There has been a long-standing debate over the role of nature (genetic variation) and nurture (environmental variation) in the etiology of myopia, which began with early observations of the link between myopia, education, and occupation.1–4 One consideration is that myopia often shows strong familial clustering, in terms of the correlations between myopia in parents and their offspring, and between siblings.5–12
It is often claimed that familial clustering suggests a substantial genetic contribution to the determination of refractive error, and particularly myopia, but this evidence is clearly ambiguous because families share environments as well as genes. As discussed by Guggenheim et al.13 and Morgan and Rose,4 sibling correlations tend to remain high, even when there have been rapid changes in the prevalence of myopia,14,15 although data from the Framingham Offspring Eye Study16 have shown that greater difference in age between the siblings significantly reduces the correlation. Guggenheim et al.17 found that, in the SCORM (Singapore Cohort study of the Risk factors for Myopia) study, the correlations between siblings in measures of environmental risk factors (e.g., using computers, reading, watching TV, playing video games, outdoor activity.) were often slightly higher than in myopia itself, emphasizing the potential for environmental explanations. In contrast to the stability of sibling correlations, when the prevalence of myopia has changed rapidly, which is generally associated with major changes in the social environment from one generation to the next, parent–offspring correlations are reduced, sometimes to trivial levels, consistent with strong environmental effects.4,6,7,13–15,18–21
Another way of looking at the relationship between myopia in parents and their children is to see whether there is an increased risk of the development of myopia in children whose parents are myopic. Several studies have shown that this is the case.22–25 This issue has been examined in a unique study of myopia over three generations in Hong Kong and China,26 which documented an increasing prevalence of myopia over the three generations, with a decline in the odds ratio (OR) for risk associated with parental myopia through the generations. We have now examined this issue using data from the population-based Guangzhou Refractive Error Study in Children (Guangzhou RESC) to determine the impact of parental myopia and possible confounding factors such as parental education, occupation, and income.27
The Guangzhou RESC survey27 carried out in 2002 used the standard RESC protocol, with randomized cluster sampling to provide a representative population-based sample of children combined with high examination participation rates and the use of cycloplegic refraction to determine spherical equivalent refraction (SER). The general RESC methodology, recruitment of subjects, and examination procedures in the Guangzhou RESC, as well as the results on prevalence of refractive error and visual impairment, have been described in detail elsewhere.27,28 In brief, 5330 children aged 5 to 15 years living in Liwan District, Guangzhou, were enumerated (based on the Guangzhou 2000 Census) using cluster sampling followed by door-to-door verification. A total of 4364 children were examined; 86.4% of those were enumerated. Informed consent for each child examined was obtained from a parent or other responsible adult. The research followed the tenets of the Declaration of Helsinki. The ethics committees of both the Zhongshan Opthalmic Center and the Liwan District Bureau of Education and Bureau of Health approved the study.
Cycloplegia was induced with 2 drops of 1% cyclopentolate administered 5 min apart to each eye, with a third drop administered after 20 min. Cycloplegia was considered complete if the pupil dilated to 6 mm or more, and a light reflex was absent after a further 15 min. Refractions were measured with a streak retinoscope (Welch-Allyn, Skaneateles, NY) and then a handheld auto-refractor (ARK-30; Nidek Corp.), after cycloplegia. For this article, only auto-refraction data from right eyes with complete cycloplegia were analyzed. Myopia was defined as an SER error of at least −0.50 diopter (D).
After enumeration to identify eligible children, for children attending schools, the project ophthalmologist visited the schools 3 days before the eye examinations to distribute parent questionnaires, which were completed by the biological parents and returned before the examination. For children not yet enrolled in school, parent questionnaires were completed by the parents in temporary community examination stations. The investigators were available to answer queries. The questionnaires investigated children's near-work activities, living habits, parental refractive status, and family socioeconomic background. In general, questions were posed with multiple-choice answers. However, parental occupation was posed as an open-ended question.
