Impact of Family History of High Myopia on Level and Onset of Myopia

Abstract

purpose. To investigate the impact of a positive family history of high myopia on the level and onset of myopia and its ocular components.

methods. A cross-sectional study was conducted. The participants (aged 17 to 45 years) were categorized into four groups: normal, mild, moderate, and high myopia. The age of first glasses for myopia was used as the onset of myopia. The impact of the family history on the level and the onset of myopia was quantified. Parental effect on corneal curvature (CC), anterior chamber depth (ACD), and axial length (AXL) was analyzed.

results. The study included 185 normal subjects, 170 mild, 140 moderate, and 392 high myopes. Family history was strongly associated with the probands’ status (P < 6 × 10^-12). When there was ≥1 highly myopic parent, the odds ratios (ORs) of developing mild or moderate myopia were between 2.5 and 3.7 (95% CI: 1.1–6.5) and the ORs of having high myopia were > 5.5 (95% CI: 3.2–12.6). A strong association (P = 2 × 10^-6) between parental myopic state and the AXL in the subjects was also found, but there was no statistical relationship for ACD or CC. There was an association between high myopia in parents and the onset of myopia in children. Siblings had a weaker association with the level of myopia and had no effect on the onset of myopia.

conclusions. This study found strong familial effects on the level and onset of myopia even after adjusting for environmental factors. The parental effect on ocular components in their offspring was primarily on AXL.

Myopia is a very common eye condition. The prevalence of myopia in the United States is approximately 25% of the population, ¹ and it is much higher in Asian countries. ^{2 3} Myopia is a significant public health problem, because it is associated with the increased risk of several ocular diseases such as glaucoma, posterior subcapsular cataract, retinal detachment, myopic retinal degeneration, visual impairment, and blindness. ^{4 5 6 7 8} Substantial resources are required for optical correction of myopia such as spectacles, contact lenses, orthokeratology, photorefractive keratectomy, and laser in situ keratomileusis (LASIK). However, these corrections do not prevent the ocular complications mentioned above. Furthermore, complications arising from the use of contact lenses, ⁹ orthokeratology ¹⁰ and surgical procedures ¹¹ also impose additional risk to myopes. Myopia is a complex trait that has multiple risk factors including genes, environment, and gene–environment interaction. While studies have found several environmental risk factors, twin studies indicate a strong genetic influence on myopia. ^{12 13 14 15 16 17} Several studies also show that family history is a risk factor for having myopia. ^{18 19 20 21 22 23} However, the magnitude of family risk on different degrees of myopia has not been well quantified. 

In Taiwan, a national survey revealed that the subjects between the ages of 16 to 18 years have a rate of myopia of 84%. ² Using the same definition of high myopia (< −6 D), the prevalence of high myopia is 18% among young Taiwanese men and 24% among young Taiwanese women, ^{2 24} both of which are even higher than 13.1% reported among young men in Singapore. ³ Furthermore, studies show the prevalence of myopia increased from 76% to 81% in the young Taiwanese population (age of 15 years) according to two national surveys conducted in 1995 and 2000. ^{2 24} It is generally believed that a disease caused mainly by genetic factors tends to have an earlier onset, more affected family members, and more severe clinical presentations compared with the same disease caused mainly by environmental factors. ^{25 26 27} Therefore, the present study focused on the impact of highly myopic parents and siblings on the level and onset of myopia. The first goal of this study was to quantify the parental and sibling effect on the level of myopia and three myopic components: axial length (AXL), anterior chamber depth (ACD), and corneal curvature (CC). The second goal was to examine the impact of family history on the onset of myopia. In addition to family history, relevant environmental factors (TV, games, computer work, education level, and outdoor activities) were also assessed in subset data, where the environmental data were available. 

