Similar results were reported by Hua Y and Funatsu H in previous studies. Wakabayashi Y. reported that high intraocular VEGF level in patients with PDR was identified as a significant risk factor for postoperative early VH. Smith JM and Steel DHW reviewed published RCTs in recent years and cautiously concluded that anti-VEGF may reduce the incidence of early postoperative haemorrhage, suggesting that VEGF may play an important role to the complications of vitrectomy for PDR patients. Gambogic-acid However, Petrovic �� MG reported different results that vitreous levels of interleukin 8 plays a role in deteriorating visual acuity by DR progression, while vitreous levels of VEGF does not. Different opinions were also reported on the research of plasma VEGF levels in patients with PDR. Kocak N reported that levels of proinflammatory cytokines include VEGF in the vitreous were higher in the diabetic patients than the non-diabetics, while the levels of plasma cytokines were similar, indicating the expression of VEGF in the blood may not accord with that in the retina and may not correlated with the severity of PDR. However, several studies reported that in patients with PDR, VEGF concentration in plasma was also elevated as it in the vitreous fluid. Other studies shows intravitreal Tubeimoside-I bevacizumab injection may decrease the level of VEGF both in the vitreous fluid and in plasma. In our study, we found that the plasma concentration of VEGF was higher in progression group than stable group. This is actually more useful in clinic practice that testing the VEGF from patients’ blood before surgery may help the surgeons to evaluate the prognosis and risks of complications after surgery, and IVB may be considered to these patients before surgery to reduce complications, such as postoperative VH. Different results among studies may due to different include/ exclude criteria or different testing methods for VEGF concentrations. As in our study, we exclude patients using ACEI or ARB and patients with a history of PRP while none of the studies above did. Many risk factors were proved to be related to the progression of PDR, such as duration of diabetes, poor glycaemic control and uncontrolled hypertension. Certain cytokines and growth factors were also considered to be correlated with the severity of PDR, including angiotensin II. In this study, we excluded patients using ACEI or ARB to exclude the influence of it may bring to 
this study. We also compared the Clinical Data of patients and found no difference of age, sex, duration of DM, HbA1c and history of hypertension between the progression group and stable group. This strengthened the effect of the role of VEGF level in the development of PDR. Other factors may affect the result including hyperlipidemia, serum creatinine, oral anticoagulant, axial length and so on were not analyzed in this study due to the limitation of patients medical records we collected. Further investigations and prospective study are required in the future. The other criteria we excluded were patients with a history of PRP. Plasma levels of VEGF were reported significantly decreased in patients with PDR after panretinal laser photocoagulation. So we also excluded patients with a history of PRP to exclude the influence it may bring to this study. As we discussed before, the relationship between VEGF concentration in vitreous fluid and in plasma were also in controversy. A significant correlation between vitreous and plasma VEGF levels in patients with DR was reported by Baharivand N and YR Jiang. Both of their studies showed VEGF in vitreous were slightly higher than in plasma, which are in consistent with the findings of our study.