AIM:To look for the minimum effective dosage of intravitreal injection with triamcinolone acetonide(TA) for diabetic macular edema(DME).METHODS:A total of 45 eyes in 38 DME patients who were diagnosed by FFA were collected.They were divided into two groups:22 eyes in group A and the other 23 eyes in group B.Different dosage of 40g/L TA(2mg or 4mg) were respectively used for intravitreal injection for the two groups.Before the treatment,the average best-corrected visual acuity(BCVA) was 0.06±0.06 and 0.06±0.05 in the two groups respectively(t=0.076,P=0.94),and all the intraocular pressure(IOP) was below 21mmHg.The average follow-up duration was 6 months after treatment.The BCVA,IOP,changes of lens and ocular fundus and FFA before and after treatment were observed.RESULTS:The average BCVA after treatment was 0.13±0.05 and 0.12±0.04 in the two groups respectively.The differences caused by treatment for each group were statistically significant(t=-4.39,P=0.00;t=-5.75,P=0.00).The hypertension rate was 9%(2 eyes) in group A and 39%(9 eyes) in group B(χ2=3.99,P=0.04).For 19 eyes in group A and 22 eyes in group B,the treatment showed effective by FFA(χ2=0.32,P=0.56).The recurrence rate was 23% and 13% respectively,the difference was not statistically significant(χ2=0.21,P=0.64).No refractory glaucoma and cataract development eye occurred.Intravitreous injection with TA again was effective if macular edema recurred.CONCLUSION:No matter the dosage is 2mg or 4mg,intravitreous injection with TA is a promising therapeutic method for DME,its short-term recurrence rate is not obviously different.But 2mg TA injection may cause less increased IOP than 4mg TA injection.