Methods

This study included data from 887 participants between ages of 17 and 45 years. The participants were recruited from Kaohsiung, Taiwan. All the participants were of Chinese descent. The participants came from four sources: 1) college or graduate students who voluntarily participated via an advertisement posted at Sun Yat-Sen University; 2) senior high school students who received vision surveys; 3) young men in the military conscripts; and 4) hospital personnel. The study was approved by the Institutional Review Boards at the Kaohsiung Municipal United Hospital and Columbia University, and all research adhered to the tenets of the Declaration of Helsinki. The participants were informed of all risks and benefits. All the participants signed an informed consent. Every participant received an examination including slitlamp, direct funduscopy, autorefractor examination, and manifest correction without cycloplegia. The measurement of the refractive error was performed using an autorefractometer (KR-8100; Topcon, Tokyo, Japan) for all eyes. The ranges of measurement by Topcon are between −25 and +22 D for spheres, −8 and +8 D for cylinders, and 33.75–67.50 D for corneal refractive power. The minimal required pupillary diameter was only 2.0 mm. The measurement of AXL was taken using an ultrasonic biometer A-scan (model 820; Humphrey Inc., San Leandro, CA). Refractive measurements were checked three times, and the final value presented was the mean of these three measurements. 

The participants were categorized into four groups according to their refractive error (the data from the higher myopic eye was used). Normal: 1.0 to −1.0 D, mild myopia: −1.25 to −3.5 D, moderate myopia: −3.75 to −4.75 D, and high myopia ≤ −5 D. Each participant was asked to complete a questionnaire, where family history, the age of first glasses for myopia, and environmental factors were sought. The questionnaire asked whether the parents and siblings use nearsighted eye glasses, and if so, whether their myopia is ≤ −5 D. If a participant did not have any siblings, he/she was not included in the analysis for sibling risk. Similarly, if family data were missing or answered “I do not know”, the participants were not included in the analysis of family risk. The reliability of family history on the questionnaire was validated in a subset of 100 participants, who were randomly selected regardless of their myopic status. We directly contacted all parents and 56 siblings by asking for their actual refraction. If siblings were not reachable, we asked the parents the same questions as we asked the participants, which was treated as the cross validation. Our validation test showed that two participants who reported to have highly myopic parents actually had only moderate myopic parents, and three participants had falsely reported highly myopic siblings. Therefore, the false report rate from the study participants was 2% for their parents and 3% for their siblings. The five false reports were from four participants: one subject was normal, two subjects were mildly myopic, and one subject was in the highly myopic category. To further validate the self-reported refraction from the participants’ family members, 35 family members were asked to come to our clinic for vision examination. All 35 members used nearsighted eye glasses. They were asked whether their myopia is above or below −5.0 D before examination. Twenty-four family members reported their myopia ≤ −5.0 D and 11 family members reported > −5.0 D. Only one family member who reported > −5.0 D had actual spherical refraction of −5.5 D in the right eye and −4.75 D in the left eye. Therefore, we considered that the family history obtained from the questionnaire was reliable. The questionnaire also included environmental risk factors including average hours/week of watching television, playing video/computer games, working on the computer, participating in outdoor activities, and the education level attained (college or higher versus less than college level). For the first four factors, each participant was asked to choose one of five scores: score 1 for ≤10 hours/week, score 2 for 11–20 hours/week, score 3 for 21–30 hours/week, score 4 for 31–40 hours/week, and score 5 for ≥40 hours/week. 

The χ² test was first used to test for the distribution of probands in accordance with parental and sibling myopic status in the contingency tables. The odds ratio (OR) was calculated to quantify the impact of family history on the myopic status of the probands. AXL, ACD, and horizontal and vertical CC were analyzed by testing for the association between the quartile of each ocular component and parental myopic state. The higher value between both eyes was used for the ocular component analysis. Normal controls and high myopes were further analyzed by logistic regression models where parental myopic state, environmental factors, and gender were included. 