Parental refractive status was determined with a question for each parent: “Is the child's father (or mother) myopic?” (YES/NO/NOT SURE). To examine whether this simple question is valid to estimate parents' refractive status, a sample of 98 consecutive subjects drawn from the parents of children who were participating for the first time in the Guangzhou Twin Study (n = 49) or who visited the Zhongshan Ophthalmic Center Optometry Clinic (n = 49) during the validation period was asked to answer the same question before undergoing non-cycloplegic auto-refraction. The response to the question and the results of non-cycloplegic refraction are shown in Table 1. The question yielded both high sensitivity (0.77) and specificity (0.84) for the detection of myopia, with a positive predictive value of 0.98 and a negative predictive value of 0.79, and a kappa coefficient of 0.77 (95% CI: 0.64–0.90) for overall agreement between questionnaire and auto-refraction data.
Family highest education level was based on the highest level of schooling achieved by the parents, in categories of no formal schooling, primary, junior high school, senior high school, and university. Because of the low numbers in the less educated categories, education levels were grouped as primary or less, secondary, and tertiary for analysis.
Family monthly income categories were ≤2000 renminbi (RMB), 2000 to 5000 RMB, 5000 to 10,000 RMB, and >10,000 RMB; they were then grouped as ≤2000 RMB and >2000 RMB for analysis because of very low numbers in the higher income groups. Average monthly income was 2100 RMB in Guangzhou in 2002 (Guangzhou Yearbook 2002), which equates to just above U.S. $300 per month.
From responses to an open-ended question, parental occupations were classified into three categories: Category A: Unemployed/Home duties/Farming; Category B: Blue collar/Physical labor/Manual worker; Category C: White collar/Administrative worker/Officer/Professional.
After being reviewed for accuracy and completeness, clinical examination forms and questionnaires were transferred to Zhongshan Ophthalmic Center for data entry. Measurement data ranges, frequency distributions, and consistency among related measurements were checked with data-cleaning programs. The χ2 test was used to compare proportions (prevalence) between different groups. Given the large sample size, ANOVA was used to compare the mean SER across different groups. The associations between myopia and child's age, gender, parental myopic status, parental education level, income, and occupation were explored by univariate and multivariate logistic regressions. Statistical analyses were performed using SPSS 16.0 (SPSS Inc., Chicago, IL). All statistical tests were two-sided, and p < 0.05 was considered statistically significant.
From a total of 4364 children with cycloplegic refractions, the biological parents of 3618 (82.9%) children responded to the parental myopia questions. The children with information on parental myopia were not significantly different from the children without information on parental myopia in terms of age (10.6 ± 2.9 years vs. 10.7 ± 12.3 years, p > 0.05), gender (boys: 51.5% vs. 52.3%, p > 0.05), or refractive error (−2.5 ± 1.9 D vs. −2.4 ± 1.7 D, p > 0.05).
A detailed breakdown of parental myopia is shown in Table 2. Overall, just <20% (95% CI: 18.9%–20.7%) of the parents were identified as myopic using questionnaire response, with a slightly higher prevalence of myopia in mothers than in fathers (p < 0.001). Also, <2% of the sample was unsure of the answer to this question, and these parents and their children were excluded from this analysis. The prevalence of myopia in the parents was less than that measured in the total sample of children (38.1%, 95% CI: 36.5%–39.7%) and considerably less than the prevalence in children aged 15 years (78.4%, 95% CI: 73.9%–82.6%).
Preliminary analysis of individual parental characteristics showed, as expected, that myopic parents were more likely to have postsecondary education (27.6% in myopic parents vs. 8.4% in non-myopic parents), to work in white collar and professional occupations (38.4% in myopic parents vs. 21.8% in non-myopic parents), and to be members of families with higher incomes (54.2% in myopic parents vs. 37.1% in non-myopic parents). Children were classified into three familial categories: no parents myopic (NPM)—66.2% (2395), one parent myopic (OPM)—28.1% (1016), and two parents myopic (TPM)—5.7% (207). Compared with what would be expected from random mating—no myopic parents of 64.3% and two myopic parents of 3.9%—there was a slight excess of no myopic parents and two myopic parents, suggesting that there was only very slight assortative mating. As shown in Table 3, parents in families with more myopic parents tended to have higher levels of education, higher incomes, and were more likely to be in white collar and professional occupations. Distributions between types of families were significantly different for each characteristic (χ2 test, p < 0.001). Thus, analysis in terms of parental myopia is confounded with parental education, income, and occupation.
