Obesity-related early structural alterations in the retina detected by optical coherence tomography

Abstract

Aims: This retrospective cross-sectional study, using retinal spectral-domain optical coherence tomography (SD-OCT) scans, investigated obesity-related structural alterations in the retina.

Materials and methods: Ninety-two eyes of 92 healthy asymptomatic participants were categorized into two groups based on body mass index (BMI) measurements: non-obese (BMI < 25, 45%) and pre-obese/obese (BMI ≥ 25, 55%) to compare imaging parameters of different retinal layers. Structural parameters, including thickness and volume values, were obtained across distinct retinal layers segmented on SD-OCT scans.

Results: The retinal nerve fibre layer volume was lower in the high-BMI group than in the low-BMI group (p = 0.048). However, the high-BMI group presented significantly higher inner nuclear layer volume than the low-BMI group (p = 0.036). In the region analysis, the retinal nerve fibre layer volume difference was prominent in the superior (p = 0.033) and inferior (p = 0.001) parafoveal and nasal perifoveal (p = 0.041) regions, while inner nuclear layer changes were prominent in the inferior (p = 0.009) perifoveal regions. A stepwise hierarchical binary logistic regression model, controlling for age and gender, pointed to significant associations of the regionally decreased retinal nerve fibre layer volume and increased inner nuclear layer volume with high BMI (p = 0.001).

Conclusions: Retinal SD-OCT imaging detected structural alterations in distinct retinal layers between healthy, asymptomatic individuals in non-obese and pre-obese/obese groups. Besides a decreased volume of the retinal nerve fibre layer, a significant increase was detected in the inner nuclear layer volume with a high BMI, possibly due to Müller glia responses to obesity-related osmotic, metabolic and inflammatory stress, awaiting further investigation.

Keywords: obesity care; observational study; weight control; weight management.

FIGURE 1

Workflow of the study. (A) 3D en face retinal infrared fundus image. (B) 3D en face retinal infrared image demonstrating SD‐OCT central section. (C) 2D en face retinal infrared image showing OCT sections passing through the macular region of the retina. (D) The OCT section passing through the center of the macula (corresponding to the green line in panel C). (E) The same OCT section in panel D after segmenting different retinal layers (the yellow‐filled area in the red box is shown in higher magnification). (F) Demonstration of the cellular organization of the retina, corresponding to the segmented retinal layers in OCT scans.

FIGURE 2

RNFL and INL volumes in BMI age groups. (A) The table shows the volumes of the retina, RNFL and INL in younger and older age groups. (B) Violin and bubble plots of RNFL volume in BMI and age groups. (C) Violin and bubble plots of INL volume in BMI and age groups. Notably, for individuals aged ≥65 years, the RNFL volume was significantly lower in the high‐BMI group than the low‐BMI group (p = 0.038), and the INL volume was significantly higher in the high‐BMI group than the low‐BMI group (p = 0.011), while there was no difference for individuals aged <65 years. Retina, RNFL and INL volumes are shown in mm³. BMI, body mass index; INL, inner nuclear layer; RNFL, retinal nerve fibre layer; SD, standard deviation.

FIGURE 3

RNFL and INL volumes in BMI gender groups. (A) The table shows the volumes of the retina, RNFL and INL in gender groups. (B) Violin and bubble plots of RNFL volume in BMI and gender groups. (C) Violin and bubble plots of INL volume in BMI and gender groups. No statistically significant differences (p > 0.05) were observed between gender groups. Retina, RNFL and INL volumes are shown in mm³. BMI, body mass index; INL, inner nuclear layer; RNFL, retinal nerve fibre layer; SD, standard deviation.

FIGURE 4

Examples of the Early Treatment Diabetic Retinopathy Study (ETDRS) charts with thickness maps and topographic comparison of RNFL and INL volumes between low‐BMI and high‐BMI groups in different retinal regions. The region analysis used the ETDRS charts. (A) ETDRS charts show a thickness map of the whole retina, RNFL or INL in subjects from low‐BMI or high‐BMI groups. (B) Mean (±SD) RNFL volumes and p‐values of RNFL comparisons in the low‐BMI and high‐BMI groups. (C) Mean (±SD) INL volumes and p‐values of INL comparisons in the low‐BMI and high‐BMI groups. RNFL and INL volumes are shown in mm³. BMI, body mass index; I, inferior; INL, inner nuclear layer; N, nasal; RNFL, retinal nerve fibre layer; S, superior; SD, standard deviation; T, temporal.