We speculated that family history might influence the level of myopia as well as the onset of myopia. The age of first glasses for myopia was used as a surrogate of the onset age. Myopes were divided into the early or late onset group according to the mean ages of the onset in each myopic category in the data. The mean age of the first glasses was 11 years for high, 13 years for moderate, and 15 years for mild myopes, respectively. The median ages for the three categories of myopia were the same as the mean ages in our data. Analysis of the association between highly myopic parents/siblings and proband’s onset age was performed. To be sure that the cutoff points would not influence the results, we also tested for ages of 11 ± 1, 13 ± 1, and 15 ± 1 years as the cutoff points. A two-tailed P value < 0.05 was considered statistically significant. 

Results

Parental and Sibling Effect on Myopic Status

One hundred eighty five subjects were classified as normal controls. The number of mild, moderate, and high myopia was 170, 140, and 392, respectively. The plots of age versus current mean refractive error, and myopia onset age versus current mean refractive error are shown in Figure 1 . The four age groups shown in Figure 1a represent the four sources of our participants. We first calculated the effect from parents and siblings and found an extremely strong association between the proband status and their parents and siblings. The overall relationship between probands and their parents and siblings yielded a P value of 5.94 × 10⁻¹² and 1.67 × 10⁻¹², respectively (Table 1) . We further dissected the strong relationship using a series of 2 × 2 contingency tables, where probands of no highly myopic parents or siblings served as the reference groups, and probands with a highly myopic family history served as the comparison groups. The OR (95% CI) was 3.09 (1.69–5.67), 3.47 (1.86–6.50), and 5.47 (3.23–9.29) for mild, moderate, and high myopia, respectively, when there was one highly myopic parent. The OR (95% CI) was 2.53 (1.07–6.02), 3.09 (1.27–7.51), and 6.03 (2.88–12.63) for mild, moderate, and high myopia, respectively, when there were two highly myopic parents. Similarly, having one highly myopic sibling resulted in the OR (95% CI) of 1.73 (0.99–3.01), 2.60 (1.45–4.67), and 5.16 (3.19–8.35) for mild, moderate, and high myopia, respectively. When the probands had ≥2 highly myopic siblings, the OR (95% CI) for the mild, moderate, and high myopia was 0.78 (0.25–2.42), 2.51 (0.90–6.95), and 2.27 (0.93–5.56), respectively. The nonsignificant results for probands with ≥2 highly myopic siblings were perhaps due to the small number of probands with ≥2 myopic siblings. 

Multivariate Analysis

The mean score of environmental factors in each myopic category is presented in Table 2 . Education level (P = 0.0001) and TV watching (P < 0.0001) were significantly different among the four myopic categories, but video/computer games, computer, or outdoor activities did not reach the significance level of 0.05. 

A multivariate logistic analysis to evaluate parental effect, five environmental factors, and gender on the proband’s myopic status (Table 3) included only normal and high myopes. The three variables yielding statistical significance were highly myopic parent (OR = 6.36, P = 1.12 × 10⁻⁹ for one myopic parent; OR = 7.70, P = 6.14 × 10⁻⁶ for two myopic parents), college or higher education level (OR = 3.26, P = 4.42 × 10⁻⁷), and women (OR = 2.21, P = 0.02 compared with men). A slight suggestion of dos–response parental effect could be seen in this multivariate logistic regression analysis as well as the above series contingency analyses. However, the 95% CIs between one and two highly myopic parents had a significant overlap. Therefore, caution should be taken in the interpretation of the dose–response effect. 

Parental Effect on Axial Length, Anterior Chamber Depth, and Corneal Curvature

The distributions of AXL, ACD, and CC are shown in Figure 2 . The AXL data were equally divided into four categories (i.e., each category had an equal number of probands) for analysis, regardless of the availability of family history. The cutoff points were 26.71, 25.78, and 24.71 mm (Table 4) . The distribution of the quartile of AXL according to parental myopic state is shown in Table 4 . Parental myopic state was strongly associated with children’s AXL (P = 1.18 × 10⁻⁶). Similarly, the quartiles of ACD and CC data were defined in the same way. The horizontal and vertical CCs were analyzed independently. The cutoff points of ACD were 3.89, 3.73, and 3.54 mm (Table 4) . The cutoff points were 43.75, 42.50, and 41.75 diopter (D) for the horizontal CC, and 45.0, 43.75, and 42.75 D for the vertical CC. There was no statistical association between parental myopic state and participants’ ACD (P = 0.15), horizontal (P = 0.19) or vertical (P = 0.11) CC (Table 4) . 