The age range of the children examined in this study was 5 to 15 years, a span of 10 years. All groups of parents were predominantly not myopic, but the proportion of parents being myopic nearly doubled from 17.2% (95% CI: 13.3%–21.4%) for the parents of 15-year-old children born in 1987, to 32.1% (95% CI: 27.2%–37.1%) for the parents of the 5-year-old children born in 1997 (p < 0.001). A similar, but somewhat larger, change was also seen in the data on highest parental education levels. The proportion of parents with no formal education or primary schooling only was very small for all age groups, but an increasing proportion of the parents had some form of tertiary education, from just over 10% (95% CI: 7.1%–13.5%) for the parents of the 15-year olds to around 30% (95% CI: 25.4%–35.1%) for the parents of the 5-year olds (p < 0.001). The different characteristics of the two cohorts of parents emphasize the rapidity of sociodemographic change in China during this period.
Impact of Parental Myopia on Myopia in the Children
The impact of parental myopia on the development of myopia in children was assessed by following the prevalence of myopia in the children by age. As can be seen in Fig. 1A, the prevalence of myopia increased in parallel in the three groups of children—from NPM, OPM, and TPM families—to reach a maximum of 88.9% (95% CI: 80.4%–93.8%) in the OPM group and 83.3% (95% CI: 58.9%–94.6%) in the TPM group at the age of 15 years, but was lower at 68.2% (95% CI: 61.6%–73.7%) in the NPM group. The myopic status of the parents had a small impact on the development of myopia in the children, as the children in the OPM and TPM groups were respectively 30% and 21% more likely to become myopic, compared with the NPM group, at the age of 15 years. The age-related increase in the prevalence of myopia by age 15 years in all groups was much greater than the difference made by parental myopia.
The impact of parental myopia can also be seen in mean SER (Fig. 1B). By the age of 15 years, the mean SER values for the OPM and TPM groups were close to −3 D, compared with close to −1.5 D in the NPM group.
Impact of Parental Education Level
Analysis of the impact of parental education is presented in terms of the highest parental education level in the family. Fig. 2A shows the prevalence of myopia in children by parental highest education level, which was categorized into primary or less (including no formal schooling), secondary, and tertiary education. Using this division, 15-year-old children with a tertiary-educated parent showed a higher risk of developing myopia, with a prevalence of 85.3% (95% CI: 77.3%–90.7%), compared with 73.3% (95% CI: 66.2%–79.6%) in those whose more educated parent had secondary education, and 60.0% (95% CI: 46.4%–71.9%) in those with parents with only primary school education. Analyses were also carried out for father's or mother's highest education level, with essentially similar results.
Fig. 2B illustrates the refractive error of children in the groups defined in terms of highest parental education level. In younger children, the increase in SER with age was similar in all groups, but after the age of 9 to 10 years, the gaps between groups became greater, with mean SER generally more myopic in the children with a parent with secondary and tertiary education.
Impact of Family Income or Parental Occupation
Children in the two income groups had almost the same prevalence data across the age range, with a slightly, but significantly, higher prevalence of myopia (81.7%; 95% CI: 75.2%–87.0%) in the higher income group vs. the lower income group (69.9%; 95% CI: 62.7%–76.6%), and a slightly more myopic mean SER at the age of 15 years in the higher income group (−2.2 D vs. −1.6 D). Similarly, parental occupation had little impact, with 81.3% (95% CI: 73.5%–88.1%) myopia at age 15 years in children from white collar families, with a relative risk of 1.04.
Logistic Regression Analysis
Table 4 shows the results of univariate and multivariate logistic regression. In univariate logistic regression, childhood myopia was associated with age, female gender, number of myopic parents, level of parental education, and more near-work time. In multivariate logistic regression, age, gender, and number of myopic parents were significantly associated with myopia in the children. Parental education, family income, father's occupation, and near work were not significant in the final model. As shown in Table 5, children from NPM, OPM, and TPM families showed significant differences in reported activity patterns. Of the statistically significant effects, children from TPM families reported that they were less likely to read and write for >2 h/d and were less likely to read or write for >2 h without a break. They were also less likely to watch television for more than an hour a day. Similar results (not shown) were obtained in relation to parental education groups. In multivariate logistic regression that included interactions between near work (read and write time) and parental characteristics (number of myopic parents, highest education level, occupation), the results showed that none of the interactions were significant in the final model.
