A meta-analysis of alterations in the retina and choroid in high myopia assessed by optical coherence tomography angiography

Authors: Wang XQ, Zeng LZ, Chen M, Liu LQ.

Geographical coverage: China and South Korea. 

Sector: Biomedical

Sub-sector: Risk

Equity focus: None

Study population: General population with myopia

Review type: Other review

Quantitative synthesis method: Meta-analysis

Qualitative synthesis method: Not applicable

Background: High myopia (HM) is a risk factor for several pathological structural changes in retinal and choroidal thickness or vessel. To date, studies examining changes in retinal and choroidal microvasculature circulations in HM have yielded inconsistent results.

Objectives: Evaluate alterations in retinal and choroidal thickness, and capillary microvasculature using optical coherence tomography angiography (OCTA) in patients with HM.

Main findings:

In summary, this review discovered that both the superficial and deep vessel densities of the macula and parapapillary regions were lower in high myopia (HM) compared to control eyes. Additionally, the thickness of the parafoveal retinal nerve fiber layer (RNFL), parapapillary RNFL, and sub-foveal choroid, as well as the chorio-capillary density, were also found to be reduced in HM.

In total, 11 eligible articles, including 525 HM eyes and 491 normal control eyes with sufficient data, were selected for the meta-analysis. Seven of the included studies had a quality score of 7 and four had quality score of 8 (7 was the minimum for inclusion). Sensitivity analysis for each outcome showed that no articles had a high influence, indicating the results of the review are stable. From the funnel plot analysis, authors found no correlation between study size and effect size or any other evidence of publication bias.

The review found that in high myopia (HM) eyes, both whole superficial and deep vessel densities were lower compared to control eyes. No significant changes were observed in foveal and parafoveal superficial vessel densities, but parapapillary superficial vessel density was lower in HM. Foveal avascular zone (FAZ) showed no significant difference, while a small increase in foveal thickness was noted in HM. Parafoveal retinal nerve fiber layer (RNFL) thickness was significantly thinner in HM by 5.36 μm, and parapapillary RNFL thickness was slightly lower. Sub-foveal choroidal thickness (SFCT) in HM was significantly lower than controls, but no significant changes were observed in choriocapillaris density.

The authors suggest that OCTA is a non-invasive and effective method for tracking the progression of myopia, implying its potential utility in future research. They also recommend that additional studies on HM should encompass a broader area to better understand changes in the macular microvasculature.

Methodology

The review included original English research articles that were either cohort or cross-sectional studies. HM was defined as a spherical equivalent (SE) refractive error ≥ -6.0 D and/or axial length (AL) ≥26.0 mm, without pathological changes. The control group had an SE between +3.0 D and -3.0 D. OCTA data was reported as mean ± standard deviation, with differences between HM and controls. The OCTA scan size was a 3×3 mm² region. At least two of the listed outcomes were included. Exclusions were case reports, conference abstracts, letters, reviews, meta-analyses, studies with insufficient data, eyes with diseases other than myopia, and animal experimental and optics studies.

Researchers searched PubMed, EMBASE, and Cochrane Library for studies on myopia using OCTA, focusing on outcomes like foveal avascular zone, vessel density, RNFL thickness, foveal thickness, sub-foveal choroidal thickness, and chorio-capillary density. The search included English articles up to June 2020. Two researchers resolved any inclusion conflicts by consensus.

Two authors independently extracted data from the included studies, including authorship, publication year, study area, number of eyes, age, AL, refractive error, OCTA device, and outcomes. Discrepancies were resolved by consensus. Study quality was assessed using the Newcastle-Ottawa Scale (NOS), with scores >7 indicating high quality and inclusion in the final analysis.

Statistical analysis was performed using Review Manager 5.3. For continuous variables, outcomes were reported as the mean ± standard deviation and the mean difference (MD) with a 95% confidence interval. To obtain reliable results, heterogeneity was evaluated using the I² test. If the homogeneity test showed p ≥0.1 and I² ≤50%, which indicated a low homogeneity between the included studies, a fixed-effects model was used for the analysis. If the value of I² was >50% or p value 1, which suggested a high heterogeneity, then a random-effects model was applied to the data. A funnel plot was employed to assess publication bias.

Applicability/external validity: The researchers note that their findings differ from those of previous similar studies. They suggest that potential methodological weaknesses in these earlier studies could have contributed to these discrepancies.

Geographic focus: While the review includes articles from a diverse range of high-income and low-to-middle-income countries, it does not explore how the findings might vary across different geographical contexts. This could be a valuable aspect to consider in future research to enhance the applicability and relevance of the results.

Summary of quality assessment:

While data analysis methods were consistently robust, there were important limitations in the search strategy process. The search was confined to published articles and those written in English, potentially excluding relevant studies in other languages. Furthermore, it’s unclear whether the reference lists of key articles were reviewed for additional sources. The time frame of the review is also not specified, which could impact the comprehensiveness of the findings. Given these limitations, our confidence in the review’s conclusions is low.