Free papers detail | 11th Asia Pacific Tele Ophthalmology Society Symposium

Title
Sleep, Melatonin, and Diabetic Retinopathy: A Mediation and Longitudinal Study

Authors
Xixuan Zhao, Haidong Zou

Presenting
Xixuan Zhao

PURPOSE:
To investigate the relationships among sleep disturbance, plasma melatonin levels, and diabetic retinopathy (DR), and to evaluate whether melatonin mediates sleep–DR associations and predicts incident DR.

METHODS:
This study was conducted within the community-based cohort using combined cross-sectional and longitudinal analyses. Sleep characteristics were assessed using the Pittsburgh Sleep Quality Index (PSQI) and Epworth Sleepiness Scale (ESS). Multivariable logistic regression examined associations with DR prevalence. Counterfactual-based mediation analyses assessed the mediating role of melatonin. Participants without DR at baseline were followed for 1 year, and Cox regression models evaluated incident DR risk. Receiver operating characteristic (ROC) analyses assessed predictive performance and the incremental value of melatonin beyond conventional risk factors.

METHODS:

This study was conducted within the community-based cohort using combined cross-sectional and longitudinal analyses. Sleep characteristics were assessed using the Pittsburgh Sleep Quality Index (PSQI) and Epworth Sleepiness Scale (ESS). Multivariable logistic regression examined associations with DR prevalence. Counterfactual-based mediation analyses assessed the mediating role of melatonin. Participants without DR at baseline were followed for 1 year, and Cox regression models evaluated incident DR risk. Receiver operating characteristic (ROC) analyses assessed predictive performance and the incremental value of melatonin beyond conventional risk factors.

RESULTS:
Among 249 participants, higher PSQI and ESS scores and lower plasma melatonin levels were independently associated with higher DR prevalence. Melatonin significantly mediated the associations between sleep disturbance and DR, with predominantly indirect effects for PSQI and partial mediation for ESS, while sleep measures did not mediate the melatonin–DR relationship. During follow-up, 11 incident DR cases occurred among 230 participants without baseline DR. Lower baseline melatonin was associated with higher incident DR risk (adjusted HR=2.744). Baseline melatonin showed moderate predictive ability (AUC=0.713), and adding melatonin to a conventional model improved discrimination (AUC: 0.653 to 0.741).

RESULTS:

Among 249 participants, higher PSQI and ESS scores and lower plasma melatonin levels were independently associated with higher DR prevalence. Melatonin significantly mediated the associations between sleep disturbance and DR, with predominantly indirect effects for PSQI and partial mediation for ESS, while sleep measures did not mediate the melatonin–DR relationship. During follow-up, 11 incident DR cases occurred among 230 participants without baseline DR. Lower baseline melatonin was associated with higher incident DR risk (adjusted HR=2.744). Baseline melatonin showed moderate predictive ability (AUC=0.713), and adding melatonin to a conventional model improved discrimination (AUC: 0.653 to 0.741).

CONCLUSIONS:
Sleep disturbance is associated with increased DR risk, and melatonin may act as a key circadian mediator linking sleep disruption to retinal microvascular damage. Lower melatonin independently predicts short-term incident DR and improves risk prediction beyond traditional factors.