A critical question for this analysis is the validity of the questionnaire for determining parents' refractive status. The results suggest that the simple questions used in the Guangzhou RESC can be used to estimate the prevalence of myopia in adults of parental age, with high sensitivity and specificity. The kappa coefficient showed substantial agreement29 between the questionnaire responses and non-cycloplegic refractions.
Two other factors should be considered in assessing the validity of the questionnaire responses. First, the parents recruited for the validation study were either participating in another refractive error study, or had brought their children to an optometry clinic for assessment, and thus may have been more aware of issues concerning refractive error than the general population. Although this raises issues concerning the application of this questionnaire to a naive sample from the general population, the parents in the Guangzhou RESC study had given informed consent to the participation of their children in a study of refractive error, which involved cycloplegia, after information about the significance of myopia, and, thus, were also likely to be better informed than the general population. It should also be noted that the validation was performed based on non-cycloplegic auto-refractions for the parents. Because non-cycloplegic refractions overestimate the prevalence of myopia in this age-group,30 it is likely that some of the responses reporting no myopia, classified as incorrect in this validation, may actually have been correct. Overall, these results show that the questionnaire responses have substantial validity.
Previous analyses of questionnaire estimation of parental myopia suggested that simple direct questions of this kind were not valid.31 The greater accuracy of the simple questions in this study may be owing to a high awareness of myopia among Chinese parents in urban areas, and because the parents enrolling their children in an eye study or seeking clinical assessment were more aware about eye problems or myopia. Either way, the use of the questionnaire in the current analysis appears likely to be valid.
Our results demonstrate a major change in the prevalence of myopia between generations in Guangzhou, and show that many children in Guangzhou become myopic, irrespective of whether their parents are myopic. Nevertheless, children with myopic parents are more likely to become myopic than children without myopic parents, with a multivariate OR of 2.84 for children with two myopic parents and 1.96 for children with one myopic parent. However, it should be noted that ORs give a reasonable approximation to relative risk only for low prevalence conditions.32 For myopia, with generally higher prevalence, ORs will systematically overestimate the deviation from a relative risk of 1.
Parental education was also associated with myopia in children in the univariate analyses, although this factor was not significant in the multivariate model. Parental income and occupation had much weaker univariate associations. Multivariate logistic regression showed that, after inclusion of parental myopia in the models, other parental characteristics including parental education did not exert significant effects.
These data show that all groups of children, from NPM, OPM, and TPM families, become substantially myopic in Guangzhou. In particular, the proportion of myopic children from families without myopic parents was 68.0% by the age of 15 years. We suggest that this high “baseline” level of myopia is associated with environmental pressures, including increased educational pressures and urbanization. Having myopic parents increases the risk of myopia over this baseline level by 20 to 30%.
The impact of parental myopia is often assumed to be genetic, although it is clear that families share environments as well as genes. As an alternative, it has been suggested that myopic parents, who are generally better educated, create myopigenic environments characterized by higher educational pressures, and perhaps by less time spent outdoors. Although we found significant associations between time spent reading or writing, and times spent reading or writing without a break by the children, and having myopic parents, the children with myopic parents reported spending less time reading or reading without a break. This is the reverse of the associations between near-work activities and myopia that have generally been reported.23,24 The relative risk and univariate OR for near work and myopia in the children were significant, but these factors were not significant in multivariate regression analysis. This weakens the case that higher near-work hours are a significant risk factor for myopia33 and suggests that the effect of parental myopia cannot be explained in these terms. Similar results were obtained in relation to parental education. We have no information on differences in the amount of near work performed by the parents when they were children, and the amount performed by their children, and therefore we cannot estimate the role that increased near work may have played in generating the inter-generational differences.
