17 ± 1.64	−5.55 ± 2.09
P value		0.034	0.023	0.011	0.008
Adjusted R 2		0.053	0.037	0.026	0.032
Corneal curvature, mm	417
Per unit increase (1.0 mm)		−1.72 ± 3.15	−5.85 ± 3.47	−9.81 ± 4.20	−13.79 ± 4.69
P value		0.584	0.093	0.020	0.004
Adjusted R 2		0.040	0.031	0.022	0.032
Anterior chamber depth, mm	416
Per unit increase (1.0 mm)		3.55 ± 3.60	4.76 ± 3.72	1.15 ± 4.18	3.55 ± 3.60
P value		0.252	0.202	0.783	0.898
Adjusted R 2		0.040	0.029	0.008	0.005

 SE, standard error.

*  Model 1 was adjusted for age, sex, and right eye spherical equivalent.

†  Model 2 was adjusted for all variables in model 1, and then further adjusted for father's education, parental myopia history, mean arterial blood pressure, body mass index, and birth weight.

Figure 1 shows the association of quartiles of axial length and quartiles of corneal curvature with retinal venular caliber after controlling for age, sex, SE, father's education, parental myopia history, MABP, BMI, and birth weight. Error plots demonstrated that highest quartile of axial length was significantly associated with narrower retinal venular caliber than the lowest quartile of axial length (P trend = 0.023; adjusted R ² = 0.168) (Fig. 1A). Highest quartile of corneal curvature was associated with narrower retinal venular caliber than the lowest quartile of corneal curvature (P trend = 0.022; adjusted R ² = 0.153) (Fig. 1B). 

Figure 1.

 

(A) Relationship between quartiles of axial length and retinal venular caliber, after controlling for age, sex, SE, father's education, parental myopia history, MABP, BMI, and birth weight. (B) Relationship between quartiles of corneal curvature and retinal venular caliber, after controlling for age, sex, SE, father's education, parental myopia history, MABP, BMI, and birth weight. CI, confidence interval.

Discussion

Among all ocular biometric parameters in our study, longer axial length was associated with narrowing in both retinal arteriolar and venular caliber. Larger corneal curvature was only associated with narrowing in retinal venular caliber. Anterior chamber depth was not associated with either retinal arteriolar or venular caliber. 

There have been only a few studies that have reported the relationship either between axial length and retinal vascular caliber or between refractive error and retinal vascular caliber, mainly in adults. ^{13 –16} Patton et al. ¹³ reported the negative association between axial length and retinal venular caliber by using Pearson's coefficient correlation (R = −0.28; P = 0.04). In the Singapore Malay Eye Study (SiMES), conducted in adults aged 40 to 80 years, Lim et al. ¹⁴ reported that per 1.0 mm increase in axial length, there was an associated 3.25 μm and a 3.20 μm narrowing in retinal arteriolar and venular caliber, respectively. Similar to axial length, SE was suggested to be positively related to wider retinal arteriolar and venular caliber in the Beaver Dam Eye Study (BDES) ¹⁵ and SiMES. ¹⁴ BDES found that each 1.0 D decrease in SE was significantly associated with a 2.8 μm decrease and 3.3 μm decrease in retinal arteriolar and venular caliber, respectively. ¹⁵ SiMES found a much smaller decrease in both retinal arteriolar caliber and venular caliber as 0.46 μm and 0.42 μm, respectively. ¹⁴ If refraction category was taken into account with the trend changed from hyperopia to myopia, the Blue Mountain Study reported an decreasing trend both in retinal arteriolar caliber (204.7 μm vs. 162.5 μm; P < 0.001) and retinal venular caliber (238.9 μm vs. 195.9 μm; P < 0.001), ¹⁶ while the SiMES study only reported an decreasing trend in retinal venular caliber from hyperopia to myopia (204.35 μm vs. 202.08 μm; P = 0.02). ¹⁴ For children, there has been only one study on axial length with retinal vascular caliber among children aged 7 to 9 years, and the findings were similar to the adult study mentioned above. ¹⁴ The SCORM reported per SD (1.02 mm) increase in axial length to be statistically associated with a 3.18 μm and a 4.62 μm decrease in retinal arteriolar and venular caliber, respectively. ¹⁷  

It has been widely suggested to model both retinal arteriolar and venular caliber simultaneously in research relating retinal vascular caliber to systemic outcomes such as hypertension and diabetes. ^{21 –24} Up until now it has not been applied to ocular outcome. ^13,14,16,17 However, the application of retinal fellow vessel model in ocular outcomes should be explored more for practical statistical analysis. To be able to relate our findings with those reported in previous studies on similar topics, ^13,14,16,17 we prefer to take the adjustment without retinal fellow vessel as our ultimate analysis. 