Parental and Sibling Effect on the Age of Onset of Myopia

There were significant associations (all P values ≤ 0.006) between highly myopic parents and the onset of high, moderate, and mild myopia in their offspring (Table 5) . Among all highly myopic probands, those with highly myopic parents tended to have an earlier onset of myopia with an OR of 2.61 (P = 3.1 × 10⁻⁵). Parental myopia was still strongly significant when ages of 10 (OR = 2.58, P = 7.98 × 10^–5) and 12 (OR = 2.03, P = 0.004) were used as the cutoff points to define “early” versus “late” onset. For moderate myopia, the parental myopic state also had significant impact on the onset of myopia in their offspring with an OR = 4.19, P = 0.001. The significant results still held when the cutoff points were 12 years (OR = 4.78, P = 0.0001) and 14 years (OR = 4.54, P = 0.002). For the mild myopes, although there was a significant association (P = 0.006) for the cutoff point of 15 years, the results were less consistent using different cutoff points (P = 0.09 and 0.03 for 14 and 16 years, respectively). There was no association between number of highly myopic sibs and the onset of myopia in any category: P = 0.79 for high myopia, P = 0.30 for moderate myopia, and P = 0.1 for mild myopia. When different ages were used as cutoff points to evaluate sibling effect, the results remained not significant for any of three myopic categories. The unequal numbers of subjects in the early and late groups (Table 5) were due to more missing family information in the late group. 

Discussion

This study showed three major points: 1) Both parental and sibling myopic states were associated with the risk of having mild, moderate, and high myopia in our participants; 2) Highly myopic parents could accelerate the onset of myopia in their myopic offspring; and 3) participant AXL, but not ACD or CC, is associated with the parental myopic state. 

Several previous studies reported the impact of family history on the development of myopia. Mutti et al. ²³ defined myopia as ≤ −0.75 D, and reported an OR of 6.4 for two, compared with no parent with myopia, which is similar to our adjusted OR of 7.70. Wu and Edwards ²² recruited Chinese children from Hong Kong and two cities in China, where they defined myopia as spherical equivalent error ≤ −1 D, and uncorrected vision of worse than logMAR 0.18. They reported that the ORs of having myopia when they had two myopic parents were between 2.96 and 12.85, depending on which data were analyzed in their multisample study. Saw et al. ²⁸ reported relatively weak parental effect (OR = 3.1 for two vs. no myopic parents) in young Singaporean men; however, myopes were classified as any one with spherical equivalent < −0.5 D. The wide range of ORs among these studies may be due to sample variation, recruitment schemes, recall bias, definition of myopia, and various risks among different populations. However, these studies consistently indicated a parental dose–response relationship and suggested that parental myopic status is an important risk factor. Although our analysis suggested that the parental effect was more important than the measured environmental risk, we must note that the parental effect in our study may include genetic factors and unmeasured shared environmental factors (such as reading habits or the early exposure to near work). 

The present study found that parents had a strong influence on their children’s AXL (P = 2 × 10^–6). However, ACD and CC were not associated with the parental myopic state. Studying Singaporean children, Saw et al. ²⁹ also reported that AXL was highly associated with parental myopic state (P < 0.001), but the mean ACD was the same among children with one, two, or no myopic parents. Zadnik et al. ¹⁹ studied the eye size in American children and reported that even before the onset of myopia, children with myopic parents had longer eyes than those without myopic parents. These studies might suggest the importance of including AXL in vision screening programs in children. Similar to our findings, corneal curvature was not related to parental myopic state in studies either by Saw et al. ²⁹ or Zadnik et al. ¹⁹  