The fact that differences in near work performed by the children does not account for the differences in prevalence of myopia seen between groups of children based on levels of parental myopia, does not preclude a role for changes in educational pressures and, possibly, the amount of near work performed between generations. In many cases, the same risk factors will be active both within and between generations, but, in principle, how important these factors are will depend on the amount of variation in the factor that exists within and between generations. School enrolment and completion data (Guangzhou Yearbook http://data.gzstats.gov.cn/gzStat1/chaxun/njsj.jsp) show that there has been a major shift between these generations. In the parental generation, a small proportion of parents did not proceed beyond primary school, and only about 10% completed high school and went on to tertiary education. In the current generation of children, 9 years of formal education is now compulsory, very close to 100% complete 12 years of formal education, and at least 80% go on to some form of further education (82% of senior high school graduates entered tertiary education level in 2010). Thus, the variation in education on these measures is high between the generations and in the parental generation, whereas it is low in the children. This social change is also reflected in the different characteristics of the parents of the older and younger children. When using a categorical variable such as myopia, particularly if a threshold level of schooling tends to lead to it, education could be an important factor in the parental generation and between the generations, but a much less important or even insignificant factor in the children. There is also a considerable body of anecdotal evidence that the current generation of children spend much more time on study than their parents, but this parameter has not, to our knowledge, been quantified. Many other potential risk factors have changed over this historical period in Guangzhou, but educational change is a highly plausible one, given the known associations between myopia and education.4
This study was carried out before time spent outdoors was shown to be negatively associated with myopia,23,34–36 and data on this parameter were not collected. Analysis of differences in the amount of time spent outdoors is required, given its strong protective effects,34,36 before it can reasonably be concluded that the effects of parental myopia are genetic.
The results of other major studies on the risk associated with parental myopia are summarized in Table 6. Although all the studies support the idea that having myopic parents increases the risk of myopia, there is no consistent relationship between childhood myopia and number of myopic parents.
An obvious feature of the data in Table 6 is that, where data are available, children of East Asian ancestry are more myopic than their parents, whereas in populations of European origin, the children are consistently less myopic than their parents. However, it should be noted that the definitions for myopia in children and their parents in these studies were different. In our study, myopia in the children was based on cycloplegic auto-refraction, whereas parental myopia was based on questionnaire. Some of the other studies used cycloplegic refraction on children, whereas others measured non-cycloplegic refraction. A variety of approaches was used for estimating parental myopia ranging from questionnaire administration, prescription analysis, and non-cycloplegic refraction. Therefore, care needs to be taken in comparing the values in the parents and children. Allowance also needs to be made for the ages at which the children were examined, often before the full development of myopia, and it is therefore possible that there will be continuing development of myopia in the children of European origin after the age at which they were studied. But, in the case of both the Orinda study23 and the Sydney Myopia Study,39 there would need to be considerable development of myopia throughout the high school years for the children to develop a similar or higher prevalence of myopia to their parents. Considerable later-onset myopia in populations of European and, more broadly, Caucasian origin40,41 has been reported, and thus this possibility cannot be dismissed.
One strength of this study is that the sample was population-based, with a high participation rate, and with statistical randomization in recruitment. These characteristics are shared with the Sydney Myopia Study.42 Other studies did not share all these characteristics. Another strength of this study was the measurement of cycloplegic refraction, as was also the case for the Sydney Myopia Study, the Orinda Study, and the SCORM (Singapore Cohort Study of the Risk factors for Myopia) study. However, where cycloplegia was not used, the prevalence of myopia may have been overestimated.40
Limitations of this study are the diversity of ages at which myopia was measured in the children, often before the full development of myopia, and that myopia in the parents was self-reported in questionnaire responses, as were other parental characteristics. The other significant limitation is in the collection of data on activity patterns in the children, as data were only collected by questionnaire on near-work factors, not on time spent outdoors. Other factors such as maternal age43 and birth order,44,45 which have been reported to be associated with myopia, have not been considered.
Despite these limitations, this study shows a marked inter-generational changes in the prevalence of myopia in Guangzhou, with many children in Guangzhou becoming myopic without myopic parents. There was a small additional risk of myopia in those with myopic parents. This additional risk could not be explained in terms of increased engagement in near-work activities by the children with myopic parents. This suggests that additional genetic risk may be involved, although further analysis of the role of time spent outdoors is required.
Department of Preventive Ophthalmology
Zhongshan Ophthalmic Center
People's Republic of China
This work was supported by World Health Organization under National Institutes of Health contract N01-EY-2103, the Fundamental Research Funds for the State Key Laboratory for Ophthalmology, Zhongshan Ophthalmic Center, Guangzhou. IGM's contribution was supported by the Australian Research Council through a grant to the ARC Centre of Excellence in Vision Science (CEO561093).
1. Curtin BJ. The Myopias. Basic Science and Clinical Management. Philadelphia, PA: Harper & Row; 1985.
2. Mutti DO, Zadnik K, Adams AJ. Myopia. The nature versus nurture debate goes on. Invest Ophthalmol Vis Sci 1996;37:952–7.
