OBJECTIVES: We performed a systematic review and meta-analysis to examine the association between HbA1c and the appearance and progression of diabetic retinopathy (DR) in T1DM. METHODS: The two electronic medical databases (MEDLINE, CENTRAL) were searched to identify all papers reporting HbA1c level and retinopathy in T1DM. Observational and randomized, controlled trials (RCTs) with at least one year of follow-up were included. Estimates were made of the adjusted relative risk (RR) of complications for an increase in HbA1c of 1%. If available data were insufficient to calculate RR, the odds ratio (OR) was estimated. Weighted mean differences (WMD) in HbA1c level  between the case group (i.e. with DR) and the control group (i.e. without DR) were also calculated. RESULTS: We identified 16 trials (4176 patients) that fulfilled the inclusion criteria. Based on four RCTs (n=1597), pooled RR for progression of DR was calculated as 1.24 (95%CI:1.01–1.52) for an increase in HbA1c of 1%. Pooled data from four observational studies (n=910) showed that RR of the incidence of DR was 1.59 (CI:1.34–1.89) for HbA1c increase of 1%. A meta-analysis of eight observational studies (n=1171) demonstrated a lower HbA1c level in patients without DR compared with patients with DR (WMD=0.82 [CI:0.69–0.96]). In addition, a meta-analysis of five observational studies revealed that mean HbA1c values were significantly lower in the group without progression of DR relative to the group with DR (WMD=1.05 [CI:0.37–1.72]). One RCT included data on visual deterioration and macular oedema; analysis demonstrated that an increase in HbA1c level of 1% increased the risk of both macular oedema (RR=1.81 [CI:1.17–2.81]) and visual deterioration (OR=2.2 [CI:1.2–3.9]). CONCLUSIONS: The results of our systematic review indicate a strong correlation between HbA1c level and appearance and progression of DR in T1DM. Thus, HbA1c may be considered an excellent surrogate endpoint for DR in T1DM.