Using the age of first glasses for myopia as a surrogate of the onset of myopia may cause some concerns. Generally speaking, in Taiwan, eye glasses would not be prescribed to a child until myopia reaches approximately −2 D. One may argue that parents who have myopia are more likely to have their children tested early on. However, because of the high prevalence of myopia in Taiwan, vision screening is popular and many schools provide compulsory vision screening at no cost to parents. Therefore, the bias from myopic parents should not be substantial. Furthermore, Zadnik et al. ¹⁹ reported that children with myopic parents had longer eyes than those without myopic parents even before the onset of myopia. Their findings indirectly suggest an earlier onset of myopia when there is family history. Using the parental and sibling myopic state as risk factors, our study found that parents had a significant effect on the onset of myopia, especially high and moderate myopia. However, sibling myopic state was not related to the onset of the condition. No sibling effect on the onset of myopia was indeed unexpected. One possible explanation is that parents are in a generation where exposure to television, computer, video/computer games, or other near work was significantly less while they were young. Accordingly, parental myopia is more likely due to underlying genetic susceptibility. It is generally believed that a disease caused by genetic factors tends to have an earlier onset than the same disease caused by environmental factors. ^{25 26 27} As a result, having highly myopic parents will accelerate the onset of myopia in their offspring. On the contrary, sibling myopia can be profoundly influenced by the above environmental factors. Therefore, the onset of myopia in probands is less likely to be predicted by their siblings. It needs to be noted that the ORs discussed in this paragraph were used to test for the association between highly myopic parents/siblings and an early onset of myopia, rather than parent-offspring or sibling recurrent risk. 

There were some limitations in this study. Similar to any retrospective epidemiologic studies, our analysis may be subject to recall bias. Subconsciously, highly myopic probands are more likely to report having myopic parents and siblings than normal controls. To reduce the impact from this bias, we used only highly myopic parents and siblings as family risks, which is more robust to recall bias. Furthermore, the extremely strong association (OR > 6) between highly myopic parents and offspring is unlikely to be explained by recall bias alone. The insignificant relationship between environmental risk factors (except for education and TV watching) and highly myopic state (Table 2) also suggests a minimal influence from recall bias. As for the information of the age of first glasses for myopia, it is impossible to verify the reported age by checking optical/medical records since no such data are available. However, the strong statistical findings in our study would provide a good hypothesis to validate this finding in future studies. 

The present study found strong parental effects on the level and onset of myopia in their offspring. Although environmental effects also partially account for having myopia, they were less important than parental and sibling effects. The parental effect on probands’ myopic components was primarily on the AXL, but not on ACD or CC. 

 Supported in part by grants from the Department of Health, Kaohsiung, Taiwan, and from Kaohsiung Municipal United Hospital, Kaohsiung, Taiwan.

 Submitted for publication September 24, 2003; revised December 7, 2003, and March 11 and May 19, 2004; accepted June 4, 2004.

 Disclosure: C.-L. Liang, None; E. Yen, None; J.-Y. Su, None; C. Liu, None; T.-Y. Chang, None; N. Park, None; M.-J. Wu, None; S. Lee, None; J.T. Flynn, None; S.-H.H. Juo, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

(a) Distribution of current mean refractive error versus current age. Data included all participants. (b) Distribution of current mean refractive error versus myopia onset age. Data included only the participants who have nearsighted eye glasses. The number of participants in each age group is shown on the top of each bar.

Table 1.

 

Distribution of Participants among Different Parental and Sibling Myopic Status

Proband Myopic Status	Number of Highly Myopic Parents*					Number of Highly Myopic Siblings, †
		0	1	2	0	1	≥2
Normal	127	20	9	100	27	7
Mild	78	38	14	92	43	5
Moderate	64	35	14	57	40	10
High	145	125	62	132	184	21

*  The P value from 3 × 4 contingency χ² test = 5.94 × 10⁻¹⁴ (χ² ₆ = 64.32, degree of freedom (df) = 6). Parental information was missing/unknown in 157 participants.