3. Saw SM, Katz J, Schein OD, Chew SJ, Chan TK. Epidemiology of myopia. Epidemiol Rev 1996;18:175–87.
4. Morgan I, Rose K. How genetic is school myopia? Prog Retin Eye Res 2005;24:1–38.
5. Sorsby A, Leary GA, Fraser GR. Family studies on ocular refraction and its components. J Med Genet 1966;3:269–73.
6. Keller JT. A comparison of the refractive status of myopic children and their parents. Am J Optom Arch Am Acad Optom 1973;50:206–11.
7. Hegmann JP, Mash AJ, Spivey BE. Genetic analysis of human visual parameters in populations with varying incidences of strabismus. Am J Hum Genet 1974;26:549–62.
8. Ashton GC. Segregation analysis of ocular refraction and myopia. Hum Hered 1985;35:232–9.
9. Lee KE, Klein BE, Klein R, Fine JP. Aggregation of refractive error and 5-year changes in refractive error among families in the Beaver Dam Eye Study. Arch Ophthalmol 2001;119:1679–85.
10. Klein AP, Duggal P, Lee KE, Klein R, Bailey-Wilson JE, Klein BE. Support for polygenic influences on ocular refractive error. Invest Ophthalmol Vis Sci 2005;46:442–6.
11. Wojciechowski R, Congdon N, Bowie H, Munoz B, Gilbert D, West SK. Heritability of refractive error and familial aggregation of myopia in an elderly American population. Invest Ophthalmol Vis Sci 2005;46:1588–92.
12. Fotouhi A, Etemadi A, Hashemi H, Zeraati H, Bailey-Wilson JE, Mohammad K. Familial aggregation of myopia in the Tehran Eye Study: estimation of the sibling and parent offspring recurrence risk ratios. Br J Ophthalmol 2007;91:1440–4.
13. Guggenheim JA, Kirov G, Hodson SA. The heritability of high myopia: a reanalysis of Goldschmidt's data. J Med Genet 2000;37:227–31.
14. Young FA, Leary GA, Baldwin WR, West DC, Box RA, Harris E, Johnson C. The transmission of refractive errors within Eskimo families. Am J Optom Arch Am Acad Optom 1969;46:676–85.
15. Alsbirk PH. Refraction in adult West Greenland Eskimos. A population study of spherical refractive errors, including oculometric and familial correlations. Acta Ophthalmol (Copenh) 1979;57:84–95.
16. The Framingham Offspring Eye Study Group. Familial aggregation and prevalence of myopia in the Framingham Offspring Eye Study. Arch Ophthalmol 1996;114:326–32.
17. Guggenheim JA, Pong-Wong R, Haley CS, Gazzard G, Saw SM. Correlations in refractive errors between siblings in the Singapore Cohort Study of Risk factors for myopia. Br J Ophthalmol 2007;91:781–4.
18. Nakajima A. Refractive elements of the eye as metric traits [in Japanese]. Nihon Ganka Gakkai Zasshi 1968;72:2059–82.
19. Nakajima A, Kimura T, Kitamura K, Uesugi M, Handa Y. Studies on the heritability of some metric traits of the eye and body. Jpn J Med Genet 1968;13:20–39.
20. Young FA, Leary GA. The inheritance of ocular components. Am J Optom Arch Am Acad Optom 1972;49:546–55.
21. Johnson GJ, Matthews A, Perkins ES. Survey of ophthalmic conditions in a Labrador community: part I. Refractive errors. Br J Ophthalmol 1979;63:440–8.
22. Pacella R, McLellan J, Grice K, Del Bono EA, Wiggs JL, Gwiazda JE. Role of genetic factors in the etiology of juvenile-onset myopia based on a longitudinal study of refractive error. Optom Vis Sci 1999;76:381–6.
23. Mutti DO, Mitchell GL, Moeschberger ML, Jones LA, Zadnik K. Parental myopia, near work, school achievement, and children's refractive error. Invest Ophthalmol Vis Sci 2002;43:3633–40.
24. Saw SM, Chua WH, Hong CY, Wu HM, Chan WY, Chia KS, Stone RA, Tan D. Nearwork in early-onset myopia. Invest Ophthalmol Vis Sci 2002;43:332–9.