Increased axial elongation in myopic individuals may lead to mechanical stretching and thinning of the choroid and retinal pigment epithelium with concomitant vascular and degenerative changes. ²¹ Regarding the possible concomitant vascular changes on retinal microcirculation, especially speculated less ocular blood flow, studies have shown some direct and indirect evidence. ^{9,10,12,14,22,23} Using Doppler ultrasonography or Doppler velocimetry, reduction of retinal blood flow velocity or reduced retinal blood flow was reported in highly myopic eyes (>−8.0 D), ^9,10 myopic eyes with open-angle glaucoma, ¹¹ and myopic eyes with choroidal neovascularization. ¹² Using retinal photography, retinal vascular caliber and retinal vessel geometry were all decreased in myopic eyes, which implied a decreased retinal microcirculation in myopic subjects. ¹⁴ By using three-dimensional magnetic resonance imaging (MRI), axial globe enlargement to achieve a prolate shape in young children's myopic eyes was seen, which implied that a stretched eye ball and increased axial elongation in myopic individuals probably led to mechanical stretching and thinning of the choroid and retinal pigment epithelium with concomitant vascular and degenerative changes. ^23,24 If a given eye begins development with a set complement of retinal vasculature, it is likely that a pathologic increase in ocular dimensions could cause stretching and elongation of the retinal vessels, leading to reduce the retinal vessel width. 

The clinical implications of our study might lie in two parts. Firstly, Saw et al. ²⁴ suggested that patients with myopia, especially high myopia, may have higher risks of cataract, glaucoma, and chorioretinal abnormalities such as retinal detachment and optic disc abnormalities. Therefore, the relationship between axial length and retinal vascular caliber can shed light on the underlying pathophysiological mechanisms on how myopic subjects progress to develop pathologic complications, like retinal detachment. Secondly, in a recent study by Lim et al., ²⁵ myopic refraction and longer axial length were associated with a lower risk of diabetic retinopathy, particularly vision-threatening retinopathy. It was hypothesized that chorioretinal thinning among highly myopic individuals may be protective both by reducing the metabolic demands of the retina and by facilitating diffusion of oxygen through the retina. Therefore, the reduced metabolic demands of the retina might be directly caused by the reduced blood flow established in this study and the previous studies. 

Longanesi et al. ²⁶ and Shimmyo ²⁷ found a highly significant correlation between corneal curvature and central corneal thickness. Thinner central corneal thickness (CCT) was reported to be associated with retinal arteriolar narrowing in the SCORM study. ²⁸ Furthermore, intraocular pressure (IOP) measurements by applanation tonometry are affected by central corneal thickness, corneal curvature, and axial length. ²⁹ Interestingly, we found a very strong correlation between retinal venular caliber and corneal curvature across both models. Based on the previous study findings, a few possible mechanisms might lead larger corneal curvature to retinal vascular caliber changes. Firstly, the corneal curvature might share the same viscoelastic properties of lamina cribrosa as CCT. ³⁰ Larger corneal curvature might share the same mechanism as thinner CCT, which was associated with a thinner lamina cribrosa and the reduced mechanical support for blood vessel passing through would in turn lead the compression of retinal vessel walls. ³⁰ Secondly, larger corneal curvature might imply a latent higher IOP, which had a direct impact on retinal vascular narrowing. ^31,32 However, our study only found the association between larger corneal curvature and retinal venular narrowing, which might not be fully explained by these two hypothesized theories mentioned above. Due to lack of literature support, it is unclear how corneal curvature was related to retinal venular caliber yet not to retinal arteriolar caliber. Whether this association is biologically proven or just statistically proven still needs more studies to identify. 

The strengths of our study design include its large sample size on very young and homogenous Chinese preschoolers, standardized protocol for in vivo ocular biometric and retinal vessel measurements, and detailed information on a range of potential confounders. The possible benefit to observe association between retinal vessel calibers and ocular parameters in such a young cohort is to provide some normative data in a young Chinese population, which can be the reference for evaluating pathologic myopia progression later in adult life. However, potential limitation still exists due to the nature of our study. As a cross-sectional study, it could not provide temporal information about the relationships. Furthermore, we were not able to grade for retinal geometric parameters such as fractals, vessel tortuosity, and branching angles as these assessments require gradable vessels outside of the standard 2 disc diameter areas that we used to measure retinal vessel diameter. It would be of great interest to examine these associations if further research can be done to provide more parameters for assessing a complete retinal vascular geometry. 

In brief, our study is the first study to investigate the effects on retinal vascular caliber by ocular biometric parameters in very young children aged 48 to 72 months. These results further support the hypothesis of a biological link between ocular axial dimension and retinal vasculature, and might also improve the understanding of the pathophysiology of hemodynamic changes on myopic eyes among children. 

Footnotes

 Supported by the National Health and Medical Research Council, Singapore (NMRC/1009/2005 and NMRC/1112/2007).

Footnotes

 Disclosure: L.-J. Li, None; C.Y.-L. Cheung, None; G. Gazzard, None; L. Chang, None; P. Mitchell, None; T.-Y. Wong, None; S.-M. Saw, None