Methodological quality of the review: Medium confidence

Author: Zhi-Peng Lu, Man-Zhao Ouyang, Ran Zhang, Xiao Tang, Hao-Jie Zhong

Region: China

Sector: Myopia

Subsector: Chinese eye exercises and onset of myopia

Equity focus: School children

Study population: School students

Type of programme: School based

Review type: Other review

Quantitative synthesis method: Meta-analysis

Qualitative synthesis method: Not applicable

Background: Myopia has emerged as a significant public health concern, with a rapidly increasing prevalence. A recent report estimated that approximately 1.4 billion individuals, worldwide, are myopic, comprising >20% of the population. Myopia has already reached epidemic levels in certain areas of East and Southeast Asia. Myopia is a risk factor for pathological ocular changes, such as cataracts, glaucoma, and myopia macular degeneration. It is an important cause of impaired vision and blindness. There is a need to introduce a simple and easy-to-use intervention for prevention of myopia onset. Chinese eye exercises, an intervention for visual protection and myopia prevention, originated from the theories of traditional Chinese medicine. However, the effects of Chinese eye exercises on prevention of myopia remain inconsistent.

Objectives: To investigate the association between Chinese eye exercises and onset of myopia.

Main findings: A total of five studies conducted in China were included in this meta-analysis. Four studies were cross-sectional and one was a case-control study. Study quality scores of the four cross-sectional studies ranged from eight to 11 stars, according to the AHRQ, while the case-control study received a score of nine scores, according to the NOS scale. This meta-analysis included data for a total of 14,590 participants (range: 60-11,138 participants per study).

According to a fixed-effects model, performing Chinese eye exercises was associated with an increased risk for onset of myopia (pooled OR=1.19, 95% CI: 1.02-1.39). Since there was no obvious heterogeneity (P=0.19, I2=37%) observed among selected studies. Comparisons based on the quality and attitude (serious or not) of eye exercises showed that children that performed high-quality eye exercises were associated with a markedly lower risk of developing myopia, compared with those that did not perform high-quality exercises (pooled OR=0.27, 95% CI: 0.11-0.71). According to present findings, Chinese eye exercises showed a trend towards increased risk of developing myopia. However, this effect did not reach statistical significance (pooled OR=1.36, 95% CI: 0.98-1.90). Funnel plots, Begg’s test (P=0.734, continuity corrected), and Egger’s test (P=0.277) did not reveal any obvious publication bias.

According to the authors, present findings suggest that performance of Chinese eye exercises is associated with a higher risk of myopia onset because most students performed low-quality eye exercises. However, performing high-quality Chinese eye exercises significantly lowers rates of myopia onset. Therefore, students should be encouraged to perform high-quality Chinese eye exercises.

Methodology: Inclusion criteria included 1) Participants (students with or without myopia); 2) Interventions (performed Chinese eye exercises); 3) Comparators (did not perform Chinese eye exercises); 4) Outcomes (risk ratio [RR] or odds ratio [OR] of outcomes was provided or could be calculated); and 5) Studies (cohort, cross-sectional, or randomised controlled trials). Inclusion was not restricted by study size, language, or publication type. Full-text articles were reviewed, independently, by two investigators (OMZ and ZR) to evaluate their eligibility.

The Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) statement was used as a guideline for this meta-analysis. PubMed, Embase, and the Cochrane Library databases were searched from inception to 24 July 2018. Studies were identified using a combination of relevant keywords, including acupoints, exercises, myopia and near-sightedness.

Study characteristics and data were independently extracted by two investigators. Extracted data included first author, year of publication, location, study design, sample size, adjusted/unadjusted OR (or RR), and adjusted factors. Conflicts in data extraction were resolved through discussion until a consensus was reached. The quality of cohort and case-control studies was assessed using the Newcastle-Ottawa Scale (NOS), while the quality of cross-sectional studies was assessed by the Agency for Healthcare Research and Quality (AHRQ). Studies with NOS scores ≥ 7 or an AHRQ scores ≥ 8 were considered high-quality. Review Manager Version 5.3 (Cochrane Collaboration, Nordic Cochrane Centre, Copenhagen, Denmark) and STATA v12 (StataCorp, College Station, TX, USA) were used for analyses. Outcomes were summarised as pooled ORs (or RRs) with 95% confidence intervals (CIs). Statistical heterogeneity was measured using the Chi-squared test and inconsistency index (I2) statistics. A P-value 0.1 or I2>50% indicates obvious heterogeneity. A random-effects model was used for sensitivity analysis. Potential publication bias was assessed, qualitatively, through visual inspection of funnel plots and, quantitatively, through calculation of Begg’s or Egger’s tests. A P-value 0.05 (two-tailed) indicates statistical significance.

Applicability/external validity: The authors highlighted four major limitations to consider when using the result externally. First, the number of studies included in this meta-analysis was limited. There were only five studies, involving 14,590 participants. Second, diagnosis of myopia was assessed through two different ways. Third, recall bias may also be present, considering that assessment of the quality of eye exercises in the included studies was performed using questionnaires. Fourth, other confounding factors (i.e., the duration of the eye exercises and hand hygiene), which may influence the effects of these exercises, were not taken into consideration.

Geographic focus: This meta-analysis focused on regions in China.

Summary of quality assessment:

In conclusion, medium confidence was attributed to the conclusions about the effects of this study. Despite the use of a good methodology, this review has some limitations, particularly in its robustness: the number of studies included in this meta-analysis was limited. Only five studies, involving 14,590 participants, were selected, none of which were randomised controlled trials. This was also acknowledged by the authors.