Dear Editor,

I have read the research article “Physical activity and risk of diabetic retinopathy: a systematic review and meta-analysis” by Ren et al. [1]. The authors conducted a meta-analysis to investigate the effect of physical activity (PA) on diabetes mellitus (DM). Pooled risk ratios (RRs) (95% confidence intervals [CIs]) of physical activity (PA) for incident diabetic retinopathy (DR) and vision-threatening DR were 0.94 (0.90–0.98) and 0.89 (0.80–0.98), respectively. In addition, pooled RR (95% CI) of sedentary behavior for incident DR was 1.18 (1.01–1.37). I have some concerns about the study.

First, the authors evaluated the effect of moderate-intensity PA on the risk of DR, and pooled RR (95% CI) was 0.76 (0.58–1.00). The level of significance was marginal, which was inconsistent with significant dose–response relationship between increased PA and lower risk of DM [2]. DR is generally observed in patients with advanced DM, and progression of vascular damage by DM might not be protected by moderate-intensity PA. To verify the association, further studies are needed for the risk assessment by a meta-analysis.

Second, PA and sedentary behavior presented the opposite results for DR, which was logical and understandable. Relating to the methodology of PA detection, objective evaluation of PA might be important for preventing incident DR and its progression [3, 4].

Finally, Yaribeygi et al. [5] reported that aerobic exercise could prevent the development of diabetic complications, which was partly explained via attenuation of the major molecular adverse effect by DM. Ren et al. [1] did not refer to the type of exercise, and molecular mechanism of suppressing inflammatory response and oxidative stress of retina by exercise should be verified by further studies.