25. Ip JM, Huynh SC, Robaei D, Rose KA, Morgan IG, Smith W, Kifley A, Mitchell P. Ethnic differences in the impact of parental myopia: findings from a population-based study of 12-year-old Australian children. Invest Ophthalmol Vis Sci 2007;48:2520–8.
26. Wu MM, Edwards MH. The effect of having myopic parents: an analysis of myopia in three generations. Optom Vis Sci 1999;76:387–92.
27. He M, Zeng J, Liu Y, Xu J, Pokharel GP, Ellwein LB. Refractive error and visual impairment in urban children in southern China. Invest Ophthalmol Vis Sci 2004;45:793–9.
28. Negrel AD, Maul E, Pokharel GP, Zhao J, Ellwein LB. Refractive Error Study in Children: sampling and measurement methods for a multi-country survey. Am J Ophthalmol 2000;129:421–6.
29. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977;33:159–74.
30. Fotouhi A, Morgan IG, Iribarren R, Khabazkhoob M, Hashemi H. Validity of noncycloplegic refraction in the assessment of refractive errors: the Tehran Eye Study. Acta Ophthalmol 2012;90:380–6.
31. Walline JJ, Zadnik K, Mutti DO. Validity of surveys reporting myopia, astigmatism, and presbyopia. Optom Vis Sci 1996;73:376–81.
32. Zhang J, Yu KF. What's the relative risk? A method of correcting the odds ratio in cohort studies of common outcomes. JAMA 1998;280:1690–1.
33. Mutti DO, Zadnik K. Has near work's star fallen? Optom Vis Sci 2009;86:76–8.
34. Rose KA, Morgan IG, Ip J, Kifley A, Huynh S, Smith W, Mitchell P. Outdoor activity reduces the prevalence of myopia in children. Ophthalmology 2008;115:1279–85.
35. Rose KA, Morgan IG, Smith W, Burlutsky G, Mitchell P, Saw SM. Myopia, lifestyle, and schooling in students of Chinese ethnicity in Singapore and Sydney. Arch Ophthalmol 2008;126:527–30.
36. Jones LA, Sinnott LT, Mutti DO, Mitchell GL, Moeschberger ML, Zadnik K. Parental history of myopia, sports and outdoor activities, and future myopia. Invest Ophthalmol Vis Sci 2007;48:3524–32.
37. Ip J, Robaei D, Rochtchina E, Rose K, Smith W, Wang JJ, Mitchell P. Can information on the purpose of spectacle use and age at first use predict refractive error type? Ophthalmic Epidemiol 2007;14:88–92.
38. Gwiazda J, Thorn F, Bauer J, Held R. Emmetropization and the progression of manifest refraction in children followed from infancy to puberty. Clin Vis Sci 1993;8:337–44.
39. Ip JM, Huynh SC, Robaei D, Kifley A, Rose KA, Morgan IG, Wang JJ, Mitchell P. Ethnic differences in refraction and ocular biometry in a population-based sample of 11–15-year-old Australian children. Eye 2008;22:649–56.
40. Hashemi H, Fotouhi A, Mohammad K. The age- and gender-specific prevalences of refractive errors in Tehran: the Tehran Eye Study. Ophthalmic Epidemiol 2004;11:213–25.
41. Cumberland PM, Peckham CS, Rahi JS. Inferring myopia over the lifecourse from uncorrected distance visual acuity in childhood. Br J Ophthalmol 2007;91:151–3.
42. Ojaimi E, Rose KA, Smith W, Morgan IG, Martin FJ, Mitchell P. Methods for a population-based study of myopia and other eye conditions in school children: the Sydney Myopia Study. Ophthalmic Epidemiol 2005;12:59–69.
43. Rahi JS, Cumberland PM, Peckham CS. Myopia over the lifecourse: prevalence and early life influences in the 1958 Br birth cohort. Ophthalmology 2011;118:797–804.
44. Peckham CS, Gardiner PA, Goldstein H. Acquired myopia in 11-year-old children. BMJ 1977;1:542–5.
45. Rudnicka AR, Owen CG, Richards M, Wadsworth ME, Strachan DP. Effect of breastfeeding and sociodemographic factors on visual outcome in childhood and adolescence. Am J Clin Nutr 2008;87:1392–9.
parental myopia; myopia; children