The article by Zhang and colleagues, 1 using data from the Hong Kong Children Eye Study, 2 presents a detailed picture of the associations between myopia and exposure to secondhand smoking (SHS) in children aged 6 to 8 years. Exposure to SHS was associated with greater myopic refraction, longer axial length, and earlier onset of myopia. The statistically significant effect sizes were small, but consistent, and showed a dose-response association with severity of exposure and younger age.Every increase in units of 10 cigarettes per day was associated with 0.07D more myopia, and overall exposure to SHS was associated with 1.8 months earlier onset of myopia. 1 The data were internally consistent, with the differences associated with exposure to SHS in spherical equivalent refraction and axial length consistent with biological explanations. This analysis shows the power of analyzing factors associated with the risk of myopia in a well-designed study that uses a large sample of children of a similar age, to minimize age-related confounding, and that uses systematic cycloplegia, a combination that is unfortunately all too rare in studies from East Asia at the moment. Performing such studies is hard work, primarily in persuading parents to give informed consent, but the effort pays off. This study 1 is the most precise evidence of an association between exposure to SHS and myopia in children currently available.A recent review 3 of factors associated with the risk of myopia concluded that previous studies have given conflicting results on smoking, probably owing to uncontrolled confounding. Although the study by Zhang et al 1 is cross-sectional and, therefore, cannot prove a causal relationship, crosssectional studies can provide the starting point for generating testable hypotheses about causation.The existence of a dose-response curve is consistent with a causal relationship, but does not prove it. Pharmacological studies 4 have shown that both cholinergic nicotinic agonists, such as nicotine, and nicotinic antagonists inhibit the development of experimental myopia; thus, a possible causal pathway is not well defined. Further longitudinal evidence would be particularly useful, but it is hard to imagine an ethics-approvable clinical trial in which children are randomized to receive SHS, so unless mendelian randomization provides additional information, the issue of causality may be hard to resolve.The authors 1 have tried to address the issue of confounding, with some detailed statistical analysis. In most studies, myopia is associated with higher socioeconomic status (SES), presumably owing to increased parental education, whereas smoking tends to be more common in lower SES groups. This pattern would suggest that exposure to SHS should tend to be associated with lower parental education and income, as the data in this article suggest, and less myopia in the children.However, there are also indications of a greater complexity to the issue in Hong Kong, since Choi et al 5 have reported that myopia in Hong Kong is associ...