Your study you linked compares the diabetic population taking GLP-1s against a healthy population. Of course they're going to find more eye problems in the diabetic population, diabetes damages the eye.
Correlation does not imply causation. Especially this one.