ObjectiveTo observe the safety and efficacy of Keluoxin capsules in the treatment of moderate to severe non-proliferative diabetic retinopathy (NPDR). MethodsAn open-label, multi-center, single-arm, phase Ⅱa clinical trial. From May 2014 to December 2016, the patients diagnosed with moderate to severe NPDR who received Keroxin treatment in General Hospital of Central Theater Command, Affiliated Eye Hospital to Nanchang University, Xiyuan Hospital of China Academy of Chinese Medical Sciences, and Eye Hospital China Academy of Chinese Medical Sciences were divided into moderate NPDR group and severe NPDR group. The baseline data of the patients were obtained, best-corrected visual acuity (BCVA), optical coherence tomography, fundus fluorescein angiography and fundus photography were performed. On the basis of maintaining the original diabetes treatment, all patients took Keluoxin capsules orally for 24 weeks; 24 weeks after treatment was used as the time point for evaluating the efficacy. BCVA letters, central macular thickness (CMT) and 6 mm diameter total macular volume (TMV), retinal vascular leakage area, and retinal non-perfusion (RNP) area within an average diameter of 6 mm were compared between the two groups at baseline and 24 weeks after treatment. Independent sample Mann-Whitney U test was used to compare continuous variables between groups. Categorical data were compared by χ2 test. ResultsA total of 60 NPDR patients and 60 eyes were included, 9 cases were lost to follow-up, and 51 cases and 51 eyes were finally included, including 37 eyes in the moderate NPDR group and 14 eyes in the severe NPDR group, respectively. At baseline, BCVA in moderate NPDR group and severe NPDR group were (80.1±6.8), (81.4±6.3) letters, respectively. CMT were (249.5±32.1), (258.9±22.2) μm, respectively. TMV were (8.79±1.09), (8.95±1.31) mm3, respectively. Retinal vascular leakage areas were (7.69±10.63), (10.45±7.65) mm2, respectively. RNP area were (2.48±5.74), (10.63±20.06) mm2, respectively. There were 11 (29.7%, 11/37) and 4 (28.6%, 4/14) eyes with diabetic macular edema (DME), respectively; 24 weeks after treatment, BCVA in moderate NPDR group and severe NPDR group increased by (1.3±5.2), (3.2±3.0) letters, respectively. Compared with baseline, there was a statistically significant difference in the severe NPDR group (t=－3.986, P=0.033). CMT were (252.1±45.6), (269.8±57.2) μm, respectively. There were no significant differences compared with baseline (t=－0.567, －0.925; P＞0.05). TMV were (9.96±1.16), (10.09±1.32) mm3, respectively. There were no significant differences compared with baseline (t=－0.996, －1.304; P＞0.05). Retinal vascular leakage area decreased (0.19±6.90), (1.98±7.52) mm2, respectively. There were no significant differences compared with baseline (t=0.168, 0.983; P＞0.05). RNP area were (3.01±6.47), (10.36±19.57) mm2, respectively. Compared with baseline, the differences were statistically significant (t=－1.267, 0.553; P＞0.05). There were 8 (21.6%, 8/37) and 3 (21.4%, 3/14) eyes with DME, respectively. Compared with baseline, the difference was statistically significant (χ2=11.919, 4.571; P=0.001, 0.033). ConclusionKeluoxin capsules can stabilize or improve BCVA, CMT, TMV and RNP area in patients with moderate and severe NPDR, and reduce the area of retinal vascular leakage.