Studies regarding OCTA parameters and its associations with BCVA in the older population remain limited. Rather than invasive diagnosis tools such as fluoresceine angiography, OCTA could be incorporated into regular eye check-ups for this population, especially in those with myopia, as implicated by our findings.
Although previous OCTA studies have been performed assessing demographic, systemic and ocular alterations on retinal vessel densities in healthy eyes (2, 10), the sample sizes were small, eliciting possible selection bias in the recruitment of participants. This is the first prospective cross-sectional and population-based study of this size providing a normative database on common OCTA parameters in Hong Kong citizens aged 50 and above, while also revealing its associations with BCVA and biomarkers predicting BCVA. Our results show increasing age, being female, having a high BMI, high FAZ area and low FAZ circularity, and low MSVD in the outer ring, specifically the nasal and outer quadrants, and low CPD in the outer superior quadrant being statistically significant biomarkers in predicting a risk of lower VA in this population.
Previous studies have been performed on the associations between age and VA, with a study by Radner et al revealing the significant threshold for this association to be after the age of 55, with those aged 55–64 having a mean BCVA LogMAR of -0.12 ± 0.06 compared to -0.10 ± 0.05 in those aged 65–74(13), with 97.5% of eyes from healthy participants aged between 25–54 achieving a BCVA LogMAR of -0.1 or lower, compared to 71.25% and 51.25% in those aged between 55–64 and 65–74 respectively(13). This may be due to the increased degenerative processes occurring in the ageing population, with the most prominent one being the development of cataracts. We have therefore excluded significant cataracts from analysis.
Associations between gender and VA have also been reported, such as in the study by Emerole et al, revealing females to have poorer vision than males in a cohort of 2606 participants(14), with the majority of them aged between 40–64. A proposed theory for this could be because increased age and subsequent menopause in women results in lower oestrogen levels, with a study by Toker et al reporting a significantly lower flow velocity (p = 0.02) and higher resistive index (RI) (p = 0.001) in central retinal arteries of post-menopausal women when compared with pre-menopausal women(15), which would affect their visual acuity, due to the higher chances of developing retinal ischemia. RI also decreased, while flow velocity increased, with increasing oestradiol levels, and decreased when testosterone levels were raised in both groups(15). This corresponds to our results, with females actually having a lower central VD and CPD, perhaps as the females in our study were over the age of 50, and thus were more likely in menopause, with lower oestrogen levels.
Associations between a high BMI and visual impairment have also been reported in previous studies performed in Hong Kong, revealing statistically significant associations between the two, with obesity (BMI > = 25) having an adjusted odds ratio (aOR) of 4.00 [95% confidence interval [CI] 1.49–11.41] and hyperlipidemia having an aOR of 3.60 [95% CI 1.13–10.97] in increasing the risks of visual impairment(16).
Few studies assessed the associations between VA and OCTA parameters in healthy eyes, as most of these studies focused on pathological eyes with DR or diabetic macular oedema (DMO). A study by Abdelshafy et al revealed significant negative correlations between VA and VD in the SCP and DCP, especially in the whole enface region for the superficial region and parafovea region in the DCP (17). This differs from our results, where only MSVD in outer ring or the parafovea region was associated with VA, instead of the whole enface region. The most likely reason for this could be due to the fact that only healthy eyes without retinal pathologies were assessed our study, as the study by Abdelshafy et al assessed eyes with and without DR, although other reasons for the differing findings could be due to the ethnicity of the population resulting in biological variation, along with different inclusion and exclusion criteria, OCTA machine with different algorithms used, and variations in sample size, as only 60 eyes of the Egyptian population were evaluated (17). Correlations between the superficial FAZ area and VA have also been reported, revealing a negative relationship between the two (r=-0.54, p = 0.03)(18), which is consistent with our findings. However, this study also assessed eyes with retinal pathologies, in particular, central retinal vein occlusion.
Associations between MSVD, MCPD and VA have also been reported in another study by Leng et al, reporting MSVD decreasing with myopia and longer axial length (p = 0.021 and 0.027) (7), which is consistent with our findings. Similarly, another study by Milani et al revealed myopic eyes being negatively correlated with MSVD in Chinese participants (p < 0.001), with the high myopia group (defined as having a spherical equivalent (SE) > 6 diopters) having an MSVD of 19.64% compared to 25.64% in the control group with no or mild myopia (SE >-3 diopters and < 3 diopters)(19). The exact mechanism of how myopia and reduced BCVA decreases MSVD and CPD remains unconfirmed, but the general hypothesis is that as the axial length is increased, retinal vessels are similarly straightened and narrowed, reducing the overall blood flow and perfusion(20). Another proposed mechanism by Wang et al is that as the eyeball elongates in progressing myopia, the retinal tissue concurrently stretches and thins out, reducing the oxygen demand, leading to reductions in the MSVD and CPD(21). Despite its unconfirmed mechanism, multiple studies have reported myopic eyes having significant reductions in MSVD and choriocapillaris circulation(22–24).
Out of all the variables associated with BCVA in our multivariate model, the OCTA biomarkers predicting BCVA includes: FAZ area and circularity (β = 0.039 and − 0.034, p = 0.01 and 0.024 respectively), MSVD in the outer ring (β=-0.513, p < 0.001), specifically in the nasal and temporal quadrants (β = -0.226 and − 0.259, p < 0.001 for both), and low CPD in the outer superior quadrant (β= -0.123, p = 0.016), although systemic factors such as age, gender and BMI should also be taken into consideration for both older aged individuals and ophthalmologists. Reasons for these specific quadrants having a significantly lower MSVD and MCPD remains unknown, and warrants further research. However, these specific regions should be given greater attention when reviewing OCTA results.
Our study is the first one assessing the associations between OCTA parameters in specific ETDRS quadrants with BCVA in healthy eyes, as well as in this specific population aged 50 and above. However, our study presents with some limitations. First, selection bias may be present, as subjects participating in this programme are voluntary and may be more risk adversed individuals. Moreover, axial length was not measured in our study, which is an important indicator for visual acuity. OCTA also requires high levels of cooperatively from participants, such as to remain still during the procedure, to reduce motion artifacts and poor quality scans, which may be difficult for older subjects with motor disorders. The DCP was also not assessed in our study, so associations of BCVA with parameters in the DCP could not be examined.
However, the major strength of our study is its large sample size, increasing the power of the study, and increasing the accuracy of our linear regression models. The population-based and prospective nature of the study, and how it specifically focuses on the older population also aids in providing a normative database and ocular and systemic factors influencing the MSVD in this population. Our strict inclusion and exclusion criteria also enabled only healthy eyes to be assessed. We also reported VD in mm− 1, instead of measuring the area occupied by perfused vessels divided by the total area (%), because this method has been suggested to improve the accuracy of measurements when quantifying finer vessels(25), by measuring the vessel length per surface area in mm/mm2 (mm− 1). OCTA scans at the level of the SCP also include large arteries and veins, instead of just assessing retinal vessels and retinal microvasculature, and assessing the total area of perfused vasculature may overestimate retinal tissue perfusion(25), so choosing this method to assess the MSVD is more precise. Therefore, our study adds significant and accurate information to the current knowledge surrounding the normal retinal microvasculature.
In conclusion, our present study revealed the average MSVD in adults aged 50 and above. Increased age, being female, having a high BMI, high FAZ area, low FAZ circularity, low MSVD in the outer ring, specifically the nasal and outer quadrants, and low CPD in the outer superior quadrant are independently significant risk factors for reductions in VA and should be considered in clinical practice and in citizens of this population, while these OCTA parameters can be used as biomarkers to predict deteriorations in VA.