†  The P value from 3 × 4 contingency χ² test = 1.67 × 10⁻¹² (χ² ₆ = 67.01, df = 6). Sibling information was missing/unknown in 173 participants.

Table 2.

 

Scores of Environmental Factors and Percent of College (or Higher) Education

Myopic Status	TV Mean (SD)	Game Mean (SD)	Computer Mean (SD)	Outdoor Mean (SD)	College Education (%)
Normal	2.19 (± 1.33)	2.15 (± 1.44)	2.12 (± 1.38)	1.88 (± 1.12)	43.3%
Mild	1.95 (± 1.02)	1.85 (± 1.14)	2.12 (± 1.21)	1.84 (± 1.01)	54.8%
Moderate	2.03 (± 1.11)	2.09 (± 1.31)	2.14 (± 1.27)	1.83 (± 1.13)	49.6%
High	2.60 (± 1.46)	2.26 (± 1.50)	2.01 (± 1.29)	1.71 (± 1.00)	63.0%

 Score 1 for ≤10 hours/week, score 2 for 11–20 hours/week, score 3 for 21–30 hours/week, score 4 for 31–40 hours/week, and score 5 for ≥40 hours/week.

Table 3.

 

Results from Multivariate Logistic Regression Model

	Coefficient	SE	P Value	OR (95% CI)
One highly myopic parent	1.85	0.30	1.12 × 10−9	6.36 (3.51–11.54)
Two highly myopic parents	2.04	0.45	6.14 × 10−6	7.70 (3.18–18.66)
Education (college or higher)	1.18	0.23	4.42 × 10−7	3.26 (2.06–5.16)
Gender (female vs. male)	0.79	0.34	0.02	2.21 (1.13–4.32)

Figure 2.

 

(a) The distribution of axial length (mm) among participants. (b) The distribution of anterior chamber depth (mm) among participants. (c) The distribution of corneal curvature among participants.

Table 4.

 

Relationship between Parental Myopic State and Axial Length (AXL), Anterior Chamber Depth (ACD), and Corneal Curvature (CC) in Participants

	Number of Highly Myopic Parents
		0	1	2
AXL (mm)*
≥26.71	61	56	35
≤25.78 and <26.71	80	56	21
≤24.71 and <25.78	89	47	21
<24.71	118	31	14
ACD (mm), †
≥3.885	39	40	22
≤3.73 and <3.885	49	34	18
≤3.54 and <3.73	37	32	24
<3.54	53	31	12
Horizontal CC (D), **
≥43.75	102	52	35
≤42.5 and <43.75	137	68	28
≤41.75 and <42.5	83	52	21
<41.75	91	45	13
Vertical CC (D), ***
≥45	112	55	39
≤43.75 and <45	108	66	26
≤42.75 and <43.75	100	53	18
<42.75	93	43	14

*  χ² ₆ = 37.89, P = 1.18 × 10⁻⁶;

†  χ² ₆ = 9.52, P = 0.15;

**  χ² ₆ = 8.77, P = 0.19;

***  χ² ₆ = 10.45, P = 0.11.

Table 5.

 

Family History and Proband’s Age of First Glasses for Myopia among Three Myopic Categories

Age at First Glasses (y)	Highly Myopic Parents			Highly Myopic Siblings
		≥1 Parent	None	≥1 Sibling	None
High myopia
≤11	116	57	105	67
>11	64	82	95	57
OR	2.61 (95% CI 1.65–4.11; P = 3.10× 10−5)			0.94 (95% CI 0.6–1.47; P = 0.788)
Moderate myopia
≤13	38	29	28	27
>13	10	32	20	29
OR	4.19 (95% CI 1.78–9.90; P = 0.001)			1.50 (95% CI 0.69–3.27; P = 0.303)
Mild myopia
≤15	35	40	32	47
>15	8	31	11	32
OR	3.39 (95% CI 1.4–8.19; P = 0.006)			1.73 (95% CI 0.77–3.92; P = 0.1)