Methodological quality of the review: Medium confidence
Author: Xiaoyu Zhang, Xinhua Qu and Xingtao Zhou
Region: Australia, China, Taiwan, Singapore, USA, Greece, UK, Turkey, Jordan, Oman, Finland
Subsector: Parental myopia as risk of myopia in a child
Equity focus: No
Study population: Children, adolescents or youth
Type of programme: Community based
Review type: Other review
Quantitative synthesis method: Meta-analysis
Qualitative synthesis method: Not applicable
Background: Myopia is a global health problem that has social, educational and economic consequences, and significantly affects the quality of life of sufferers. There is growing evidence to suggest that the prevalence of myopia is increasing; it is one of the five ocular conditions that are considered an immediate priority by the World Health Organization’s Global Initiative for the Elimination of Avoidable Blindness. A number of epidemiological studies have demonstrated a positive association between the prevalence of myopia in parents and a child’s risk of developing myopia. The association between parental myopia and a child’s risk of developing myopia has not yet, to the best of our knowledge, been investigated through meta‑
Objectives: To quantitatively investigate the association between parental myopia and a child’s risk of developing myopia.
Main findings: A total of 16 studies were eligible for final inclusion. Of these, six were prospective cohort, eight were cross‑sectional and two were case‑control studies. Over the 16 studies, 8,393 cases of myopia were studied on 31,677 participants. All of the studies recruited male and female participants aged ≤31 years. Seven of the studies identified the participants’ myopia as a spherical equivalent refraction (SER) ≤‑5 diopters (D) and two studies identified the participants’ myopia as a SER ≤‑0.75 D (26,32). Three of the studies reported the participants’ myopia as a SER ≤‑0.75 D in the horizontal and vertical meridians following cycloplegic autorefraction. The participants’ myopia in another four studies was identified as SER ≤‑1.5 D, a visual acuity ≥6/9 following correction with a concave lens >0.5 D, a SER between ‑0.75 and ‑2.99 D, and self‑reported wearing of ‘minus’ glasses.
There was a statistically significant positive association between myopia in one or two parents and a child’s risk of developing myopia. The ORs and 95% CIs in the prospective cohort, cross‑sectional and case‑control studies were 1.53 (95% CI, 1.21‑1.85), 1.96 (95% CI, 1.53‑2.39), and 2.13 (95% CI, 1.79‑2.46), respectively, for myopia in one parent and 2.10 (95% CI, 1.42‑2.77), 2.96 (95% CI, 2.21‑3.71), and 2.13 (95% CI, 1.79‑2.46), respectively, for myopia in two parents. No heterogeneity was observed in the case‑control studies with myopia in one or two parents (P=0.39, I2=0%). Visual inspection of the funnel plot revealed a certain level of asymmetry. The Begg’s test (P=0.18 and P=1.00 for included cohort and cross‑sectional studies, respectively) and Egger’s test (P=0.05 and P=0.46 for included cohort and cross‑sectional studies, respectively) did not suggest any evidence for publication bias in the analysis of participants with one parent with myopia. For the prospective cohort studies, age, geographical location, recruitment date or years of follow‑up did not significantly influence the association between myopia in one parent and a child’s risk of developing myopia. Geographical location and recruitment date were identified to be possible sources of heterogeneity (P<0.01 and P=0.03, respectively) in the studies of myopia in two parents. Age and geographical location were observed as sources of heterogeneity (both P<0.01) in studies with myopia in one parent; however, there was no evidence that age or geographical location were a source of heterogeneity in studies with myopia in two parents.
Studies were included in the meta‑analysis if they fulfilled the following criteria: i) children, adolescents or youth were included as participants; ii) the exposure of interest was parents with myopia; iii) the outcome of interest was myopia among children (prevalent or incident) and; iv) risk estimates, including relative risks (RRs), ORs, hazard ratios (HRs), or other measures that it was possible to transform into ORs with 95% confidence intervals (CIs), were reported.
The Medline (articles from 1966 to June 1, 2013), Embase (articles from 1980 to 1 June 2013), and Ovid (articles from 1950 to 1 June 2013) databases were searched for prospective cohort, cross‑sectional and case‑control studies that did not have access restrictions. All relevant studies using Medical Subject Headings (MeSH) or free text words were selected. The MeSH search strategies were followed and the search terms for exposure (parent, parental, family, history), and outcomes (myopia, myopic, short‑sight, short sight, near‑sight, near sight, refractive errors) were combined. Furthermore, a number of potential studies were identified electronically by searching the reference lists of the relevant publications. These publications were scrutinised in an effort to identify additional relevant studies.
The OR was used to assess associations across studies. RRs and HRs were transformed into ORs using a previously described method. The OR was pooled to summarise the associations between one or two parents with myopia and a child’s risk of developing the condition. Pooled estimations and complete analyses across studies were obtained using random‑effects models throughout the meta‑analysis. The heterogeneity of the studies was assessed using Cochran’s Q test and the I2 statistic. As suggested by Higgins and Thompson, I2 values of 25, 50 and 75% were considered to indicate low, moderate and high heterogeneity, respectively. Stata 10 (StataCorp, College Station, TX, USA) was used to carry out all the analyses. P<0.05 was considered to indicate a statistically significant difference.
Applicability/external validity: Regarding the external validity, the authors reported that this meta-analysis should be treated with cautionary attention to its limitations. Firstly, as it is based on the results of observational studies, the possibility that other factors may explain the observed associations between parental myopia, and a child’s risk of developing myopia cannot be excluded. Therefore, the possibility of residual confounders remains. It is also difficult to completely rule out that either genetic factors or a shared parent‑child environment was responsible for the observed associations. Secondly, data deficits, data restriction and data of variable quality were used with varying definitions for myopia, and this may have weakened the strength of the associations observed. Publication bias existed in our analysis, as shown by the funnel plot and the Egger’s and Begg’s tests. Finally, methodological differences in the designs of the studies may have introduced heterogeneity.
Geographic focus: No study from LMIC was identified, nonetheless, the authors included studies from Asia, Europe, Middle-East and United States of America with different ethnic groups. The result can be applied in other settings from those same regions.
Summary of quality assessment: In conclusion, this meta-analysis was attributed medium confidence. Several limitations were identified by the authors which can undermine the quality of this meta-analysis. Additionally, the authors did not clearly explain how studies’ quality assessment was done and criteria used. The low number of studies included have reduced the strength of the results obtained in the current study.