Xuxi Zhang

with interventions with follow ‐ up of 5 years or less, interventions with follow ‐ up of over 5 years had better effect on preventing DR. A previous meta ‐ analysis also had similar result that “more intensive glucose control over 5 years reduced both kidney and eye events” among patients with type 2 diabetes. 29 Moreover, we explored the effect of interventions targeting modifiable risk factors of DR on its worsening specifically, which was rarely studied by previous meta ‐ analysis studies. We found blood ‐ pressure ‐ control intervention was effective in slowing down DR worsening. However, controlling blood glucose alone had no significant effect on the control of DR worsening. A systematic review on DR also suggested that there is no evidence that rapid improvement of blood glucose control will reduce the risk of DR worsening. 30 As for follow ‐ up intervals, our results showed that compared with interventions with follow ‐ up of less than 2 years, interventions with follow ‐ up of 2 to 5 years and over 5 years had better effect on reducing the risk of DR worsening. According to the analysis on the follow ‐ up intervals, the effect of interventions on preventing DR can be observed after over 5 years, while the effect on slowing down DR worsening can be observed after 2 years, indicating that effect of interventions on delaying DR worsening could be observed earlier than that on preventing DR development. Regarding DR progression (new onset or worsening), our results indicated that multifactorial intervention also had better effect on reduction of DR progression compared with the blood ‐ pressure ‐ control intervention, glycemic ‐ control intervention and lipid ‐ control intervention. Individualized methods were adopted in the multifactorial intervention to control the progression of DR. 11, 12, 16 For example, if patients could not reach the blood pressure goal and/or blood glucose goal set at the beginning after three months, stepwise approaches were adopted based on patients’ situation. 12, 16 Additionally, we found glycemic ‐ control intervention could reduce the risk of DR progression, which is consistent with previous meta ‐ analysis. 1, 29 The control of blood pressure or lipid level alone had no significant effect on reduction of DR progression among type 2 diabetes according to our pooled results. A recent subgroup meta ‐ analysis of 4 RCTs found a borderline significant reduction in DR progression with more intensive blood pressure lowering, which is different from our finding. 31 However, they did not focus on diabetic patients and also reported substantial heterogeneity of subgroup analysis. More studies on the effect of blood pressure control on DR would be needed. As for follow ‐ up intervals, our results showed that compared with interventions with follow ‐ up of over 5 years, interventions with follow ‐ up of 2 to 5 years had better effect on reduction of DR progression. However, the heterogeneity among interventions with follow ‐ up of over 5 years on DR progression was substantial. More studies are still needed to verify this finding. Strengths and limitations This meta ‐ analysis is the first to report variation among different intervention types targeting modifiable risk factors of DR, and among different follow ‐ up intervals of interventions in 7 175 Prevention and control of diabetic retinopathy